E FFECTS OF MATERNAL REGULATORY ATTEMPTS AND MIND - MINDEDNESS ON - REGULATION By Neda Senehi A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Human Developm ent and Family Studies Doctor of Philosophy 2018 ABSTRACT EFFECTS OF MATERNAL REGULATORY ATTEMPTS AND MIND - MINDEDNESS ON - REGULATION By Neda Senehi The toddler years are a watershed developmental period for emergence of emotional an d behavioral regulation (i.e., self - regulation; SR), and the gradual transition from other - supported self - regulation to more autonomous forms of regulation. Negative emotional expression and delay of gratification are developmentally salient aspects of sel f - regulation, particularly since expectations for frustration tolerance and waiting increase during toddlerhood. As toddlers develop regulatory capacities, Maternal Regulatory Attempts (MRAs: the ways in which parents ehaviors) are thought to play an important role in self - regulatory development. While MRAs have been linked to SR in toddlers, the role of maternal Mind - has been ne glected despite recent theoretical and empirical evidence for its regulatory effects. G ratification (DG: behavioral regulation) during a two - minute delay of gratification task. All study predictors (MRA, MM, and TRS) and study outcomes (ENE and DG) were observed and coded in twelve 10s intervals. A series of multilevel models with 10s - repeat ed measurement occasions nested within 134 mother - toddler dyads (67 girls; Mage = 25.77 months, SDage = 1.60) enrolled in the National Early Head Start Research and Evaluation P roject, were used. gged (e.g., previous 10s interval) MRAs, lagged MM, and lagged TR S in random coefficient models . Findings provide evidence for effects of MRAs, specifically more robust effects of lagged maternal positive emotional reactions (e.g., laughs with child) on to and lagged maternal negative emotional reactions (e.g., becomes upset herself) on both ENE and and dependent use of non - verbal distractions and lagged mother - directe d verbalizations of self - - dependent physical comfort were related to higher ability to wait. Study results did not provide evidence for effects of lagged MM on ENE or DG, but can be used to inform future directions. Study findings provide evidence for contributions of caregiving context and point to important temporal contingencies and interactional effects to be considered i n future investigations of MRAs, MM, and SR. iv This dissertation is dedicated to my parents for carrying me on their back. And to my brother for his vision, courage, and paving the way. v ACKNOWLEDGEMENTS I would like to extend my deepest gratitude to my co mmittee members without whom I could not have comp leted this dissertation or my doctoral program. Thank you all for helping me stay grounded and focused during this long journey. Your mentorship has helped me overcome barriers both within and outside of my self, and, for that I will forever be grateful to you. I would like to thank my one - in - a - million graduate advisor, Dr. Brophy - Herb, for her unrelenting support and confidence in my vision. Her fierce leadership, guidance, and expertise, have shaped me in to a scholar beyond my wildest dreams. I would like to thank Dr. Vallotton, for passionately entertaining every one of my ideas and being deeply instrumental in my ability to critically observe, think, and write about human development. I would like to tha nk Dr. Bowles, for his consistent mentorship and for challenging me to examine and fine - tune my thinking about developmental processes, and, Dr. Lonstein for his continued support, constructive feedback, and investment in my success throughout this process . I am enormously grateful to my BEES and HDFS fami lies for holding my hand, having my back, and providing opportunities for my personal and professional growth every step of the way. I would like to thank my undergraduate coding team for their diligent c oding efforts that brought this data to life. I would also like to thank my eight - grade English teacher, Mrs. Kaminsky, for making learning infinitely empowering, and, my first graduate mentor, Dr. Shepherd - Look, for her endless support and inspirational w isdom. I also want to thank my personal mentors, Daren and S a rita, for their unconditional love, and Dr. Snyder, for her brilliant teachings that gave me wings. vi TABLE OF CONTENTS LIST OF TABLES x L IST OF FIGURES xvi CHAPTER 1: Introduction 1 The Importance of Self - 1 Origins of Individual Differenc es in Self - Regulation: The Caregiving Contex t 2 Defining Self - Regulation 5 Emotion Regulation 7 Delay of Gratification 8 Self - Regulation during Toddler Years 9 - Regulation 10 Purpose and Rationale 11 C HAPTER 2: Literature Review 14 Maternal Regulatory Attempts (MRAs) 14 Unique Contributions of MRA to Toddler Emotion Regulation and Delay of Gratification 15 Mind - Mindedness (M M) 18 Unique Contributions of Maternal Mind - ion and Delay of Gratification 21 Attuned M ind - R elated C omments 21 Non - A ttuned M ind - R elated C omments 25 Contributions of Maternal Regulatory Attempts and Mind - Mindedness 26 Emotion Regulation and Delay of Gratification 27 Current Study 28 Covariates 32 Overview of Research Ques tions 33 Question 1A and 1B 33 Hypothesis 1A 33 Hypothesis 1B 33 Question 2A 33 Hypothesis 2A 33 Question 2B 34 Hypothesis 2B 34 Question 3A and 3B 34 Hypothesis 3A 34 Hypothesis 3B 34 Question 4A 34 Hypothesis 4A 35 Question 4B 35 Hypothesis 4B 35 vii CHAPTER 3: Method 36 Participants 3 6 Procedure 3 7 Delay of Gratification Task 37 Measures 3 8 Maternal Regulatory Attempts 38 Maternal Mind - Mindedness 40 Toddler Regulatory Strategies 45 motion 50 52 Covariates 52 Coding 53 Plan of Analysis 57 - O) 58 62 CHAPTER 4: Resul ts 66 Missing Data 66 Assumptions 67 Descriptive Statistics 68 Toddler Expression of Negative Emotion and Delay of Gratification Per 10s Interval and Overall Task 68 Maternal Regulatory Attempts and Mind - Mindedness Per 10s Interval and Overall Task 69 Per 10s Interval and Overall Task 73 Independent Regulatory Strategies Per 10s Interval and Overall Task 73 Dependent Regulatory Strategies Per 10s Interval and Overall Task 74 Correlational Analyses 76 Lagged Maternal Variables and Overall Expression of Negative Emotion 76 78 Lagge d Maternal Variables and Delay of Gratification 79 81 Expression of Negative Emotion, Delay of Gratification, Child Gender, and Express ive Language 82 Multilevel Analyses 83 Research Question 1A: The Random Intercept Model Predicting Expression of Negative Emotion from Lagged Maternal Regulatory At tempts and Mind - Mindedness 85 Lagged Maternal Verbal Comfort 88 Lagged Maternal Initiated Distraction 89 Lagged Maternal Positive Commands 90 Lagged Maternal Posit ive Emotional Reactions 91 Question 1A Summary of Results 93 Research Question 1B: The Random Intercept Model Predicting Expression of Negative Emotion from Lagged Maternal Regulatory Attempts and Min d - Mindedness 94 Lagged Maternal Punitive Reactions 95 Lagged Maternal Minimizing 96 viii Lagged Maternal Physical Restraint 97 Lagged Maternal Negative Emotional Rea ctions 99 Question 1B Summary of Results 100 Research Question 2A: The Random Intercept Model Predicting Expression of Negative 102 Lagged Toddler Initiated Verbal Distraction 103 Lagged Toddler Initiated Verbal Keys 104 Lagged Toddler Initiated Verbal Self - Contro l 105 Lagged Toddler Initiated Verbal Desire 105 Lagged Toddler Self - Comfort 107 Question 2A Sum mary of Results 108 Research Question 2B: The Random Intercept Model Predicting Expression of Negative Emotion from Lag 109 Lagged Toddler Verbal Distraction Bids to Mom 111 Lagged Toddler Joined Verbal Distraction 112 Lagged Toddler Joined Non - Verbal Distraction 113 Lagged Toddler Verbal Self - Control Bids to Mom 115 Question 2B Summary of Results 116 Research Question 3A: The Random Intercept Model Predicting Overall Touch from Lagged Maternal Regulato ry Attempts and Mind - Mindedness 118 Lagged Maternal Physical Comfort 120 Lagged Maternal Initiated Distraction 121 Lagged Maternal Positive Commands 122 La gged Maternal Positive Emotional Reactions 123 Question 3A Summary of Results 125 Research Question 3B: The Random Intercept Model Predicting Overall Touch from Lagged Maternal Regulatory Attempts and M ind - Mindedness 127 Lagged Maternal Negative Commands 128 Lag ged Maternal Physical Restraint 129 Lagged Maternal Negative Emotional Reactions 130 Question 3B Summary of Results 131 Research Question 4A: The Random Intercept Model Predicting Overall Touch from Lagged 133 Lagged Toddler Initiated Ver bal Distraction 134 Lagged Toddler Initiated Non - Verbal Distraction 135 Lagged Toddler Initiated Verbal Keys 136 Lagged Toddler Initiated Verbal Self - Control 138 Lagged Toddler Self - Comfort 138 Question 4A Summary of Results 139 Research Question 4B: The Random Intercept Model Predicting Overall Touch from Lagged 140 La gged Toddler Verbal Distraction Bids to Mom 142 Lagged Toddler Physical Comfort Seeking 143 Question 4B Summary of Results 144 CHAPT ER 5: Discussion 147 Overview 147 ix Unique Effects of Maternal Regulatory Attempts on Emotion Regulation 148 Lagged Maternal Emotional Reactions 149 Lagged Mate rnal Punitive Reaction s and Minimizing 150 Lagged Maternal Verbal Comfort 151 Lagged Physical Comfort and Phy sical Restraint 152 Lagged Positive and Negative Commands 153 Lagged Maternal Distraction and Verbal Orientation to Delay 153 Unique Effects of Mind - Mindedness on Emotion Regulation and Delay of Gratification 155 Appropriate Mind - Related Comments 155 Non - Attuned Mind - Related Comments 157 Gratification 158 istraction 159 - Comfort and Comfort Seeking 159 160 Strengths and Limitations 161 Coding Related Challenges 163 Future D irections 164 APPENDICES 172 APPENDIX A Maternal Reg ulatory Attempts & Mind - Mindedness Overlap 17 3 APPENDIX B Coder Traini ng and Inter - Coder Reliability 17 6 APPENDIX C Equations 1 80 APPENDIX D Patterns of Study Variables across 12 Inte rvals 219 APPENDIX E Descriptive Statistics across Task (120s) 228 APPENDIX F Correlations of Outcomes with Current - Interval Predictors 234 APPENDIX G Independence of Errors and Intraclass Correlations 241 APPENDIX H Model Specification of Covariates 244 APPENDIX I Models of Non - Significant Predictors 247 APPENDIX J Acknowledgement of Grants 277 R EFERENC ES 279 x LIST OF TABLES Table 1 Conceptual and Operational Definition s of Maternal Variables 30 Table 2 Conceptual and Operational Definitions of Toddler Variables 31 Table 3 Description of Maternal Regulatory Attempts 43 Table 4 Description of Maternal Mind - mindedness 45 Table 5 Description of Toddler Independent Regulatory Strategies 48 Tabl e 6 Description of Toddler Dependent Regulatory Strategies 49 Table 7 Description of Toddler Expression of Negative Emotion and Delay of Gratification 5 2 T able 8 Interobserver Reliability Estimates for Maternal Regulatory Attempts, Appropriate Mind - Related, and Non - Attuned Mind - Related Comments 55 Table 9 56 Table 10 Interobserver Reliability Estimates for Toddler Expression o f Negative Emotion and Delay of Gratification 57 Table 11 Maternal and Toddler Predictors of Expression of Negative Emotion al Expression and Overall Touch 59 Table 12 Descriptive Statistics for Toddler Expression of Negative Emotion and Overall Touch 69 Table 13 Descriptive Statistics of for Maternal Regulatory Attempts, Appropriate and Non - Attun ed Mind - Related Comments 72 75 Table 15 Correlations for Lagged Maternal Regulatory Attempts, Appropriate and Non - Attuned Mind - R elated Comments with Expressi on of Negative Emotion (n=1437) 77 Table 16 Correlations for Lagged Toddler Regulatory Strategies with Expression of Negative Emotion (n = 1437) 79 Table 17 Correl ations for Lagged Maternal Regulatory Attempts, Appropriate and Non - Attuned Mind - Related Comments with Overall Touch (n=1437) 80 xi Table 18 Correlations for Lagged Toddler Regulatory Strategies with Overall Touch (n = 1437) 82 Table 19 Intercorrelations for Expression of Negative Emotion with Overall Touch, Toddler Gender, and Vocabulary Production 83 Table 20 Overview of Research Questio ns: Predictors X Outcomes 85 Table 21 Question 1A. Do Lagged Maternal Regulatory Attempts and Lagged Mind - Mindedness - 3) 87 Table 22 Effects of Lagged Ma ternal Verbal Comfort on Overall Expression of Negative Emotion 89 Table 23 Effects of Lagged Maternal Initiated Distraction on Overall Expression of Negative Emotion 90 Table 24 Effects of Lagged Maternal Positive Commands on Overall Expression of Negative Emotion 91 Table 25 Effects of Lagged Maternal Positive Emotional Reactions on O verall Expression of Negative Emotion 92 Table 26 Question 1B. Do Lagged Maternal Regulatory Attempts and Lagged Mind - Mindedness - 3) 94 Table 27 Effects of Lagged Punitive Reactions on Overall Expression of Negative Emotion 96 Table 28 Effects of Lagged Minimizing on Overall Expression of Negative Emotion 97 Table 29 Effects of Lagged Maternal Physical Restraint on Overall Expressio n of Negative Emotion 98 Table 30 Effects of Lagged Negative Emotional Reactions on Overall Expression of Negative Emotion 100 Table 31 Expression of Negative Emotion (C ontinuous: 0 - 3) 102 Table 32 Effects of Lagged Toddler Initiated Non - Verbal Distraction on Overall E xpression of Negative Emotion 104 Table 33 Effects of Lagged Toddler Initiated Verbal Keys on Overall Expression of Negative Emotion 105 xii Table 34 Eff ects of Lagged Toddler Initiated Verbal Desire on Overall Expression of Negative Emotion 106 Table 35 Effects of Lagged Toddler Self - Comfort on Overall Expression of Negative Emotion 108 Expression of Negative Emotion (Continuous: 0 - 3 ) 110 Table 37 Effects of Lagged Toddler Verb al Distraction Bids to Mom on Overall Expression of Negative Emotion 112 Table 38 Effects of Lagged Toddler Joined Verbal Distraction on Overall Expressio n of Negative Emotion 113 Table 39 Effects of Lagged Toddler Joined Non - Verbal Distraction on Overall Expression of Negative Emotion 114 Table 40 Effects of Lagged Toddler Self - Control - Bids to Mom on Overa ll Expression of Negative Emotion 116 Table 41 Question 3A. Do Lagged Maternal Regulatory Attempts and Lagged Mind - Mindedness 119 Table 42 Effects of Lagged Maternal Physical Comfort on Overall Touch 121 Table 43 Effects of Lagged Maternal Initiated Distraction on Overall Touch 122 Table 44 Effects of Lagged Positive Commands on Overall Touch 123 Table 45 E ffects of Lagged Maternal Positive Emotional Reactions on Overall Touch 124 Table 46 Question 3B. Do Lagged Maternal Regulatory Attempts and Lagged Mind - Mindedness 127 Table 47 Effects of Lagged Maternal Negative Commands on Overall Touch 128 Table 48 Effects of Lagged Maternal Physical Restraint on Overall Touch 130 Table 49 Effects of Lagged Maternal Negative Emotional Reactions on Ov erall Touch 131 Touch (Count) 133 Table 51 Effects of Lagged Toddler Initiated Verbal Distraction on Overall Touch 135 xiii Table 52 Effects of Lagged Toddler Initiated Non - Verbal Distraction on Overall Touch 136 Table 53 Effects of Lagged Toddler Initiated Verbal Keys on Overall Touch 137 Table 54 Effects of Lagged Toddler Self - Comfort on Ove rall Touch 139 Touch (Count) 141 Table 56 Effects of Lagged Toddler Verbal Distraction - Bids to M om on Overall Touch 143 Table 57 Effects of Lagged Toddler Physical Comfort Seeking on Overall Touch 144 Table A1 Percentages of Appropriate and Non - Attuned Comments Further Coded for MRA Overlap and Mental State Category 175 Table D1 Analysis of Variance (ANOVA) Between Toddler Expression of Negative Emotion and Overall Touch with Elapsed Time Across 12 Intervals 222 Table D2 Analysis of Variance (ANOVA) Between Maternal Regulatory Strategies and Elapsed Time Across 12 Intervals 225 Table D3 Analysis of Variance (ANOVA) Between Toddler Regulatory Strategies and Elapsed Time Across 12 Intervals 227 Table E1 Descriptive Statistics for Toddler Expression of Negative Emotion and Overall Touch Across Task (All 120 Seconds) 229 Table E2 Descriptive Statistics of for Maternal Regulatory Attempts, Appropriate a nd Non - Attuned Mind - Related Comments across Task (All 120 Seconds) 230 Table E3 Descriptive Statistics for Toddler Regulatory Strategies Across Task (A ll 120 Seconds) 232 Table E4 Strategies and Dependent Regulatory Strategie s Across Task (All 120 Seconds) 233 Table E5 Comparison between means for frequency of toddlers Independent Regulatory Strategi es and Dependent Regulatory Strategies per 10 - Second Interval 233 Table F1 Correlations for Current Interval Maternal Regulatory Attempts, Appropriate and Non - Attuned Mind - Related Comments with Expression of Negative Emotion (n=1571) 236 Table F2 Correlations for Current Interval Toddler Regulatory Strategies with Expression of Negative Emotion (n = 1571) 237 xiv Table F3 Correlations for Current Interval Maternal Regulatory Attempts, Appropriate and Non - Attuned Mind - Related Comments with Touch (n=1571) 239 Table F4 Correlations for Current Interval Toddler Regulatory Strategies with Expression of Negative Emotion and Overall Touch (n = 1571) 2 40 Table G1 Intraclass correlations for Intercept - Only Models for Expression of Negative Emotion and Delay of Gratification 243 Table H1 Variable names and labels for study covariates 24 6 Table I1 Effects of Lagged Maternal Physical Comfort on Overall Expression of Negative Emotion 249 Table I2 Effects of Lagged Maternal Joined Distraction Comfort on Overall Expression of Negative Emotion 250 Table I3 Effects of Lagged Maternal Verbal Orientation to Delay on Overall Expression of Negative Emotion 251 Table I4 Effects of Lagged Appropr iate Mind - Related Comments on Overall Expression of N egative Emotion 252 Table I5 Effects of Lagged Negative Commands on Overall Expression of Negative Emotion 254 Table I6 Effects of Lagged Non - Attuned Mind - Related Comments on Overall Expression of Negative Emotion 255 Table I7 Effects of Lagged Initiated Verbal Distraction on Overall Expression of Ne gative Emotion 256 Table I8 Effects of Lagged Toddler Verbal Keys Bids to Mom on Overall Expression of Negative Emotion 258 Table I9 Effects of Lag ged Toddler Verbal Desire Bids to Mom on Overall Expression of Negative Emotion 259 Table I10 Effects of Lagged Toddler Physical Comfort Seeking on Overall Expression of Negative Emotion 260 Table I11 Effects of Lagged Maternal Verbal Comfort on Overall Touch 262 Table I12 Effects of Lagged Maternal Joined Distraction on Overall Touch 263 Table I13 Effects of Lagged Maternal Ve rbal Orientation to Delay on Overall Touch 264 xv Table I14 Effects of Lagged Appropriate Mind - Related Comments on Overall Touch 265 Table I15 Effects of Lagged Punitive Reactions on Overall Touch 266 Table I16 Effects of Lagged Minimizing on Overall Touch 267 Table I17 Effects of Lagged Maternal Non - Attuned Mind - Related Comments on Overall Touch 269 Table I18 Effects of Lagged Toddler Initiated Ve rbal Desire on Overall Touch 270 Table I19 Effects of Lagged Toddler Joined Verbal Distraction on Overall Touch 271 Table I20 Effects of Lagged Toddler Joined Non - Verbal Distraction on Overall Touch 272 Table I21 Effects of Lagged To ddler Verbal Keys Bids to Mom on Overall Touch 274 Table I22 Effects of Lagged Toddler Verbal Self - Control Bids to Mom on Overall Touch 275 Table I23 Effects of Lagged Toddler Verbal Desire Bids to Mom on Overall Touch 276 xvi LIST OF FIGURES Figure D1 . Linear and Quadratic fit for Overall Expression of Negative Emotion and Overall Touch Across 12 intervals 223 1 CHAPTER 1: Introduction The purpose of this introduction is to presen t the study focus, identify the theoretical contexts of the study, and articulate key study concepts. Study concepts as they are housed in the current literature are more thoroughly addressed in the literature review in Chapter 2. Formal research questions and hypotheses are presented in the current study description at the conclusion of Chapter 2. The Importance of Self - Self - regulation is a multidimensional system of cognitive, emotional, and behavioral control that unde rgoes rapid shifts during early periods of development, has profound implications for developmental psychopathology, and is significantly affected by early environmental factors including cumulative risk and quality of interactions with primary caregivers. growing capacity for self - regulation. For example, self - regulation promotes social, emotional, and behavioral competence in toddlers (Spinrad et al., 2007b), school readiness in preschoolers (Blair, 2002), academic success in school aged children (McClelland & Cameron, 2012), and positive psychosocial adjustment in youth from low - income families (Buckner, Mezzacappa, & Beardslee, 2009). Moreover the acquisition of self - regulat ory skills is a defining characteristic of resilient youth living in poverty (Buckner, Mezzacappa, & Beardslee, 2003). Self - regulation moderates the relationship between life stressors and social - emotional and behavioral outcomes (Lengua & Long, 2002; Leng ua & Sandler, 1996), and, thus, functions as a protective factor in the context of cumulative risk (Lengua, 2002). One explanation for the protective effects of self - regulation may lie in the relationship between self - ubsequent resiliency (Lengua & Long, 2002). 2 For instance, children with greater ability to focus attention, inhibit automatic or dominant responses, and lower impulsivity are more likely to utilize positive internal coping strategies, including cognitive r eappraisal and problem solving, to manage stressful life events; in turn, they demonstrate fewer internalizing and externalizing problem behaviors (Lengua & Long, 2002). Coping responses to stressful circumstances reflect behavioral mechanisms related to d ifferences in regulation of attentional, emotional, and control systems (Eisenberg, Valiente, & Sulik, 2009), making the examination of contributing factors to these differences particularly relevant to the study of social - emotional development. On the ot her hand, d isruptions in physiological and behavioral regulation place toddlers at risk for aggressive and destructive behavior problems (Calkins & Dedmon, 200 0 ). Children with regulatory deficits are more prone to anger, impulsivity, anxiety, social withd rawal and depression (Eisenberg et al., 2001) Additionally, children from at risk populations (e.g., poverty, maternal psychosocial problems) are more likely to experience disruptions in normative regulatory development (Blair & Raver, 2012) , and more like ly to experience negative parenting practices that contribute to regulatory competence. Thus, identification of contributing mechanisms which promote or disrupt self - regulation has significant implications for long - term mental health. Origins of Individual Differences in Self - Regulation: The Caregiving Context Individual differences in self - regulation point to both intrinsic (e.g., temperament) and extrinsic mechanisms (e.g., parental support of emerging regulatory skills) that contribute to its developmen tal trajectory (Posner & Rothbart, 2000; Fox & Calkins, 2003). In support of this notion, recent biological perspectives in self - control, an early form of self - regulation, suggest psychophysiological and neural differences between internal mechanisms of in tentional self - control and externally facilita ted self - control (Schel, Scheres , & Crone, 2014). Similarly, 3 Swingler, Perry, and Calkins (2015) highlight the joint role of intrinsic factors such as biological mechanisms, specifically neural plasticity, and extrinsic factors, specifically the caregiving environment, on the development of attention, attentional control, and emotion regulation systems in the first year of life and subsequent development. This study focuses on the caregiving environment, specif - regulation in a low - income population. This focus is informed by several theoretical frameworks including (1) developmental models of self - regulation as a relational construct and its continuous s usceptibility to socialization (Cole, Martin, & Dennis, 2004; Cole, 2014; Eisenberg, Cumberland, & Spinrad, 1998; Morris, Silk, Steinberg, Myers, & Robinson, 2007), (2) attachment theory which designates the parent - child relationship as an organizing and r egulating construct (Bowlby, 1969; Cassidy, 1994; Schore & Schore, 2008), and (3) developmental models which highlight the role of maternal social cognition (e.g. tendency to treat the young child as a psychological agent) in l development (Ainsworth, Blehar, Waters, & Wall, 1978; Fonagy, 2006; Gottman, Katz, & Hooven, 1996; Koren - Karie, Oppenheim, Dolev, Sher & Etzion - Carasso, 2002; Meins, Fernyhough, Fradley, & Tuckey, 2001; Sharp & Fonagy, 2008 ). Understandings of self - regu lation as a relational construct are informed by Eisenberg et - related parenting regulat ory competence. Emotion - related parenting practices refer to a broad range of parenting behaviors (e.g., punitive reactions to negative emotions, cognitive reframing of a stressful tive emotions (Brophy - Herb, Stansbury, Bockneck, & Horodynski, 2012; Morris, Silk, Morris, Steinberg, 4 Aucoin, & Keyes, 2011; Morris et al., 2007). The socializing role of the family context, in gative emotions, have been conceptualized as influential mechanisms in the development of self - regulation and its subcomponents (Eisenberg et al., 1998; Morris et al., 2007), and continue to be a point of emphasis for current investigations of self - regulat ion (Bridgett, Burt, Edwards, & Deater - Deckard, 2015). Within an attachment framework, a central component of the caregiving environment is the parent - child relationship. The characteristics of the parent - child relationship continually s opportunities to practice and build self - regulation skills. For example, the parent - emotional states may be mutually communicated and regulated (Tronick, 1989). For infants and toddlers, who mostly depend on their external caregiving environment for regulation of internal states of arousal (Schore & Schore, 2008), interactions with primary caregivers serve a regulatory function in modulation and organization of activ ated emotional states (Cole et al., 2004; Schore & Schore, 2008). Additionally, social - cognitive theories of parenting, suggest that parent - child relationships of optimal quality (e.g. secure attachments) are characterized by maternal tendencies to accura and accepting of these states( Fonagy, Steele, Steele, Moran et al., 1991; Gottman et al., 1996; Oppenheim & Koren - Karie, 2002; Slade, Grienenberger, Bernbach, Levy, & Locker, 2005 ). A dominant social - cognitive parenting construct is mentalization, a form of social cognition that underlying mental states including emotions, cognitions, be liefs, and intentions (Fonagy et al., 5 1991; Fonagy, 2006). Parental mentalization has been suggested as the mechanism through which the parent - directs the organization and re emotionally relevant states (Fonagy, Gergley, Jurist, & Target, 2002). The parent - child relationship contributes to attentional regulation as well as the capacity to understand and interpret mental states in to understand themselves and others in terms of mental states facilitate self - re gulation through p.187). Within this framework, self - organi zation including affect regulation, impulse control, self - monitoring, and self - secure attachment relationship (Fonagy & Target, 1997; Fonagy et al., 2002). Defining Self - Regulation The c onstruct of self - regulation has been described as multidimensional with differentiated subcomponents, which overlap and become an integrated process as development unfolds (Bell & Deater - Deckard, 2007; Bridgett et al., 2015; Cole et al., 2011; McClelland a nd Cameron, 2012). Despite current lack of consensus on the exact definition of self - regulation (Burman, Green, & Shanker, 2015), emotion regulation and delay of gratification, the foci of this study, are construed as two constructs that reflect the emotio nal and behavioral dimensions of self - regulation (Kopp, 1982). Both emotion regulation and delay of gratification are relevant developmental shifts in toddler years ( Rai kes, Robinson, Bradley, Raikes, & Ayoub, 2007) . 6 Specifically, for toddlers from low - income families, self - regulation, including emotion regulation and delay of gratification, partly develop in the context of the parent - child relationship during daily inter actions in a variety of routine and play contexts (Brophy - Herb et al., 2012). Next, I turn to effortful control as an underlying construct for both emotion regulation and delay of gratification. Although effortful control is not a measured construct in this study, it is discussed here to conceptualize the overlapping concepts of emotion regulation and delay of gratification as jointly espoused by the underlying construct of effortful control. This provides theoretical and empirical support for the ration ale that maternal behaviors contribute to differences in emotion regulation and delay of gratification by influencing underlying mechanisms (e.g., effortful control) shared by emotion regulation and delay of gratification. Both emotion regulation (Eisenber g & Spinrad, 2004; Eisenberg, Spinrad, & Eggum, 2010a) and delay of gratification (Spinrad, Eisenberg, & Gaertner, 2007a) involve effortful control. Effortful control represents a voluntary (vs. reactive) form of control, which involves the ability to shif t and sustain attention, inhibit dominant responses, and activate subdominant responses (Rothbart & Bates, 2006). Individual differences in effortful control emerge toward the end of the first year of life while rapid developmental shifts occur in the seco nd and third years (Eisenberg et al., 2010a; Spinrad et al., 2007a). In terms of emotion regulation, effortful control is suggested to facilitate production of strategies that help reduce negative affect or overarousal such as attention shifting (for revie w see Eisenberg et al., 2010a). Delay of gratification is one behavioral manifestation of effortful control (Spinrad et al., 2007a). Additionally, effortful control has been found to play a mediating role in the relationship between maternal emotion - relate d socialization - emotional functioning (e.g., externalizing behaviors, separation distress, inhibition to novelty, and social competence) concurrently (18 months) and 7 longitudinally (a year later; Spinrad et al., 2007b). Emot ion regulation and delay of gratification are each discussed next. Emotion Regulation. emotions using strategies to initiate, maintain, evaluate, and modify emotional intensity and express ion in a goal - oriented and socially adaptive manner (Cole et al., 2004; Thompson, 1994). implies that emotion is being regulated via the behavioral strategies an individual employs. For emotion ) decreases (frustration is regulated ) as a function of seeking comfort from a caregiver ( regulat ory strategy ). Thus, emotion is regulated via the strategy for coping with the negative emotion. Within heuristic models for emotion regulation (Eisenberg & Morris, 2002; Eisenberg, Morris, & Spinrad, 2005; Eisenberg & Spinrad, 2004) emphasis has been plac ed on theoretical and methodological distinctions between regulation of internal processes (e.g., modulation of internal feeling states, attention shifting) and behavioral concomitants of emotion (e.g., voluntary inhibition or activation of behavior linked to emotion and overt expression of emotion). Additionally, emotion is increasingly understood as a pro cess that is socially regulated with intensive (e.g., intensity of expression) and temporal features (e.g., relationship between strategy use and latency , duration, frequency, and intensity of expression) that reflect regulatory processes (Cole, 2014). To avoid making faulty inferences regarding emotion regulation (e.g., take low levels of emotion expression as evidence for emotion regulation; Eisenberg, C hampion, & Ma, 2004), recent advancements in conceptualization of emotion regulation emphasize that emotion regulation strategies be measured independently from activated emotions (e.g., emotion expression) and emphasize analysis of temporal relations 8 betw een such strategies and their corresponding activated emotions (Cole et al., 2004; 2011). Cole et al. (2004) suggest that emotion regulation should be measured in such a way that and that regulatory operationalized as a combination of toddler - initiated regulatory strategies (predictor variable) ression of negative emotion (outcome variable). intensity of expression of negative emotion provides methodological advancement over studies in which tempora l distinction between regulatory strategies (maternal attempts or toddler y strategies study of emotion regulation (Thompson, 1996; Cole, 2014; Cole et al., 2004; Cole et al., 2011). Delay of Gratification. In addition to emotion r egulation, the focus of this study is to examine behavioral self - will hereafter be referred to as delay of gratification. The ability to wait is a behavioral manifestation of self - re gulation as it reflects the integration of emotional, attentional, and behavioral control (Cole et al., 2011, Bell & Deater - Deckard, 2007). Behavioral self - regulation involves interrelated processes including self - control, effortful control, and executive functioning (Bridgett et al., 2015). Behavioral manifestations of self - control, emerge in the second year of life, and involve the ability to monitor behavior, comply with demands of caregiver, and delay gratification upon request (Kopp, 1982), which repre sent developmentally salient, albeit 9 Self - Regulation during Toddler Years Toddlerhood represents a key period for the examination of individual differences in early self - regulation. Toddlers are often expected to delay gratification and tolerate frustration in the course of daily family life. As language abilities increase in th e second and third years of life, toddlers gain more regulatory competence (Cole, Armstrong, & Pemberton, 2010; Roben, Cole, & Armstrong, 2013; Vallotton & Ayoub, 2011) and learn to shift from external sources of regulation to internal sources (Kopp, 1982; Raikes et al., 2007). - regulation skills become salient to the everyday experiences of toddlers. Perhaps accordingly, and 30 months, while their use of distraction and soothing decrease between 18 and 30 mont hs (Spinrad, Stifter, Donelan - McCall, & Turner, 2004). Typically, toddlers use a range of strategies that rely on the self (independent strategies) or the caregiver (dependent strategies) to regulate negative emotions and delay gratification (Calkins & Joh nson, 1998; Feldman, Dollberg, & Nadam, 2011; Grolnick, Bridges, & Connell, 1996; Gilliom et al., 2002; Spinrad et al., 2004). Independent strategies typically include distraction and self - comforting behaviors, whereas, dependent strategies include physica l and verbal bids to caregivers (Calkins & Johnson, 1998; Gilliom et al., 2002). Distraction involves object such as looking away or exploring other objects in the room. Self - comforting includes 10 verbal self - - soothing behaviors such as thumb sucking or other forms of self - touch (e.g., self - hug) (Feldman et al., 2011; Grolnick et al., 1996). Bids to caregiver s refer to toddler initiated behaviors that attempt to involve the mother in regulation such as physical comfort seeking or verbalizations directed toward or in response to object/task (Ekas, Braungart - Rieker, Lickenbrock, Zentall, & Maxwell, 2011; Gilliom et al., 2002). - Regulation negative emotions m been studied in the context of maternal availability by manipulating levels of materna l involvement during frustrating tasks (Diener & Mangelsdorf, 1999). Generally, mothers tend to affect (Spinrad et al., 2004). Thus, frustrating tasks which induce provide a rich context for the examination of effects of maternal regulatory strategies on regulation and delay of gratification in childhood highlights the role of parenting behaviors in Corapci, & Benga, 2015a), hereafter called Maternal Regulatory Attempts (MRAs), and, - 11 M indedness (MM; Meins et al., 2001) have recently emerged as two parenting constructs gaining attention for their links to early social - emotional development. As will be described next, MRAs refer to specific strategies parents use with the intention to pro behaviors during emotion eliciting experiences, such as during a frustrating task. Mind - own internal mental states, which c 2001; Meins & Fernyhough, 2015). Given the prevailing focus on mothers in the literature and Alt hough I recognize that fathers are important socializing agents, the socializing role of fathers is beyond the scope of this study and is being pursued elsewhere. Purpose and Rationale The general aim of this study is to examine how the relationship share d by toddlers and - regulation using a multilevel modeling approach. - re gulation. Specifically, maternal contributions as defined by the effects of Maternal Regulatory Attempts, maternal mind - mindedness, and overall intensity and predo minance) , and (2) ability to wait upon request or delay of gratification (measured as sum of touch and attempted touch of the attractive keys), during a 2 - minute frustrating task. As stated earlier, emotion regulation refers to changes in intensity, durat ion, and frequency of emotional expression as a function of a temporally contingent strategy (Cole et al., 2004, Cole, 2014; Thompson, 1994). In the current study, emotion regulation was 12 operationalized as within - individual differences 1 in intensity and pr edominance in expression of negative emotion as a function of a previous - interval strategy. Behavioral regulation was defined as self - 1982). In the current study, b ehavioral regulation was operationalized as within - individual differences in total number of times toddlers refrained from touching or attempting to touch a desirable object as a function of a previous - interval strategy. A very limited number of studies h ave used multilevel modeling to examine self - regulation using longitudinal data. Cole et al. (2011) investigated age - related progress (18 - 48 months) in ability to wait as operationalized by latency and duration of anger expression and strategies (e.g., attention focus and bids to caregiver) during an 8 - min gift delay task. However, Cole et al., (2011) did not examine temporal contingency between regulatory strategies and emotional expression (e.g., anger expression), maternal regulator y strategies or maternal mind - mindedness. In another study, Morris et al. (2011) examined children (preschool - 2 nd grade) in an observational task while nesti ng repeated measurement occasions (a form of longitudinal data structure) within mother - child dyads. Similarly, the current study nested repeated measurement occasions (12 units of 10s intervals) within mother - child dyads, and examined effects of maternal mind - strategies in addition to effects of Maternal Regulatory Attempts as level - one predictors. regulation strategy and emotional ex pression (e.g., effects of strategy from previous interval on outcome in current interval), in 24 - month old toddlers from low - income families, a population 1 measured over 10s repeated measurement occasions 13 that was not examined in previous studies but is a sensitive developmental period for emergence of s elf - regulation. Examining the occurrence of Maternal Regulatory Attempts, Appropriate and non - attuned mind - negative emotion (average intensity and predominance) and toddle (delay of gratification/ability to wait), separately and over repeated measurement occasions (i.e., intervals in this study), allowed me to (1) demonstrate within - individual differences (changes) in on of negative emotion over time (2) demonstrate within individual - differences in toddlers via temporally contingent maternal regulatory attempted, maternal mind - minde considering temporal associations can provide evidence for effectiveness of a regul atory strategy on overall expression of negative emotion, and, thus partly explain change in intensity and predominance of expressed emotion as a function of a regulatory strategy. Examination of the trategies and overall touch, considering temporal associations can provide evidence for effectiveness of regulatory strategy on delay of gratification, and, thus partly explain change in ability to wait as a function of a regulatory strategy. 14 CHAPTER 2: Literature Review The following sections in this chap ter clarify (a) definitions of maternal regulatory a ttempt s; (b) unique contributions of maternal regulatory a regulation and delay of gratification; (c) definitions of mind - mindedness; (d) unique contributions of mind - fication, (e) contributions of maternal regulatory a ttempts and mind - mindedness (f) unique contributions of toddler - initiated regulatory strategies t The chapter concludes with study research questions and hypotheses. Maternal Regulatory Attempts (MRAs) Mothers provide scaffolding for emotional development in multiple contexts of daily life in cluding playtime, daily routines, and social interactions (Brophy - Herb et al., 2012). MRAs refer to parental behaviors, or parent - initiated regulatory strategies, that attempt to regulate ors (Friedlmeier, Corapci, & Benga, 2013; Calkins & Johnson, 1998; Spinrad et al., 2004; Putnam, Spritz, Stifter, 2002). MRAs are strategies for facilitating regulation of child distress induced, for example, by request to wait for an attractive toy, and m s al., 2004, Spinrad et al., 2007b). While these maternal behaviors are traditionally assessed in observational paradigms (Calkins & Johnson, 1998), others have examined MRAs by investigating self - sadness, fear and anger (Eisenberg et al., 2010b; Davidov & Grusec, 2006, Spinrad et al., 15 2007b). These studies suggest that maternal respon negative emotions include differences in the use of supportive strategies such as encouragement of emotional expression and problem solving and non - supportive strategies including minimizing, ignoring, or punishment o negative affect or in the context of frustrating tasks including toy removal, still - face paradigm, and delay of grat ification tasks (Feldman et al., 2011; Putnam et al., 2002). Unique Contributions of MRA to Toddler Emotion Regulation and Delay of Gratification A number of studies have established the link between specific MRAs and differences in regulatory capacities in toddlers. Generally, physically and verbally warm and comforting strategies are related to positive development of self - regulation. For instance, Spinrad et al. (2004) found that maternal supportive strategies, specifically soothing and acceptance of th eir 18 - month - - initiated or independent regulation strategies at age 5, and, in particular, soothing and acceptance was - distraction as an emotion regulatory strategy in response to disappointment. Additionally, higher maternal warmth is associated with higher behavioral regulation in 3 year olds in at risk families (Eiden, Edwards, & Leonard, 2007), and predicts greater growth in sustained attention from 2 to 4.5 years (Graziano, Calkins, and Keane, 2011).When mothers use positive feedback and guidance (e.g., physical affection, warmth when reminding child not to touch delay object in gentle voice) during delay of gratification tasks, their toddlers tend to use self - distraction as a regulatory strategy in delay of gratification tasks (Calkins & Johnson, 1998; Feldman et al., 2011). 16 Findings with regard to maternal use of distraction as an effective strategy in regulating distress during frus tration have been mixed (Spinrad et al., 2004). While some studies have expression (Spinrad et al., 2004), others have found positive relationship between maternal use Rivkin, & Bridges, 1998). Grolnick et al., (1998) found that when mothers use distraction while actively engaging with their toddlers, their toddlers display higher l evels of distress when they are required to regulate independently (e.g. parent - passive paradigms) which can suggest a temporary effectiveness of distraction Feldman et al., (2011) argue that although di straction may serve a regulatory function in context maternal use of distraction may be effective for anger regulation in older children (Morris et al., 2011) or delay of gratification in toddlers (Putnam et al., 2002). Morris et al., (2011) found that maternal use of distraction was effective in reducing intensity of anger expression in a cross - sectional sample (6 - 9 year olds). Putnam et al., (2002) found tha t higher maternal use of distraction was associated with higher ability to delay gratification in 30 month olds and a greater ability to allocate attention on others instead of a desirable toy . The bidirectional nature of maternal use of distraction and to Johnson, 1998) may play a role in these inconsistent findings. It may also be the case that maternal use of distraction is more of an adaptive strategy with older children compared to younger children systems of control (Cole et al., 2011). For instance, Cole et al., found that compared to 2 year - olds, 3 - - initiated distractions for longer perio ds of time was 17 related to shorter duration of anger expression. Together, these findings suggest mixed effects for maternal use of distraction and delay of grat ification. and punishment of negative emotions, have been associated with emotional dysregulation, including externalizing problem behaviors concurrently and longitudin ally at 18, 30, and 42 their negative emotions may lead to over arousal of negative emotion and therefore compromise their ability to develop and employ effective regulatory strategies (Spinrad et al., 2007b). The use of minimizing emotional expressions in toddlerhood may have negative implications for developing self - regulatory skills as development continues. For instance, children who do not use (or do not have access to) behavioral strategies to regulate disappointment at age 5 tend to have mothers who questioned their negative affect at 18 months, suggesting that the mechanism for development of poor regulatory capacities (Spinrad et al., 2004). Maternal harsh control (e.g., use of harsh and angry discipline, scolding, insulting or hitting) is related to higher anger expression in toddlers (Feldman et al., 2011). Additionally, E isenberg and colleagues (2010b) found that greater maternal unsupportive responses (i.e., minimization and punishment attention shifting, attention focusing , inhibitory control, and observed ability to delay gratification) at 30 months of age. In sum, MRAs refer to maternal behaviors directed towards the child that contribute to emotion regulation and delay of gratification. Although maternal verbal support such as 18 included under Maternal Regulatory Attempts (Spinrad et al., 2004), they do not necessarily nterpreted as motives underlying mind - mindedness. Mind - Mindedness (MM) Mind - Minded tendency to consider emotions, cognitions, and intentions as mental states that motivate behaviors in self and others (Meins, Fernyhough, & Harris - Waller, 2014). In order to accurately p erceive mental states in others as underlying motives of behaviors, one must first recognize others as separate psychological agents with intentionality. In parents, this tendency is indicative of a range of mentalization capacities including mind - mindedne ss (Sharp & Fonagy, 2008). Mind - individual with mental states, which directly motivate his behaviors (Meins et al., 2001; Meins & Fernyhough, 2015). MM has been operationa mind - emers, Bernier, Tarabulsy, & Provost, 2010a) presumed internal states; (2) of non - 98; Meins & Provost, 2010b) and preschooler (Meins, Fernyhough, Russell, & Clark - Carter, 1998) using mental attributes (e.g., curious, nervous) when asked to des 19 The current study, utilizes the interactional operationalization of MM. Interactional measures of MM operationalize the construct of MM as the maternal tendency to make appropriate or non - attuned mind - related comments that reflect the c Appropriate comments refer to comments which seem to be an accurate interpretation of the with waiting. Non - mental s appropriate mind - ays with giraffe; Meins & Fernyhough, 2015). To date, MM has been studied primarily relative to parental sensitivity and child - centered responses (Meins et al., 2001; Meins, Fernyhough, Arnott, Turner & Leekman, 2011) during parent - child interactions (Lara njo, Bernier, Meins, & Carlson, 2014; Meins et al., 2001). It is important to clarify that appropriate mind - related comments have been studied most often as predictors of early manifestations of theory of mind understanding in 2 - year - olds (Laranjo, Bernier , Meins, & Carlson, 2010), theory of mind performance in 4 - year - olds (Meins, Fernyhough, Wainwright, Gupta, Fradley, & Tuckey 2002), and 6 - and 10 - year - Devine, Marks, & Hughes, 2014). However, MM has recently been associated with fewer behavi oral problems including tendency to experience negative emotions, and hyperactivity in children (44 and 61 months; Meins, Centifanti, Fernyhough, Fishburn, 2013) and toddlers (18 - 36 months) from low - income families (Brophy - Herb et al., 2015) and is the foc us of a few studies 20 - regulatory skills (Bernier, Carlson, & Whipple 2010). - child conversational contexts. Mental state language refers to a specific form of communication that utilizes references to internal states, including cognition (e.g. think/know), emotions (e.g., sad/excited), and desires/goals (e.g. like, want ) in order to facilitate social - emotional understanding, identify internal states, and facilitate meaning - making (Taumoepeau & Ruffman 2006, 2008; Fivush & Baker - Ward, 2005). Maternal mind - as children interact with their environment (e.g. remember how you were crying and felt s ad after grandmother went back home?). However, not all mental state language is manifested as mind - mindedness. 2 utilization of mental state terms to reflect putative, present, and ongoing mental activity of the child accurately - minded caregiver often utilizes mental state language as a tool in order to verbalize her al states. An essential distinction between MM and merely - 2 - related comments that align 21 orientation to non - attuned mind - related comments represent distinct dimensions of mind - mindedness as not all mind - 12). Appropriate mind - maternal sensitivity (Meins et al., 2001) whereas non - attuned mind - related comments have been found to be unrelated to maternal sensitivity (Arnott & Meins, 200 7) and independently predict attachment disorganization in infants (Meins et al., 2012). The majority of research on MM has as theory of mind (Laranjo et al., 201 0; Meins et al, 2002; Ensor et al. 2014), while less focus has been placed on elucidation of the effects of maternal MM on regulatory competence in young empiric al value in its potential contributions to the extant literature and applied significance via its possible implications for parenting/caregiver educational programs. Unique Contributions of Maternal Mind - and Del ay of Gratification Attuned Mind - Related C omments. To our knowledge, the direct effects of mind - mindedness, operationalized as appropriate - attuned, and non - attuned mind - related comments, on toddler expression of negative emotion or delay of gratification have not been examined in toddlers with the exception of one study which examined the effects of mind - mindedness (appropriate mind - - regulation (Bernier et al., 2010). I will discuss this study shortly. There are on ly a handful of studies which have examined associations between maternal mind - 22 instance, exposure to appropriate mind - related comments in infancy is negatively related to dysregulatory behaviors, sp ecifically, internalizing and externalizing problem behaviors at 44 and 61 months (Meins et al., 2013) and conduct problems at age 10 (Centifanti, Meins, & Fernyhough, 2016). Furthermore, appropriate mind - related comments directed toward infants mitigate t he negative impact of low - socioeconomic status on behavior problems in childhood after controlling for maternal sensitivity, depressive symptoms, perceived social support, child language abilities and child gender (Meins et al., 2013). Although these prio r studies make important contributions to the field, Bernier et al. (2010) were the first to study potential mechanisms through which mind - related comments are related to regulatory outcomes. In the only study of mind - - re gulation, they investigated the effects of maternal sensitivity, autonomy support, and mind - mindedness (operationalized as number of appropriate mind - related comments during a 10 - minute free play session) at 12 and 15 months on dimensions of executive func tion including working memory and impulse control at 18 and 26 months (N=80). Compared to toddlers exposed to fewer appropriate mind - related comments, those exposed to more appropriate mind - related comments had better working memory at 18 months and better impulse control at 26 months. Additionally, Bernier and colleagues (2017) have provided evidence for the longitudinal indirect effects of appropriate mind - related la mind - related comments at 18 months predict better inhibitory control in delay tasks for 3 year olds (Gagne, Bernier, & McMahon, 2017). These findings provide emerging evidence that exposure to appropriate mind - related comments are associated with processes underlying emotional and behavioral regulation such as effortful control. 23 Similarly, despite the dearth of research on the effects of maternal mind - mindedne ss on mind - mindedness, particularly maternal differences in use of and references to mental states s, including emotion regulation and delay or gratification. For instance, maternal mental state language, specifically the number of emotion, cognition, and desire words mothers use with toddlers during a book sharing task, has been associated with toddler - income samples (Brophy - Herb et al., 2012). Maternal emotion bridging, a type of emotion talk which aims to interpret behaviors of others in terms of their underlying mental states (e.g., she is sad bec ause her pet is lost), and to make meaningful connections between internal states of toddlers and others (e.g., do you remember you were sad when grandma left?) has been found to moderate the ally for children from higher risk families and with greater earlier behavioral problems (Brophy - Herb et al., 2015). Broadly, these findings suggest that maternal appropriate mind - delay of gratification by facilita tive use of processes underlying behavioral regulation including working memory and impulse control. Additionally, maternal mentalization - related parenting behaviors indicated by use of mental state words, emotion bridging, and representational mind - minded - Herb, & Vallotton, 2018). Although, maternal sensitivity and maternal representations of the parent - child relationship, two constructs closely related to maternal mind - mind edness (Arnott & Meins, (Feldman et al., 2011), the exact nature of the contribution of maternal mind - mindedness to intensity of emotion expression is not known in the existing literature. 24 To date, most explanatory mechanisms relative to the links between these maternal exposure to mental state references, Ruffman 2008), including understanding of mental states in self and others (Symons, Fossum, & Collins, 2006), and advance ca (Taumoepeau and Ruffman 2008), which serves regulatory functions by allowing the child to express and identify emotions and other internal states. Thus, maternal references to mental states, an indicat or associated with mind - mindedness, may promote the development of self - emotions in themselves and others (Eisenberg, Sadovsky, & Spinrad, 2005). Research with pres chool aged children provides similar evidence of such mechanisms. For example, understanding of emotion regulation strategies (which requires emotion understanding) in preschoolers has been associated with preschooler - initiated regulatory strategies (e.g. distraction and support seeking), and behavioral regulation (e.g. persistence) during a frustrating task (Cole, Dennis, Smith - Simon, & Cohen, 2009). Additionally, emotion understanding at 51 months is negatively related to problem behaviors reflective of r egulatory deficits in 10 year olds including externalizing problem behaviors, conduct problems, and impulsivity (Centifanti et al., 2016). However, other possible mechanisms may be the skills that are modeled and promoted by mothers via mind - mindedness. - mindedness as characterized by attuned mind - self - regulation capacity by providing them with social, emotional, and cognitive tools for 25 cognitive reappraisal of emotional arousal which can serve a regulatory function (Cole et al, 2004). For instance, Morris et al., (2011) found that cognitive reappraisal of a frustrating event was related to lower intensity of anger expressions in school - aged children. Theoretically, narration of these motives may be an essential me and understanding of emotional, cognitive, and intentional states (Meins et al., 2001). Therefore, exposure to appropriate mind - related comments during frustrating situations may facilitate early development of toddlers to gain attentional control over internal states during this developmental period when they are transitioning from external - regulation towards self - regulation. The proposed s tudy is the regulatory strategies, mind - emotion expression and delay of gratification. Following a brief discussion of non - attuned mind - related comments, I turn toward further explanation of how both Maternal Regulatory Attempts and mind - Non - Attuned Mind - Related Comments. Contrary t o the promotion of skills associated with attuned mind - related comments, non - attuned mind - related comments may be negatively - attuned mind - effects of non - attuned mind - gratification, i - 26 sition which is than allowing child to do activity on his own) leads to higher intensity of negative emotion expression in toddlers in response to a frustratio n task (Calkins & Johnson, 1998). However, the exact nature of this relationship remains to be empirically validated. Contributions of Maternal Regulatory Attempts and Mind - Mindedness Although there is emerging evidence of the moderating role of social c ontext (e.g., book sharing, mealtimes, conversation about past events) on individual differences in child and maternal use of mental state language in preschoolers (Howe, Rinaldi, Recchia, 2010), maternal mind - mindedness has not been examined during frustr ating tasks with toddlers, leaving a gap in understanding of the effects of maternal reflection of mental states (i.e., mind - mindedness) on Regulatory Attempts. It has been suggested that maternal mind - mindedness may facilitate - regulation (specifically executive function) by promoting cognitive skills (e.g., linguistic tools), and effortful control (Bernier et al., 2017) necessary to shift from reliance o n external - regulation to self - regulation (Bernier et al., 2010; Carlson, 2003). However, empirical evidence for the contribution of mind - of gratification is lacking. Additionally, the degree to which m ind - mindedness, in tandem with other maternal regulatory strategies, facilitates or hinders emotional and behavioral regulation, remains to be understood. Examining the effectiveness of maternal mind - mindedness and maternal regulatory strategies in the sam e study, will inform our current understanding of the nature of the relationship among maternal regulatory strategies and maternal mind - mindedness 27 regulation and delay of gratification. Maternal Regulatory Attempts and mind - mindedness reflect aspects of parenting that mechanisms through which they may promote emotion regulation and delay of gratification. Maternal Regulatory Attempts have been examined in tasks that elicit distress in toddlers and preschoolers. Mind - mindedness, on the other hand, has traditionally been examined in tasks that induce positive or neutral affect in infants. To date, no study has examined the contributions of each approach simultaneously in at risk populations in the context of a frustrating task. Although - regulati on, immediacy of effects and predictive strengths of each (MRA or MM) are not known. The examination of each approach in separate multilevel models will provide evidence for immediacy of effects and predictive strengths of each as a regulatory strategy. U Regulation and Delay of Gratification Developmental changes in anger expression and attention focus underlie the ability to wait and partly occur as a function of maturational pr ocesses (Cole et al., 2011). However, differences in toddler - initiated use of regulatory strategies may explain some variation in to quickly distract themsel ves for long periods of time predicts shorter duration of anger expression in waiting tasks (Cole et al., 2011). In a longitudinal study, Gilliom et al., (2002) ed to decreased anger (emotion regulation) while having to wait to eat a cookie (delay of 28 related to greater sustained attention (Graziano et al., 2011). In a sample of 20 - month old toddlers (N=116), Ekas et al., (2011) found that toddlers who made more bids to their inattentive negative affect compared to toddlers who did not use parent - focused strategies. These findings distress inducing while exposed to unsupportive maternal regulatory attempts such as being ignored. However, D effective in reducing fear and anger across four laboratory episodes designed to elicit fear and anger, while distraction only served to maintain levels of anger intensity and was thu s deemed as a less effective strategy. Broadly, these findings suggest that toddler - initiated distraction and bids to caregiver may be effective in reducing negative emotion and helping toddlers wait but are influenced by levels of maternal involvement. C urrent Study The current study aim ed to examine the effects of M aternal R egulatory A ttempts (MRAs) , m aternal M ind - M indedness (MM) , and T R egulatory S trategies (TSR) on (ENE - O) , and (2) ability to wait upon request or delay of gratification (Overall Touch) , during a 2 - minute delay of gratification task. Conceptual and operational definitions are presented in Table 1 (maternal variables of interest) and Table 2 (toddler variables of interest) . The f rustrating task used in the current study (i.e. touch an attractive set of toy keys upon maternal request. All study variables were coded observationally usin g the delay of gratification task. The duration of the delay of gratification 29 task was 120 seconds. Coding of independent and dependent study variables were conducted along 12 intervals. Each interval was defined as a 10s unit of time. Twelve 10s - repeated measurement occasions were nested within 134 mother - toddler dyads. This was the first study to simultaneously examine MRAs, MM, and TSR (predictors) as they were ENE - O and Overall Touch (outcomes) . Thus, hypotheses were only proposed for the relationship of each pred ictor to each outcome variable and analyzed in separate two - level models with random intercepts. T he overarching goal of the study was to identify temporal associations between predictors in previous interval s (lagged 10s) ENE - O and Overall Touch in the current interval. Thus, my focus was not on which maternal behavior or toddler behavior was the most effective predictor of ENE - O or Overall Touch , in presence or interaction of other predictors, but whether each maternal behavior and toddler behavior predicted ENE - O or Overall Touch. Therefore, hypotheses about the effectiveness of strategies in comparison to each other or as they interact ed with each other were not made in this initial study. I do expect to exa mine such questions in subsequent studies. 30 Table 1 Conceptual and Operational Definitions of Maternal Variables Overarching Theoretical/Concept ual or Developmental Framework Conceptual Definitions Operational Definitions Independent Variables Mater nal Regulatory Attempts Emotion - Related Socialization Behaviors (Eisenberg et al., 1998, Morris et al., 2007) Specific strategies parents (mothers in the current study) demonstrate to promote young - eliciti ng experiences, such as during a frustrating task. These strategies include verbal and physical comfort of the child, distraction, orientation to delay, commands, punitive reactions in response physical rest emotional reactions to the child. 1. Sum of verbal comfort 2. Sum of physical comfort 3. Sum of initiated distraction 4. Sum of joined distraction 5. Sum of verbal orientation to delay 6. Sum of positive commands 7. Sum of negative command s 8. Sum of punitive reactions 9. Sum of minimizing 10. High/low physical restraint 11. Presence/absence of positive emotional reactions 12. Presence/absence of negative emotional reactions 1. Verbal Comfort 2. Physical Comfort 3. Initiated Distraction 4. Joined Distraction 5. Verbal Orientation to Delay 6. Positive Commands 7. Negative Commands 8. Punitive Reactions 9. Minimizing 10. Physical Restraint 11. Positive Emotional Reactions 12. Negative Emotional Reactions Mind - Mindedness Attachment Theory (Meins, et al., 2 001); Mentalization, Affect Regulation, & Self - Organization (Fonagy et al., 2002) Tendency to treat child as a separate individual with psychological states that govern his behavior. Mind - minded comments may be appropriate (e.g., accurate and aligned with the - attuned (inaccurate 1. Sum of appropriate mind - related comments 2. Sum of non - attuned mind - related comments 1. Appropriate Mind - Related Comments (1) 2. Non - Attuned Mind - Related Comments (2) for each variable is given in Tables 3 and 4. 31 Table 2 Conceptual and Operational Definitions of To ddler Variables Overarching Theoretical/Conceptual or Developmental Framework Conceptual Definition Operational Definitions Independent Variables Dependent Variables Emotion Regulation Developmental frameworks for operationalization of emotion - regulat ion (Cole, Martin, & Dennis, 2004; Eisenberg and Spinrad, 2004; Thompson, 1994) Emotion regulation refers to activated emotions using strategies to initiate, maintain, evaluate and modify emotional intensity and expression in a goal - oriented and socially adaptive manner. Emotion regulation includes verbal and nonverbal behaviors. Some strategies may not involve the parent (independent of the parent) and some strategies involve the parent (termed dependent). Sum of toddler init iated: 1. verbal distraction 2. non - verbal distraction 3. verbal - keys 4. verbal - self - control 5. verbal - desire 6. self - comfort Sum of toddler joined: 7. verbal distraction - bids 8. verbal distraction - joined 9. non - verbal distraction joined 10. verbal - keys - bi ds 11. verbal - self - control - bids 12. verbal - desire - bids 13. comfort seeking 1. Toddler IRS (sum of 1 - 6) 2. Toddler DRS (sum of 7 - 13) Intensity of Expression of Negative Emotion 0 = no cue for emotion 1 = slight 2 = moderate 3 = strong intensity Delay of Gratification Developmental frameworks for operationalization of behavioral self - regulation (Kopp, 1982) Ability to wait to touch the keys upon request. 1. sum of discreet events of touching the keys 2. sum of discreet attempts to touch the keys - Ove rall Touch (sum of Touch and Attempted Touch) for each variable is given in Tables 5 and 6. IRS = Independent Regulatory Strategies, DRS = Dependent Regulatory Strategies. 32 Covariates. strategies, emotion regulation, and delay of gratification is limited. However, a few studies point to gender differences related to toddler - initiated regulatory strategies. Two - year - old girls display more caregiver - focused regulation strategies, self - soothing, and higher distress when observed in low - threat novelty episode (e.g., interaction with a female clown or puppet sho w) (Premo and Kiel, 2014). In one study by Premo and Kiel (2014), child gender moderated the association between toddler initiated regulation strategy and m aternal r egulatory a ttempts such that 2 - year old boys who sought more caregiver focused regulations strategies (e.g., seeking caregiver contact) during mildly frustrating laboratory experiments (e.g., interacting with a clown), received less non - supportive regulatory attempts (e.g. minimization, punitive reactions) at age 3. Smith, Calkins, and Keane (20 06) found that maternal controlling behaviors including threats and criticisms of their 2 - relationship was not significant for boys. These findings point to gender differences in mat ernal emotion socialization practices which broadly suggest supportive respon and fear, and thi rd years of life and facilitate emergence of self - regulatory faculties (Cole et al., 2010) including effortful control (Bernier et al., 2017). Additionally, gender differences in expressive - regulation (Val lotton & Ayoub, 2011). Thus, intercepts in the current study and will be pursued as moderators for slope differences (e.g., effects of MRAs on outcomes) in future ana lyses. 33 Overview of Research Questions. 3 An overview of research questions and accompanying hypotheses is provided below. Questions are operationalized within multilevel modeling framework under Plan of Analyses in Chapter3: Method. Questions are presente d in order of Question 1A and 1B. Do Maternal Regulatory Attempts from the pre vious interval Hypothesi s 1A. Maternal Regulatory Attempts from previous interval including lagged Verbal Comfort, lagged Physical Comfort, lagged Initiated Distract ion, lagged Joined Distraction, lagged Verbal Orientation to Delay, lagged Positive Commands, lagged Positive Emotional Reactions, and lagged Appropriate Mind - Related Comments will each be related to low Overall Expression of Negative Emotion in current in terval. Hypothes i s 1B. Maternal Regulatory Attempts from previous interval including lagged Negative Commands, lagged Punitive Reactions, lagged Minimizing, lagged Physical Restraint, lagged Negative Emotional Reaction, and lagged Non - Attuned Mind - Relate d Comments will each be related to high Overall Expression of Negative Emotion in current interval. Question2A. the current interval? Hypothesi s 2A. including lagged Initiated Verbal Distraction, lagged Initiated Non - Verbal Distraction, lagged Initiated Verbal Keys, lagged Initiated Verbal Self - Cont rol, lagged Initiated Verbal Desire, and 3 Se e Supplementary Summaries A and Table 20 on page 80. Table 20 is also provided on page 2 of Supplemental Materials. All supplementary materials are included in a supplemental file in the electronic ProQuest account. 34 lagged Self - Comfort will each be related to low Overall Expression of Negative Emotion in current interval. Question 2B. (lagged) predict tod Hypothesi s 2B. including lagged Verbal Distraction - Bids to Mom, lagged Joined Verbal Distraction, Toddler Joined Non - V erbal Distraction, lagged Verbal Keys Bids to Mom, lagged Verbal Self - Control - Bids to Mom, lagged Verbal Desire Bids to Mom, and lagged Physical Comfort Seeking will each be related to low Overall Expression of Negative Emotion in current interval. Question 3A and 3B. Do Maternal Regulatory Attempts from the previous interval Hypotheses 3A. Maternal Regulatory Attempts from previous interval including lagged Verbal Comfort, lagged Phys ical Comfort, lagged Initiated Distraction, lagged Joined Distraction, lagged Verbal Orientation to Delay, lagged Positive Commands, lagged Positive Emotional Reactions, and lagged Appropriate Mind - Related Comments will each be related to low Overall Touch in current interval. Hypothesi s 3B. Maternal Regulatory Attempts from previous interval including lagged Negative Commands, lagged Punitive Reactions, lagged Minimizing, lagged Physical Restraint, lagged Negative Emotional Reaction, and lagged Non - Attun ed Mind - Related Comments will each be related to high Overall Touch in current interval. Question 4A. 35 Hypothesi s 4A. To including lagged Initiated Verbal Distraction, lagged Initiated Non - Verbal Distraction, lagged Initiated Verbal Keys, lagged Initiated Verbal Self - Control, lagged Initiated Verbal Desire, and lagged Self - Comfort will each be related to low Overall Touch in current interval. Question 4B. Hypothesi s 4B. Tod including lagged Verbal Distraction - Bids to Mom, lagged Joined Verbal Distraction, Toddler Joined Non - Verbal Distraction, lagged Verbal Keys Bids to Mom, lagged Verbal Self - Control - Bids to Mo m, lagged Verbal Desire Bids to Mom, and lagged Physical Comfort Seeking will each be related to low Overall Touch in current interval. 36 CHAPTER 3: Method Participants The data used in the current study were collected as part of the Early Head Start Research and Evaluation P roject (Love et al., 2005). A total of 3,001 low - income families were recruited nationally across 17 research sites. The current sample is comprised of 134 toddlers (67 girls; M age = 25.77 months, SD age = 1.60) with available video data from one Midwestern site. There were 2 cases with fathers that were not included in the current sample. Thus, all parental 4 variables are maternal variables. Families were randomly assigned to receive Early Head Start (EHS) services (N=69) or to a c omparison group (N = 65). Families in the comparison group were free to access non - EHS programs. Data were collected in the home at study enrollment, at - months, 24 months, 36 months, 60 months birth - related assessments, and when children were 10 years old. The current study utilizes the data collected for Michigan research site with available videotaped interviews at the 24 - months assessment. About 68.7% (n= 92) of the sample was Caucasian, 14.9 % (n=20) African - American, 3.0% (n=4 ) Hispanic, and 3.7% . On average, mothers were 22.97 (SD = 4.92) years old at enrollment. About 42.9 % (n=51) mothers had less than high school education, 33.6 % (n= 40) mothers had a high school diploma, and 23.5 % (n=28) mothers had a col lege education. Cumulative risk was calculated as a sum score (Love et al., 2002) indicated by the presence or absence of five risk factors (single parenting, welfare receipt, teen pregnancy, high - school dropout, and unemployment) and ranged from 0.00 - 5.00 , with mean of 2.81 ( SD = 1.20). 4 37 Procedure Delay of Gratification Task. MRAs, MM, TRS, ENE - O and Overall Touch were observed during the D elay of G ratification T ask assessed during the 24 - month parent - child start of the task, the camera was placed 4 feet in front of the parent - child dyad with both faces visible on the tape. The interviewer said to the toy anim al (e.g., Pooh Bear, turtle). I would like you to keep (child) from touching it for 2 minutes. We are doing this because we are interested in learning about a situation that happens to toddlers a lot, which is that they have to wait for something they want right away. For this part, The interviewer asked the parent to sit on mat with child. Once the parent and child were sitting on mat, the interviewer shook the keys several times until they perceived the child to be fully attentive to keys. The set of keys had a small toy (e.g., Pooh bear, turtle, tiger) attached to them to make them more attractive. The interviewer placed the keys only 2 feet from the child and said, making any eye contact with the child or interact with the parent at any tim e point during the task. Study protocol indicated that the task be stopped if the child cried hard for more than 30 seconds or if the parent restrained the child for more than 1 minute. Restraining was defined as the mother holding back the child as the ch ild was actively trying to move toward the keys. In the current study we also coded maternal physical restraint as a regulatory attempt. 38 Measures 5 Maternal Regulatory Attempts. Maternal Regulatory Attempts were coded using the Coding System for Delay of Gratification Task - Friedlmeier, The coding system was adapted for the current study and included 12 types of attempts described belo w (Table 3). This adaptation is based on previous findings that have found most variation in gratification (Calkins & Johnson, 1998; Morris, et al. 2011, Putnam et al., 2002; Spinrad et al., 2004). Maternal behaviors coded in the current study included: (1) Verbal Comfort (V - Comf) , eassure the child that he/she will soon get the keys (e.g., Physical Comfort (Phy - Comf) , behaviors that provide physical comfort for t he child (e.g., hugging), (3) Maternal Initiated Distraction (M - INI - Dis) , mother shifts attention (distracts) child by holding a conversation about a non - task related topic or by pointing Maternal Joined Distraction (M - JNT - Dis), mother joins child in distracting conversation or activity that child has initiated, (5) Maternal Verbal Orientation to Delay (VO2D) , statements Positive Commands (Pos - Comd), statements that indicate rules of the task without reasons (e.g., 5 For list of variable names and abbreviations as they appear throughout the Method and Results sections please refer to Supplementary Summaries A and B and Table 11. 39 provide expla Negative Commands (Neg - Comd), explicit rule Punitive Reactions (Pun), mother scolds child or threatens child about consequences if Minimizing (Min) , statements that Distraction, Verbal Orientation to Delay, Positive Commands, Negative Commands, Punitive Reactions, and Minimizing were counted as frequency sum scores for each 10s interval. Additionally, codes include (10) Physical Restraint (Phy - RST), using physically restrictive keys, hol ding child firmly or forcefully in lap). Physical restraint was coded for each 10s interval with 2 = high restraint indicating forceful and continuous holding of child (e.g., pulling child back forcefully, holding child firmly in lap), 1 = low restraint in dicating gentle, non - continuous low or high physical restraint, (11) Positive Emotional Reactions (Pos - Aff) , mother shows overt positive affect to the chi ld (e.g., smiling, laughing), and (12) Negative Emotional Reactions (Neg - Aff), sighs, harsh tone) or displays positive affect when child is crying (e.g. laughs at child in m ocking way). Maternal affect was coded as an ordinal variable ( - 1 = negative, 0 = neutral, 1= positive) and dichotomized to create two categorical variables (1) Positive Emotional Reactions (1 = Positive affect, 0 = Neutral or Negative affect), and (2) Neg ative Emotional Reactions (1 = Negative affect, 0 = Neutral or Positive affect). The coders were instructed to code for the most 40 predominant affect during each 10s interval thus each interval was coded with mutual exclusivity of either positive, neutral, o r negative affect. In cases during which mothers displayed positive affect for majority of the interval with any instance of observed negative affect, the interval was coded as negative. Descriptions are presented in Table 3. 6 Maternal Mind - Mindedness. Ma ternal mind - mindedness was coded using Mind - Mindedness Coding Manual, version 2.1 (Meins & Fernyhough, 2015) 7 . The coding system for maternal mind - mindedness yields two dichotomous and mutually exclusive codes including (1) Appropriate Mind - Related Comment s (AMM) (2) Non - Attuned Mind - related Comments (NMM). Because Appropriate Mind - Related Comments and Non - Attuned Mind - related Comments have not been found to share significant association and have distinctly explained variance in attachment organization, Mei ns and colleagues suggest that each tendency represents separate indices or dimensions of mind - mindedness (2012). Therefore, both Appropriate and Non - Attuned Mind - R elated Comments were coded to examine the unique effects of each on regulation and delay of gratification. Mind - related comments were coded based on the following criteria outlined by Meins & Fernyhough (2015). Any comment that (a) uses an explicit internal state term to comment on what the toddler may be thinking, exp eriencing, or feeling. Mind - related comments include internal state terms that reflect emotions (e.g., happy, sad), cognitions (e.g., think, obsessed), desires and preferences (e.g., like, want), epistemic states (e.g., teasing) or (b) comments that reflec 6 Fried lmeier et al., (2015a) does not include Ignoring in their coding manual, however, Spinrad et al., (2007a) consider Ignoring as an unsupportive strategy, and therefore, Ignoring (e.g., verbally or non - verbally) was c oded. However, due to the low number of observed instances of ignoring , (M = .00, SD = .06, Skewness = 14.9, 0.004 % of intervals ), it was not included in final analyses. Additionally, due to the low number of observed instances of Removing Keys (e.g., mot SD = .23, Skewness = 3.81, 5.7 % of intervals) it was not included in final analyses 7 For descriptions on exploratory overlap of MRAs and MM see Appendix A. 41 Statements that indicate include any utterance that is meant to be said or though by the toddler, and may not always contain inter After identification of all mind - related comments, they were dichotomously categorized as appropriate or non - attuned based on the following criteria. Mind - related comments were said while the toddler is smiling or laughing (Meins & said while the toddler is fussing and reaching for keys, or (b) the comment linked current activity with similar events in the past or future (e.g., ing at the keys with Pooh Bear attached to it), or (c) the comment served to clarify how to proceed after a said after toddler has been gazing around the room not focused on any object or activ ity for 2 - manual, mind - related comments were coded as Non - Attuned if (a) the coder disagreed with the after the toddler has shown no over t signs of tiredness), or (b) the comment refers to a past or future event you tomorrow having not previously mentioned Grandma), or (c) the caregiver asks what the toddler wants to do or suggests that the toddler wants to be involved in a new activity when the count? when the toddler is attending to keys) or (d) the mother seems to be attributing internal 42 when the toddler is not playing with or attending to any particular object or event). To ensure accuracy in coding, all mind - related comments were identified by the first author before th ey were coded as appropriate or non - attuned by trained undergraduate research assistants. Descriptions are presented in Table 4. 43 Table 3 Description of Maternal Regulatory Attempts Variable Name Description Example 1. Verbal Comfort (V - Comf) · Stat efforts to wait. · Statements that reassure the child that he/she will soon get the keys. · Statements that encourage the expression of internal states. 2. Physical Comfort (Phy - Comf) · Behaviors that provide physical comfort for the child. hand 3. Maternal Initiated Distracti on (M - INI - Dis) · Mother shifts attention (distracts) child by holding a conversation about a non - task related topic or by pointing out objects in the room, making suggestions for activities. Playing. 4. Maternal Joined Distraction (M - JNT - Dis) · Mother joins child in distracting conversation or activity that child has initiated. Child plays with toes and mom says 5. Maternal Verbal Orientation to Delay (M - VO2D) rd the delay task. 6. Positive Commands (Pos - Comd) · Statements that indicate rules of the task without reasons. · Suggestive statements · May provide explanations for compliance based on norms, values, or consequences 7. Negative Commands (Neg - Comd) · Explicit rule statements or requests that specify t desired action in negative terms lated to the Key Task. 44 T able 3 ( c ) 8. Punitive Reactions (Pun) · Mother scolds child or threatens child about consequences if she/he is not obedient when he reaches for keys 9. Minimizing (Min) · Statements t fun of, or tease child crying 10. Physical Restraint (Phy - RST) · Using physically restrictive behavior to stop child from touching ke ys ( 0 = no physical restraint, 1 = low physical restraint (gentle, non - continuous), 2 = high physical restraint (forceful, continuous) Low mother may gently tap High firmly holding child in lap 11. Positive Emotional Reac tions (Pos - Aff) . Mother shows overt positive affect to the child - Smiling at child, Laughing with child, expressing joy 12.Negative Emotional Reactions (Neg - Aff) positive affect when child is crying - Heavy sighs, harsh tone, laughs to mock child, yells at child 45 Table 4 Description of Maternal Mind - mindedness Variable Name Description Example Appropriate Mind - Related Comments (AMM) Mind - related comments which: . Accurately id entify/interpret the internal states of the toddler including emotional, cognitive, desires, goals, preferences based on observable behaviors . Link current activity with similar events in the past or future, the comment. . Serve to clarify how to proceed after a lull in the interaction. said while child is looking at keys or saying keys said while child is frustrated after she to Non - Attuned Mind - Related Comments (NMM) Mind - related comments which: . Non - accurately identify/interpret the internal states of the toddler based on observable behaviors . Refer to past or future events, unrelated to activity. . Ask what the toddler wants to do or suggest that the toddler wants to be involved in a new activity when the toddler is actively engaged in playing with or attending to something else. . Attribute internal states that are not implied by the tod states onto the child said while toddler shows no overt signs of fatigue said when toddler is attending to keys Toddler Regul atory Strategies. Toddler regulatory strategies were coded based on Coding System for Delay of Gratification Task Child Behavior and Regulatory Attempts (Friedlmeier, Corapci, & Benga, 2015b) as a frequency sum score for each 10s interval. Friedlmeier et exclusive categories (1) independent regulatory strategies and (2) dependent regulatory strategies. The coding system was adapted for the current study and identified 13 strat egies that were classified into two mutually exclusive categories (1) independent regulatory strategies (1 - 6), and (2) dependent regulatory strategies (7 - 13) described below (Tables 5 and 6). Strategies were based on established coding schemes from previou s studies (Grolnick et al., 1996; Gilliom 46 et al., 2002). Strategies were coded in 10s intervals and were calculated as the sum of total strategies for each strategy per 10s interval. initiated attemp ts to regulate by him/herself that do not involve the mother and include 6 mutually exclusive strategies. Strategies include (1) Toddler Initiated Verbal Distraction (T - INI - Vdis), verbal behaviors that shift attention away from keys/task by describing or t alking about other things in the room or other Toddler Initiated Non - Verbal Distraction (T - INI - nonVDis) , toddler shifts focus away from keys for more than 3 seconds (e.g., looks away, walks out of the room, engages directed towards the delay task, in the form of (3) Toddler Initiated Verbal Keys(T - INI - Vkeys), Toddler Initiated Verbal Self - Control(T - INI - Vctrl) Toddler - Initiated - Verbal Desire (T - INI - Vdes), any self - talk expressing de Toddler Self - Comfort (T - SComf) , child physically comforts self (e.g., thumb sucking). Descriptions are presented in Table 5. r involving the mother and include 7 mutually exclusive strategies. Strategies include (7) Toddler Verbal Distraction - Bids to Mom (T - VDis - 2M), toddler - initiated verbal behaviors that shift attention away from keys/task by describing or talking about other things in the room or other activities that were also looking at mom), these behaviors are the same as Toddler - Initiated Verbal Distraction (T - INI - VDis) except that they are also di rected towards or engage mother and are thus categorized under 47 dependent regulatory strategies, (8) Toddler Joined Verbal Distraction (T - JNT - VDis) , toddler joins in alternative verbal activity that mother has initiated (e.g., mom starts counting and child joins in), (9) Toddler Joined Non - Verbal Distraction (T - JNT - nonVDis) , child joins in alternative non - verbal activity that mother has initiated (e.g., mom initiates play and child joins in, mom points to a direction away from keys and child looks where mom is pointing). Dependent also directed towards mother (looking at mother or in response to mother), in the form of (10) Toddler Verbal Keys - Bids to Mom (T - Vke ys - 2M), verbalizations that describe said while looking at or in response to mother), (11) Toddler Verbal Self - Control - Bids to Mom (T - Vctrl - 2M) said while looking at or in response to mother) and (12) Toddler Verba l Desire - Bids to Mom (T - Vdes - 2M) , any self - said while looking at or in response to mother), and (13) Toddler Physical Comfort Seeking (T - ComfSk) , child seeking physical closeness to mother (e.g., reac hing arms up to mother, hugging mother, leaning on 48 Table 5 Description of Toddler Independent Regulatory Strategies Variable Name Description Example 1. Toddler Initiated Verbal Distraction (T - INI - VD is) Toddler - initiated verbal behaviors that shift attention away from keys/task by describing or talking about other things in the room or other activities 2. Toddler Initiated Non - Verbal Distraction (T - INI - nonV - Dis) Todd ler shifts focus away from keys for more than 3 seconds looks away, walks out of the room, engages in play 3. Toddler Initiated Verbal Keys (T - INI - Vkeys) verbalizations that describe the keys or reframed the task turtle is sleeping, wake up 4. Toddler Initiated Verbal Self - Control (T - INI - Vctrl) reference to rules of the task 5. Toddler Initiated Verbal Desire (T - INI - Vdes) Any self - talk expressing desire for the keys 6. Toddler Self - Comfort (T - SComf) Child physically comforts self Thumb sucking, rocking back and forth Note. The attractive set of keys was attached to diffe 49 Table 6 Description of Toddler Dependent Regulatory Strategies Variable Name Description Example 7. Toddler Verbal Distraction - Bids to Mom (T - Vdis - 2M) Todd ler - initiated verbal behaviors that shift attention away from keys/task by describing or talking about other things in the room or other activities that were also directed towards mother singing while looking at mom. 8. Toddler Joined Verbal Distraction (T - JNT - VDis) Toddler joins in alternative verbal activity that mother has initiated Mother starts counting and child joins in. 9. Toddler Joined Non - Verbal Distraction (T - JNT - nonV - Dis) Toddler joins in altern ative non - verbal activity that mother has initiated Mother initiates play and child joins in, Mother points to a direction away from keys and child looks where she is pointing. 10. Toddler Verbal Keys - Bids to Mom (T - Vkeys - 2M) verbalizat ions that describe the keys or reframed the task said while looking at or in response to mother 11. Toddler Verbal Self - Control - Bids to Mom (T - Vctrl - 2M) orm of self direction or in reference to rules of the task said while looking at or in response to mother 12. Toddler Verbal Desire - Bids to Mom (T - Vdes - 2M) Any self - talk expressing desire for the keys said while looking at o r in response to mother 13. Toddler Physical Comfort Seeking (T - ComfSk) Toddler seeking physical closeness to mother e.g., reaching arms up to mother, body Note. The at Key Task. 50 of negative emotion was coded base d on the Coding System for Delay of Gratification Task - Intensity of Emotion Expression (Friedlmeier, Corapci, & Benga 2015c). The coding system was designed to distinguish valence (Happy, Sad, and Angry) and intensity (0 - 3) of emotions, from low to high in been primarily designed to elicit anger and used to assess anger expression, thus criteria for alence was reflected lowering of voice from previous volume without intention to whisper or dropped off at the end of utterance, and if child was whining. Facial beginning to pull down, bottom lip appearing loose, droopy eyes, oblique shaped (^) brows. shoulders/body were slumping or slacking, eye coded when vocal cues reflected harsh voice, protest, irritation, and pitch of voice was loud and mouth squared off if op en, and lips pressed or tightened if mouth was closed. Posture/gesture cues for anger reflected arms akimbo (fists placed on each hip), finger wagging or jabbing. Aggressive behaviors (e.g. punching) were not codable without additional anger cues. Emotion intensity was coded on a four point scale (0 - kicking feet to the ground). 51 In order to capture maximum variation in expression of negative emotions, considering both intensity and temporal features (e.g. duration) of emotional expression, the final coding was done as the following to yield an intensity of expression of negative emotion (ENE - I), and a predominance of expression of negative emotion (ENE - P) score for each interval. Using the four point scale, we coded the highest observed intensity during the 1 st 5s and the highest observed intensity during the 2 nd 5s for each 10s interval. Then, an average intensity score was calculated for the entire 10s interval using the highest observed intensity scores for the 1 st 5s, and the 2 nd 5s for each interval. Using the four poi nt scale, we additionally coded the most predominant intensity level (i.e., the intensity level with longest duration in seconds) observed during the 1 st 5s and most predominant intensity level (i.e., the intensity level with longest duration in seconds) d uring the 2 nd 5s for each interval. Then, an average predominance score was calculated for the entire 10s interval using the predominance scores (longest observed intensity level) for the 1 st 5s, and the 2 nd 5s for each interval. This coding scheme was don e to distinguish highest intensity (ENE - I) expressed from most enduring level of intensity (ENE - P). For instance, during a 5s interval, a toddler who screamed (level - 3 intensity) for 1 second and whimpered (level - 1 intensity) for 4 seconds would receive a score of 3 for intensity and a score of 1 for predominance . Finally, an average overall expression score reflective of both intensity and predominance was calculated using total ENE - I and ENE - P for each interval. Overall Expression of Negative Emotion (ENE - O) was calculated as the average of intensity scores for the 1 st 5s, the 2 nd 5s, predominance score (longest observed intensity level) for the 1 st 5s, and the predominance score for the 2 nd 5s for each interval (Table 7). Descriptive, correlational, and m ultilevel analyses are reported on ENE - O. 52 variable and reflected a frequency score comprised of touching and attempting to touch the keys as follows. Touch was scored as having occurred each time the toddler touched the keys and Attempted Touch was scored as having occurred each time the toddler attempted to touch the keys but was physically restrained by mother for each 10s interval. Behaviorally, both Touch and Attempted Touch reflect lower capacity to wait. Therefore, Touch and Attempted Touch were summed to create an Overall Touch score for each interval with higher scores indicating lower delay of gratification (Table 7). Descriptive, correlational, and multilevel anal yses are reported on Overall Touch. Table 7 Description of Toddler Expression of Negative Emotion and Delay of Gratification Variable Name Description Scale ICC ENE - O Average of ENE - I 1 st 5s, ENE - I 2 nd 5s, ENE - P 1 st 5s, and ENE - P 2 nd 5s .96 ENE - I - 1st 5s Expression of Negative Emotion Intensity - ENE - I - 2nd 5s Expression of Negative Emotion Intensity - ENE - I Average of ENE - I 1 st 5s and 2 nd 5s .91 ENE - P - 1st 5s Expression of Negative Emotion Predominance - ENE - P - 2nd 5s Expression of Neg ative Emotion Predominance - ENE - P Average of ENE - P 1 st 5s and 2 nd 5s .92 Touch Sum of times child touched the keys - Attempted Touch Sum of times child tried (reached) to touch the keys but was restrained by mother - Overall Touch Sum of Touch and Attempted Touch - Note. ENE - O = Overall Expression of Negative Emotion, ENE - I = Intensity of Expression of Negative Emotion, ENE - P = Predominance of Expression of Negative Emotion, s = seconds. Covariates 8 . ariable at baseline (1=male), - month MacArthur Communicative Development Inventories Vocabulary Short Form (Fenson et al., 1993). 8 Note to committee: For list of cova riate names and abbreviations as they appear throughout the Method and R esults sections please refer to Table H1 in Appendix H (also provided on page 2 of Supplemental Materials). . 53 Mothers were asked to respond with a yes (1) or no (0) in dicating if they had heard their child use the word. All 100 items were summed to yield a Vocabulary Production score. Additionally, in order to examine effects of verbal MRAs and MM beyond effects of maternal talkativeness, maternal verbosity was used as a level - 1 covariate in models with verbal maternal predictors including Verbal Comfort, Verbal Orientation to Delay, Positive Commands, Negative Commands, Minimizing, as well as Appropriate Mind - Related Comments, and Non - Attuned Mind - related Comments (Que stions 1 & 3). Before coding, all comments made by mothers and toddlers were transcribed verbatim from the videotaped interactions during the delay of gratification task by a trained coder. Maternal verbosity was computed as sum of all maternal comments. A comment was defined as any utterance of one - or multiple - - mintue delay of gratification task ( M = 3.01, SD = 2.30, Range = 0 - 15 comments/interval). To contr ol for linear and quadratic dependency of outcomes on time, when all predictors were zero (e.g., expected fatigue), Elapsed Time - Linear (ET) and Elapsed Time Quadratic (ET 2 ) were calculated as time varying predictors in seconds per interval and used as l evel - 1 covariates in all random coefficient models. 9 Coding 10 All variables were coded using event - based coding conducted along 10s intervals, which yielded a sum of 12 intervals for the entire task. Coding occurred during three phases including (1) initial training phase, (2) baseline training phase, and (3) independent coding phase. Coders 9 See Appendix D for results of descriptive statistics and ANOVA conduc ted to describe direction of change in study variables including Overall Expression of Negative Emotion and Overall Touch (Table D1), Maternal Regulatory Attempts, Appropriate, and Non - Attuned Mind - Regulatory Stra tegies (Table D3). Curve Estimations examining the relationship between time and outcome variables are presented in Figure D1. 10 For detailed description for Coder Training and Inter - Coder Reliability procedures refer to A ppendix B. 54 reached inter - coder reliability of 0.80 or greater for all variables before they could be considered reliable to code independently in phase 3. During phase 3, 35% of ca ses were randomly double coded for all variables to ensure consistent and accurate coding. Inter - coder reliability estimates for all study variables ranged between acceptable to very good (0.66 - 1.00). For all reliability estimates refer to Tables 8 - 10. B ased on standards by Altman (1991), MedCalc for Windows, version 16.2.0 (MedCalc Software, Ostend, Belgium) was used to calculate the inter - rater agreement statistics (weighted kappa and intraclass correlation estimates) with 95% confidence interval for no minal and ordinal variables. Kalpha ( Krippendorff, 1997; 2004) was used to estimate inter - coder reliability for all variables with Reliability Calculator for Ordinal, Interval, and Ratio Data (ReCal OIR; Freelon, 2013). Kalpha allows for the estimation of inter - coder reliability with specificity of the measurement scale (Krippendorff, 2004). Thus, Kalpha is reported for all variables according to - Ratio, Expression of N egative Emotion is reported as Kalpha - Interval, Maternal Emotional Reactions and Physical Restrained are reported as Kalpha - Cohen, 1960) is reported for nominal data. 55 Table 8 Interobserver Reliability Estimates for Maternal Regulatory Attempts, Appropriate Mind - Related, and Non - Attuned Mind - Related Comments Variable Name ICC Kappa Weighted Kappa Kalpha Mean (Range) Mean (Range) Mean (Range) Mean (Range) V - Comf 1.00(1.00 - 1.00) .76 (.74 - .77) .85 (.83 - .83) .76 (.74 - .76) Phy - Comf .95 (.82 - 90) .86 (.70 - .96) .88 (.70 - .91) .87 (.70 - .90) M - INI - Dis .94 (.92 - .98) .82 (.62 - .95) .84 (.71 - .96) .85 (.67 - .96) M - JNT - Dis .91 (.80 - .95) .84 (.66 - .90) .84 (.66 - .90) .84 (.66 - .90) M - VO2D 1.00 (.99 - 1.0) .86 (.81 - .88) .91 (.90 - .92) .89 (.88 - .89) Pos - Comd .99 (.90 - .99) .78 (.75 - .78) .84 (.80 - .84) .86 (.80 - .91) Neg - Comd 1.00 (1.00 - .97) .84 (.83 - .84) .94 (.89 - .96) .93 (.92 - .95) Pun 1.00 0.76 0.87 0.82 Min 1.00 0.82 0.90 0.82 Ph y - RST .94 (.93 - .93) .83 (.81 - .86) .86 (.84 - .87) .88 (.87 - .88) Affect .89 (.88 - .92) .82 (.77 - .89) .81 (.77 - .87) .81 (.77 - .82) AMM .91 (.87 - .97) .80 (.73 - .94) .81 (.74 - .94) .80 (.72 - .94) NMM .96 (.96 - .97) .85 (.82 - .91) .89 (.8 7 - .92) .87 (.85 - .91) Note. ICC = Intraclass Correlation, V - Comf = Verbal Comfort, Phy - Comf = Physical Comfort, M - INI - Dis = Maternal Initiated Distraction, M - JNT - Dis = Maternal Joined Distraction, M - VO2D = Maternal Verbal Orientation to Delay, Pos - Comd = Positive Commands, Neg - Comd = Negative Commands, Pun = Punitive Reactions, Min = Minimizing, Phy - RST = Physical Restraint, AMM = Appropriate Mind Related Comments, NMM = Non - Attuned Mind Related Comments. Affect was coded as an ordinal variable and late r dichotomized to create presence/absence of positive and negative affect; (Weighted Kappa 0.61 - 0.80 = Good, 0.81 - 1.00 = Very Good, Altman, 1991; Kalpha .80 = Very Good, 0.67 = Acceptable) 56 Table 9 Interobserver Reliability Estimates for Tod dler Regulatory Strategies Variable Name ICC Kappa Weighted Kappa Kalpha Mean (Range) Mean (Range) Mean (Range) Mean (Range) Independent Regulatory Strategies T - INI - VDis 1.0 (1.0 - 1.0) .70 (.67 - .80) .80 (.76 - .81) .73 (.69 - .79) T - INI - nonVDis 1.0 (1.0 - 1.0) .67 (.63 - .72) .74 (.69 - .77) .74 (.71 - .76) T - INI - Vkeys 1.0 (1.0 - 1.0) .96 (.91 - .98) .98 (.95 - .98) .96 (.91 - .98) T - INI - Vctrl 1.0 (1.0 - 1.0) .95 (.90 - 1.0) .98 (.96 - 1.0) .95 (.90 - 1.0) T - INI - Vdes 1.0 (1.0 - 1.0) 1.0 (1.0 - 1. 0) 1.0 (1.0 - 1.0) 1.0 (1.0 - 1.0) T - SComf 1.0 (.99 - 1.0) .75 (.71 - .78) .82 (.75 - .87) .82 (.78 - .85) Dependent Regulatory Strategies T - VDis - 2M 1.0 (1.0 - 1.0) .68 (.66 - .68) .80 (.75 - .81) .71 (.66 - .75) T - JNT - VDis 1.0 (1.0 - 1.0) .73 (.67 - .81) .84 (.69 - .89) .77 (.72 - .84) T - JNT - nonVDis 1.0 (1.0 - 1.0) .74 (.67 - .81) .86 (.72 - .87) .77 (.70 - .82) T - Vkeys - 2M 1.0 (1.0 - 1.0) .98 (.97 - 1.0) .99 (.98 - 1.0) .98 (.97 - 1.0) T - Vctrl - 2M 1.0 (1.0 - 1.0) .95 (.90 - 1.0) .98 (.96 - 1.0) .95 ( .90 - 1.0) T - Vdes - 2M 1.0 (1.0 - 1.0) 1.0 (1.0 - 1.0) 1.0 (1.0 - 1.0) 1.0 (1.0 - 1.0) T - ComfSk 1.0 (.99 - 1.0) .70 (.62 - .77) .79 (.66 - .84) .74 (.69 - .78) Note. ICC = Intraclass Correlation, T - INI - VDis = Toddler Initiated Verbal Distraction, T - INI - non VDis = Toddler Initiated Non - Verbal Distraction, T - INI - Vkeys = Toddler Initiated Verbal Keys, T - INI - Vctrl = Toddler Initiated Verbal Self - Control, T - INI - Vdes = Toddler Initiated Verbal Desire, T - SComf = Toddler Self - Comfort; T - VDis - 2M, = Toddler Verbal Dis traction - Bids to Mom, T - JNT - VDis = Toddler Joined Verbal Distraction, T - JNT - nonVDis = Toddler Joined Non - Verbal Distraction, T - Vkeys - 2M = Toddler Verbal Keys - Bids to Mom, T - Vctrl - 2M = Toddler Verbal Self - Control - Bids to Mom, T - Vdes - 2M = Toddler Verbal D esire - Bids to Mom, T - ComfSk = Toddler Physical Comfort Seeking. (Weighted Kappa 0.61 - 0.80 = Good, 0.81 - 1.00 = Very Good, Altman, 1991; Kalpha .80 = Very Good, 0.67 = Acceptable) 57 Table 10 Interobserver Reliability Estimates for Toddler Expression of Negative Emotion and Delay of Gratification Variable Name ICC Kappa Weighted Kappa Kalpha Mean (Range) Mean (Range) Mean (Range) Mean (Rang e) ENE - I - 1st 5s 1.0 (.98 - 1.0) .82 (.63 - .86) .89 (.78 - .92) .92 (.79 - .95) ENE - I - 2nd 5s 1.0 (1.0 - 1.0) .86 (.68 - .90) .92( .81 - .95) .95 (.83 - .97) ENE - P - 1st 5s 1.0 (.97 - 1.0) .85 (.78 - .89) .91 (.86 - .93) .93 (.87 - .94) ENE - P - 2nd 5s 1.0 (1.0 - 1.0) .86 (.69 - .91) .82 (.82 - .94) .94 (.83 - .96) Touch 1.0 (1.0 - 1.0) .90 (.86 - .95) .92 (.89 - .96) .93 (.89 - .97) Attempted Touch 1.0 (1.0 - 1.0) .70 (.69 - .71) .79 (.76 - .80) .78 (.77 - .78) Note. ENE - I = Expression of Nega tive Emotion Intensity, ENE - P = Expression of Negative Emotion Predominance, s = seconds. (Weighted Kappa 0.61 - 0.80 = Good, 0.81 - 1.00 = Very Good, Altman, 1991; Kalpha .80 = Very Good, 0.67 = Acceptable) Plan of Analyses A multilevel modeling ( Byrk & Raudenbush, 1992) approach was used to examine the effects of maternal regulatory attempts, mind - mod eling allows for the distinction between individual - and group - level sources of variance to explain a single individual outcome and can be used to analyze data that has been collected over repeated measures (Snijders & Bosker, 2012). This study utilized tw o - level multilevel models with random intercepts with 12 repeated measurement occasions nested within 134 mother - child dyads. Heterogeneity of within - person variance (e.g., significant differences in Overall Expression of Negative Emotion and Overall Touch ) was examined using intraclass correlations of the intercept - only models (no predictors). Multilevel analyses are well suited to analyze nested data in which error terms may be dependent, which violate assumptions of independence of errors underlying ordi nary least square (OLS) estimation. Thus, in multilevel analyses, concerns related to power and effect sizes are adequately handled as long as there are large numbers of level - 2 units (134 dyads in this study) (Snijders, 2005). 58 Each outcome variable was p redicted in a separate model. Models were specified with level - 1 and level - 2 predictors, random intercepts (intercepts were allowed to vary among toddlers), and fixed (Model 1) or random slopes (Model 2). This structure allowed for the examination of speci fic Maternal Regulatory Attempts, Appropriate and Non - Attuned Mind - gender and expressive language as control variables meaning gender differences or expressive language differences were allowed to partly explain amount of variation in random intercepts over groups. Toddler gender was specified as a level - 2 covariate in all mode ls for Questions 1 - 3 - 2 covariate for models in Question 4 but not Questions 1 - 3. No hypotheses regarding cross - level interactions were made and were not included in analyses. Outc - O). Research Questions 1 and 2 examine the effects of maternal predictors (Question 1) and toddler predictors - O). First, a ba se model was used to test whether there was significant within - toddler variance (across all intervals) - O. Therefore, the first analysis concerns the amount of varian ce in intensity of expression of - O differ across the two - minute task). Operationalized in multilevel modeling framework to test whether there were intraindividual differences in means of ENE - O within toddlers across 12 intervals, the following equation was employed, Level - 1 Equation: The Intercept - Only Model ENE - O ij = 0j +e ij 59 where I indicates the time interval within a dyad and j indicates a mother - child dyad, 0j indicates the intercept (mean) for ENE - O in dyad j (varies over j dyads), e ij indicates random errors of prediction for level - 1 equation (deviation for interval I from its dyad j ). Level - 2 Equation: The Intercept - Only Model 0j = 00 + u 0j where 0j, an intercept (mean) for group j is the sum of the overall intercept, the grand mean of the ENE - O scores across all dyads when all predictors = 0, 00 , and the random error associated with the group intercept, u 0j (deviation for group j from overall intercept) 11 Next, the base models were tested to determine whether Maternal Regulatory Attempts and maternal Mind - N egative Emotion in a current interval. See Table 11 below for the specific list of maternal and - O. Table 11 Maternal and Toddler Predictors of Expression of Negative Emotional Expression and Overal l Touch Observed Predictor Label Predictor Name Maternal Regulatory Attempts (MRA) Verbal Comfort V - Comf Physical Comfort Phy - Comf Maternal Initiated Distraction M - INI - Dis Maternal Joined Distraction M - JNT - Dis Maternal Verbal Orientation to Delay M - VO2D Positive Commands Pos - Comd Negative Commands Neg - Comd Punitive Reactions Pun Minimizing Min Physical Restraint Phy - RST Positive Emotional Reactions Pos - Aff Negative Emotional Reactions Neg - Aff Maternal Mind - Mindedness (MM) Appropriate Mind - Re lated Comments AMM 11 In addition to i ntraclass correlations (between group variance/within + between variance; i.e., u 0j/ e ij+ , u 0j ) , descriptive statistics were used to describe direction of change in levels of ENE - O across 12 intervals (e.g., over time) in A ppendix D (Table D 1). 60 Table 11 ( c Non - Attuned Mind - Related Comments NMM Independent Regulatory Strategies (IRS) Toddler Initiated Verbal Distraction T - INI - VDis Toddler Initiated Non - Verbal Distraction T - INI - nonVDis Toddler Initiated Verbal Keys T - INI - Vkeys Toddler Initiated Verbal Self - Control T - INI - Vctrl Toddler Initiated Verbal Desire T - INI - Vdes Toddler Self - Comfort T - SComf Dependent Regulatory Strategies (DRS) Toddler Verbal Distraction - Bids to Mom T - VDis - 2M Toddler Joined Verbal Distraction T - JNT - VDis Toddler Joined Non - Verbal Distraction T - JNT - nonVDis Toddler Verbal Keys - Bids to Mom T - Vkeys - 2M Toddler Verbal Self - Control - Bids to Mom T - Vctrl - 2M Toddler Verbal Desire - Bids to Mom T - Vdes - 2M To ddler Physical Comfort Seeking T - ComfSk As operationalized in multilevel modeling framework, the following equation was employed to determine what level - one variables predict ENE - O, Level - 1 Equation: The Random Coefficients Model with Predictors at Both Levels ENE - O ij 0j 1j (X i - 1j ) + e ij where i indicates the time interval within a dyad (current interval), i - 1 indicates the current time interval minus one time lag within a dyad (previous interval or lagged) and j indicates a mother - child dyad, 0j indicates the intercept (mean) for ENE - O that varies over dyads j, 1j indicates the slope for the relationship in dyad j between ENE - O and level - 1 predictors (e.g., ElapsedTime, lagged maternal Verbal Comfort, lagged toddler Self - Comfort), and e ij indicat es the random errors of prediction for the level 1 equation (deviation for interval i from its dyad j ). Similarly, we employed the following equation to test what level - two variables predict ENE - O, Level - 2 Equation: The Random Intercept Model with Predict ors at Both levels 0j 00 01 W 1j + u 0j where 00 indicates the overall intercept, the grand mean of the ENE - O scores across all dyads when all predictors = 0 W 1j indicates level - 2 predictor (Gender) 01 indicates the overall regression coefficient for the relationship (slope) between Gender and ENE - O 61 u 0j indicates the random error associated with group intercept ( deviation for group j from overall intercept) 1j = 10 + u 1j 10 indicates the overall regression coefficient between the level one pred ictor and the ENE - O u 1j indicates the random error associated with group slope (deviation for group j from overall slope.) 12 Separate models were specified for all predictors as the following example. Example: Level - 1 Predictor X, Outcome Y, Level - 1 Cova riate W, Level - 2 Covariate Z Level - 1 Equation: Y ij 0j X,j (X i - 1j ETL,j (ET ij ET 2 ,j (ET 2 ij W,j (W i - 1j ) + e ij Level - 2 Equations: 0j 00 Z (Z j ) + u 0j X,j = X + u 1j ETL,j = ETL ETL,j = ETL 2 W,j = W Mixed Model: Y ij 00 Z (Z j ) X (X i - 1j ETL (ET i j ET 2 (ET 2 ij W (W i - 1j ) + u 0j + e ij where 00 indicates the overall intercept, the grand mean of the Y scores across all dyads when all predictors = 0 Z (Z j ) = the average slope between Y and Z times score of Z in dyad j X (X i - 1j ) = the average s lope between Y and lagged X times score of lagged X in dyad j ETL (ET ij ) = the average slope between Y and linear elapsed time times score of linear elapsed time in interval i in dyad j ET 2 (ET 2 ij ) = the average slope between Y and quadratic elapsed t ime times score of quadratic elapsed time in interval i in dyad j W (W i - 1j ) = the average slope between Y and W times score of W in interval i in dyad j u 0j = deviation for dyad j from overall intercept, e ij = deviation for interval i from its dyad j 12 Random slope variance components were estimated and included if they varied between dyads and if model fit was improved; however, no hypotheses were made regarding differences in IV - DV slopes among dyads. Also, no level - 2 explanatory variables were hypothesized t o interact with level - 1IVs to explain differences in ra ndom slopes between toddlers. For all equations of model specification refer to Appendix C . 62 In summary, for Research Question 1: The Random Intercept Model Predicting Expression of Negative Emotion from lagged Maternal Regulatory Attempts and lagged Mind - Mindedness, a total of 14 two - level hierarchical models examined the effects of lagged Maternal Regulatory Attempts, lagged Appropriate and Non - Attuned Mind Related Comments on - O while controlling for effects of linear elapsed time, quadratic elapsed time, lagged maternal verbosity, and toddler gender. For Research Question 2: The Ra ndom Intercept Model Predicting Expression of - level hierarchical and dependent reg - O while controlling for effects As previously mentioned, we used Overall Touc h (sum of Touch and Attempted Touch) to reflect delay of gratification such that higher Overall Touch scores indicate lower ability for delay of gratification. Research Questions 3 and 4 examine the effects of maternal predictors (Question 3) and toddler p test whether there was significant within - toddler variance (across all intervals) before we could all Touch. Therefore, the first analysis concerns the amount of variance in intensity of expression of negative emotion - minute task). Operationalized in multilevel model ing framework to test whether there were intraindividual differences in means of Overall Touch within toddlers across 12 intervals, the following equation was employed, 63 Level - 1 Equation: The Intercept - Only Model Overall Touch ij = 0j +e ij where i indicates the time interval within a dyad and j indicates a mother - child dyad, 0j indicates the intercept (mean) for Overall Touch in dyad j (varies over j dyads), e ij indicates random errors of prediction for level - 1 equation (deviati on for interval i from its dyad j ). Level - 2 Equation: The Intercept - Only Model 0j = 00 + u 0j where 0j, an intercept (mean) for group j is the sum of the overall intercept, the grand mean of the Overall Touch scores across all dyads when all predic tors = 0, 00 , and the random error associated with the group intercept, u 0j (deviation for group j from overall intercept) 13 Next, the base models were tested to determine whether Maternal Regulatory Attempts and maternal mind - mindedness (Research Questi Please refer to Table 11 on pages 59 - 60 for the specific list of maternal and toddler behaviors tested as predictors o As operationalized in multilevel modeling framework, the following equation was employed to determine what level - one variables predict Overall Touch, Level - 1 Equation: The Random Coefficients Model with Predictors at Both Leve ls Overall Touch ij 0j 1j (X i - 1j ) + e ij where i indicates the time interval within a dyad (current interval), i - 1 indicates the current time interval minus one time lag within a dyad (previous interval or lagged) and j indicates a mother - child dyad, 0j indicates the intercept (mean) for Overall Touch that varies over dyads j, 1j indicates the slope for the relationship in dyad j between Overall Touch and level - 1 predictors (e.g., ElapsedTime, lagged AMM, lagged NMM), and e ij indicates the random er rors of prediction for the level 1 equation (deviation for interval i from its dyad j ). 13 In addition to intraclass correlations (between group variance/within + between variance; i.e., u 0j/ e i j+ , u 0j ) , descriptive statistics were used to describe direction of change in levels of Overall Touch across 12 intervals (e.g., over time) in Appendix D (Table D1). 64 Similarly, we employed the following equation to test what level - two variables predict Overall Touch, Level - 2 Equation: The Random Intercept Model with Predictors a t Both levels 0j 00 01 W 1j + u 0j where 00 indicates the overall intercept, the grand mean of the Overall Touch scores across all dyads when all predictors = 0 W 1j indicates level - 2 predictor (Gender) 01 indicates the overall regression coefficient for the relationship (slope) between Gender and Overall Touch u 0j indicates the random error associated with group intercept ( deviation for group j from overall intercept) 1j = 10 + u 1j 10 indicates the overall regression coefficient between th e level one predictor and the Overall Touch u 1j indicates the random error associated with group slope (deviation for group j from overall slope). 14 In summary, for Research Question 3: The Random Intercept Model Predicting Overall Touch from lagged Mater nal Regulatory Attempts and lagged Mind - Mindedness, a total of 14 two - level hierarchical models examined the effects of lagged Maternal Regulatory Attempts, lagged Appropriate and Non - controll ing for effects of linear elapsed time, quadratic elapsed time, lagged maternal verbosity, For Research Question 4: The Random Intercept Model Predicting Overall Touch from - le vel hierarchical models examined the 14 Random slope variance components were estimated and included if they varied between dy ads and if model fit was improved; however, no hypotheses were made regarding differences in IV - DV slopes among dyads. Also, no level - 2 explanatory variables were hypothesized to interact with level - 1IVs to explain differences in ra ndom slopes between todd lers. For all equations of model specification refer to Appendix C 65 elapsed time, quadratic elapsed 15 and expressive language. 15 66 CHAPTER 4: Results Descriptive and correlational statistics were conducted prior to multilevel analyses. Data were analyzed using SPSS (version 24) and HLM software (version 7.1). In order to examine th e effects of independent variables from previous intervals on study outcomes in current interval, lagged predictors were created (lagged = previous 10s interval). Descriptive statistics for outcome variables are presented on pages 68 - 69 , for maternal predi ctor variables on pages 69 - 71 , for toddler predictor variables on pages 73 - 74 . Results specific to Question 1 are presented on pages 7 6 - 7 7 (Correlational) and on pages 85 - 100 (Multilevel); Question 2 are presented on pages 78 - 79 (Correlational) and on page s 102 - 116 (Multilevel); Question 3 are presented on pages 79 - 81 (Correlational) and on pages 118 - 131 ( Multilevel); and, Question 4 are presented on pages 81 - 82 (Correlational) and on pages 133 - 144 (Multilevel). Summaries of the results for each research qu estion are included after each set of re s ults is presented (Q1A on page 93 ; Q1B on page 100, Q2A on page 108, Q2B on page 116 , Q3A on page 118 , Q3B on pages 131 , Q4A on page 139, and Q4B on page 144 ). Missing Data All cases were coded for all study vari ables. During intervals in which the toddler stepped out of camera frame or walked out of the room and could not be observed (20 out of Toddler Initiated Non - Verbal Distraction for walk ing out of the room, expression of negative emotion was coded using vocal cues and zero touch was counted. The Delay of Gratification task lasted 120 seconds or 12 intervals for all cases (n=134), except for cases in which the toddler had been in severe di stress for more than 30 seconds (3.73 %, n = 5), or the experimenter ended the task before the end of the 2 minutes for no apparent reason (2.23 %, n = 3). In dyads for which the experimenter did not 67 end task after 12 intervals (17.16%, n = 23), the remain ing intervals were not included in the final analyses. Additionally, specification of multilevel models with previous interval (lagged 10s) predictors resulted in missing data for 134 intervals (first interval data deleted in each dyad). Level - 1 missing da ta was handled using HLM - 1 level - 2 explanatory variable for overall mean in Overall Touch (Question 4), 3 cases were missing scores for expressive vocabulary and were handled via listwise deletion prior to running models reducing sample size to 131 in Question 4. Assumptions Maternal Regulatory Attempts (except for maternal Physical Restraint, maternal Positive Em otional Reactions and Negative Emotional Reactions), Appropriate and Non - Attuned Mind - and Overall Touch (outcomes) were count variables, where scores reflect ed the number of times the behaviors were observed in each 10s interval. While, ENE - O was slightly positively skewed (1.03, SE = .06), positive skewness was observed in Overall Touch (2.94, SE = .06). Tests of normality were conducted to assess violations of assu mptions of normality for outcome variables . All p values associated with Shapiro - Wilk and Kolmogorov - Smirnov tests were less than .001 indicating violations of assumptions for distributional normality for both outcome variables. Boxplots were obtained and indicated 15 intervals with extreme outliers (data values that were 3 times the interquartile range) for Overall Touch, but no outliers were identified for ENE - O. Given the moment to moment measurement approach of the current study design, the large number of zeros for all count variables was expected as behaviors were not realistically occurring repeatedly in a short amount of time. As noted by Raudenbush, Bryk, Cheong, 68 Congdon, and du Toit (2011) assumptions of normality for count and binary data are not realistic and transformations may not help linearize the data specially when there are many zeros as is the case with the current data. However, to ensure results accuracy, I compared model results between untransformed Overall Touch and transformed (squar e root transformation) and found that model results did not did not change significantly for transformed variables. Additionally, I compared model results between untransformed Overall Touch and Overall Touch after removing extreme outliers. Marginal p - val ues for previous lagged maternal Negative Commands and lagged Physical Restraint fell under .05 in models without outliers. However, significant p - values for lagged maternal Initiated Distraction and lagged Positive Emotional Reactions were no longer signi ficant in models without outliers while all other results remained the same. Therefore, untransformed scores were retained in final analyses. No tests concerning collinearity were conducted as we did not model any interactions. Descriptive Statistics Desc riptive statistics were conducted for outcome variables per 10s interval (Tables 12 - 14) 16 . Toddler Expression of Negative Emotion and Delay of Gratification Per 10s Interval and Overall Task . Descriptive statistics for ENE - O and Overall Touch are reporte d (Table 12). Per 10s interval, on average, toddlers expressed a mild level of intensity ( M = 0.76, SD = 0.99) on a scale of 0 - 3 (0 indicating no cues for negative emotion, 1 = mild, 2 = moderate, 3 = severe level of intensity of negative emotion). Scores for average ENE - O fell between 0 - 1 16 For descriptive statistics for study variables in overall scores across the enti re 2 - minute delay of gratification task (all 120 seconds) for each parent - child dyad (sums and averages for all 12 intervals) see A PPENDIX E (Tables E 1 - E4 ). The only difference between descriptive statistics that are presented in Tables E 1 - E4 and Tables 1 2 - 1 4 is in how variables were calculated across units of time. Tables E 1 - E4 variables were calculated and described across the entire 120 second task (wide format) , while Tables 1 2 - 1 4 variables were calculated and described per 10s interval (long format). D ifferences in means of observed toddler independent regulatory strategies and dependent regulatory strategies wer e analyzed across task (Table E4 ) and per interval (Table E5 ) using paired sample t - tests. 69 (mild) intensity level for majority of intervals (71.8 %, n = 1128), greater than 1 or equal to 2 (moderate) intensity level for 15.3% of the intervals (n = 241), and greater than 2 or equal to 3 (severe) intensity level for 12.9 % of the intervals (n = 202). For the delay of gratification task as a whole , s cores for average ENE - O was zero for 9.7 % of toddlers (n = 13) , fell between 0 - 1 (mild) intensity level for slightly more than half of all toddlers ( 59 %, n = 79 ), grea ter than 1 or equal to 2 (moderate) intensity level for 22.3 % of all toddlers (n = 31 ), and greater than 2 or equal to 3 (severe) intensity level for 9 % of the toddlers (n = 11 ). Overall Touch scores revealed that on average toddlers either touched or at tempted to touch the keys 0 - 6 times per interval ( M = 0.27, SD = 0.61). Out of all 10s intervals for all toddlers (n = 1571), most toddlers either touched or attempted to touch the keys (Overall Touch) 0 times per interval for majority of the intervals (79 .4%, n = 1247), 1 time for 15.8% of intervals (n = 249), 2 times per interval for 3.5 % of intervals (n = 55), and 3 6 times for 1.4 % of intervals (n = 20). For the delay of gratification task as a whole , 23.1 % (n = 31) of toddlers did not touch or att empt to touch the keys at all, while 22.4% (n = 30) of toddlers touched or attempted to touch the keys 1 time, and 54.5% (n = 73) of toddlers touched or attempted to touch the keys 2 or more times. Table 12 Descriptive Statistics for Toddler Expression of Negative Emotion and Overall Touch Per 10s Interval Min Max Mean SD Skewness (SE) ENE - O 0.00 3.00 0.76 0.99 1.03 (.06) Overall Touch 0.00 6.00 0.27 0.61 2.94 (.06) Note . ENE - O = Overall Expression of Negative Emotion (average of ENE - I and ENE - P); O verall Touch = sum of Touch and Attempted Touch Maternal Regulatory Attempts and Mind - Mindedness Per 10s Interval and Overall Task . Descriptive statistics are reported in Table 13 . Maternal Regulatory Attempts, Appropriate and Non - Attuned Mind - Related Comments were used infrequently and on average occurred less than 1 time per internal. Maximum variation in frequency of strategies used was 70 observed for Negative and Positive Commands, while minimum variation in frequency of strategies used was observed for maternal Joined Distraction, Appropriate and Non - Attuned Mind - Related Comments. These results were similar to those found in previous studies (Meins et al., 2001; Spinrad et al., 2004). The most frequently used regulatory attempts were Negative Command least frequently used regulatory attempts were Negative Emotional Reactions (e.g., yells at distractions. Punitive reactions were more frequently use d than Minimizing statements. Mothers did not use any form of Physical Restraint majority of the time (59.4% of all intervals, n= 933), however, when mothers did use Physical Restraint as a form of regulatory strategy (40.6 % of all intervals, n = 638), ma jority of time high restraint was used (23.8% of all intervals, n = 374) compared to low restraint (16.8 % of all intervals, n = 264). For majority of the time, mothers did not display either positive or negative emotional reactions (58.1 % of all interval s, n = 913). Negative emotional reactions were expressed in 2% of all intervals (n =32) while positive emotional reactions were expressed in 39.8 % of all intervals (n = 626). Overall means for maternal mind related comments per interval suggest that on av erage mothers made more Appropriate Mind - Related than Non - Attuned Mind - related Comments (Table 13). Additionally , the percent and number of mothers who did not use any strategies and those who used each strategy at least one or more times during the task as a whole (as opposed to the per interval results presented previously) is summarized in Table 13. As noted in the table, more than half of mothers used the following strategies at least once during the 2 - minute task: v erbal and physical 71 comfort, initiat ed distraction, verbal orientation to delay, positive and negative commands, physical restraint, and positive emotional react ions. About one third of mothers used mind - minded comments at least once during the delay task. 72 Table 13 Descriptive Statistics of for Maternal Regulatory Attempts, Appropriate and Non - Attuned Mind - Related Comments Per 10s Interval Overall Task Frequency % (n) Variable Name Min Max Mean SD Skewness (SE) 0 > or = 1 Verbal Comfort 0 6 0.14 0.5 4.94 (.06) 48.5 (65) 51.5 (69) Physical Comfort 0 4 0.18 0.48 3.06 (.06) 45.5 (61) 54.5 (73) Initiated Distraction 0 5 0.3 0.53 2.04 (.06) 32.1 (43) 67.9 (91) Joined Distraction 0 2 0.07 0.26 3.83 (.06) 66.4 (89) 33.6 (45) Verbal Orientation to Delay 0 7 0.36 0.98 3.13 (.06) 46.3 ( 62 ) 53.7 ( 72 ) Positive Commands 0 10 0.42 0.94 3.09 (.06) 26.1 ( 35 ) 73.9 ( 99) Negative Commands 0 11 0.59 1.26 3.12 (.06) 15.7 ( 21 ) 84.3 ( 113 ) Punitive Reactions 0 3 0.05 0.27 6.55 (.06) 77.6 ( 104 ) 22.4 ( 30 ) Minimizing 0 4 0.03 0.22 10.32 (.06) 84. 3 ( 113 ) 15.7 ( 21 ) Physical Restraint 0 2 0.64 0.84 0.75 (.06) 15.7 ( 21 ) 84.3 ( 113 ) Positive Emotional Reactions 0 1 0.4 0.49 0.42 (.06) 14.2 ( 19 ) 85.8 ( 115 ) Negative Emotional Reactions 0 1 0.02 0.14 6.8 (.06) 89.6 ( 120 ) 10.4 ( 14 ) Appropriate Mind Relat ed Comments 0 3 0.08 0.31 4.61 (.06) 60.4 ( 81 ) 39.6 ( 53 ) Non - Attuned Mind Related Comments 0 3 0.07 0.31 5.44 (.06) 68.7 ( 92 ) 31.3 ( 42 ) Note. Frequency for each variable in Overall Task indicates percentage and number of mothers out of n = 134 who did no t use strategy at all (0) or used it at least once (> or =1) , for Physical Restraint 0 indicates NO restraint, 1 or 2 indicate low or high restraint, for Positive or Negative Emotional Reaction 0 indicates Emotional Reaction was not observed at all, > or = 1 indicates Emotional Reaction was observed at some point during the overall task. 73 Per 10s Interval and Overall Task . M eans for the frequency of toddler regulatory strategies are reported (Table 14). Additionally , the percent and the number of toddler s who did not display any strategies and those who displayed each strategy at least one or more times during the task as a whole (as opposed to per interval) is presented in Table 14. Independent Regulatory Strategies Per 10s Interval and Overall Task . In general, Independent Regulatory Strategies were used infrequently and on average occurred less than 1 time per interval. Maximum variation in frequency of strategies used was observed for toddler initiated verbalizations - comforting behaviors (e.g., thumb sucking, rocking back and forth). Minimum variation in frequency of st The most frequently displayed independent strategy by toddlers was Self - Comfort (e.g., thumb sucking, rocking back and forth), while the least frequently displ ayed independent strategy was Non - Verbal Distraction (e.g. looks away for more than 3 seconds) per 10s interv al. Toddlers the form of self - direction or in reference to the towards or involved mother (described below). More than 75% of toddlers used nonverbal distraction independently and self - comforting one or more times during the entire 2 minute task. Other strategies oc curred less frequently. 74 Dependent Regulatory Strategies Per 10s Interval and Overall Task . In general, Dependent Regulatory Strategies were used infrequently and on average occurred less than 1 time per interval. Maximum variation in frequency of strate gies used was observed for toddler Joint Non - Verbal Distraction (e.g., distracting activity). Minimum variation in frequency of strategies used was observed for toddler verbalizations about keys (e. - Comfort - fr equently displayed strategy directed toward or in response to mother was Physical Comfort displayed strategy directed toward or in response to mother were verb alizations in the form of self - On average, toddlers displayed more Joined Non - Verbal Distraction than Joined Verbal Distraction. On average, toddlers vocalized more words or phrases de scribing the keys or - direction or in reference to the rules of the task (e.g. - 20% of all intervals were scored with maximum number of frequencies for all dependent regulatory strategies. In terms of independent strategies, more than 60% of toddlers sought physical comfort and more than 40% utili zed non - verbal distraction directed at or in response to the mother. Other strategies occurred less frequently. 75 Table 14 Descriptive Statistics for Toddler Regulatory Strategies Per 10s Interval Overall Task Frequency % (n) Variable Name Min Ma x Mean SD Skewness (SE) 0 > or = 1 Independent Regulatory Strategies (IRS) Initiated Verbal Distraction 0 2 0.01 0.11 12.30 (.06) 93.3 ( 125 ) 6.7 ( 9 ) Initiated Non - Verbal Distraction 0 4 0.45 0.59 1.12 (.06) 23.1 ( 31 ) 76.9 ( 103 ) Initiated Verbal Keys 0 5 0.04 0.25 9.96 (.06) 82.8 ( 111 ) 17.2 ( 23 ) Initiated Verbal Self - Control 0 0 0 0 0 0 0 Initiated Verbal Desire 0 5 0.02 0.18 17.93 (.06) 93.3 ( 125 ) 6.7 ( 9 ) Self - Comfort 0 5 0.5 0.68 1.26 (.06) 11.9 ( 16 ) 88.1 ( 118 ) Dependent Regulatory Strategies (DR S) Verbal Distraction - Bids to Mom 0 4 0.06 0.3 6.44 (.06) 74.6 ( 100 ) 25.4 ( 34 ) Joined Verbal Distraction 0 4 0.06 0.31 6.60 (.06) 79.9 ( 107 ) 20.1 ( 27 ) Joined Non - Verbal Distraction 0 5 0.14 0.41 4.32 (.06) 59.7 ( 80 ) 40.3 ( 54 ) Verbal Keys - Bids to Mom 0 3 0.08 0.38 5.24 (.06) 70.1 ( 94 ) 29.9 ( 40 ) Verbal Self - Control - Bids to Mom 0 3 0.01 0.13 16.03 (.06) 96.3 ( 129 ) 3.7 ( 5 ) Verbal Desire - Bids to Mom 0 4 0.03 0.21 8.95 (.06) 81.3 ( 109 ) 18.7 ( 25 ) Physical Comfort Seeking 0 3 0.24 0.48 1.91 (.06) 37.3 ( 50 ) 62.7 ( 84 ) Note. The relatively wide ranges of frequency for Initiated Non - Verbal Distraction and Self - Comfort per interval, but relatively low means, may be related to the different frequencies with which behaviors were observed. For majority of interval s (60%) toddlers did not display any Initiated Non - Verbal Distraction or any Self - Comfort and displayed the maximum number of Initiated Non - Verbal Distraction (4) and Self - Comfort (5) for only 10% of intervals. In fact, less than 10% of all intervals were scored with maximum number of frequencies for all independent regulatory strategies . Frequency for each variable in Overall Task indicates percentage and number of toddler out of n = 134 who did not use strategy at all (0) or used it at least once (> or = 1 ) at some point during the task. 76 Correlational Analyses Correlational analyses were conducted to examine associations between predictor and outcome variables. A total of five correlation tables are reported (Tables 15 - 19). First, correlations for lagg ed Maternal Regulatory Attempts, Appropriate and Non - Attuned Mind - Related Comments with Expression of Negative Emotion are presented Table 15 (Question 1). Next, correlations for lagged Toddler Regulatory Strategies with Expression of Negative Emotion are presented in Table 16 (Question 2). Next, correlations for lagged Maternal Regulatory Attempts, Appropriate and Non - Attuned Mind - Related Comments with Delay of Gratification are presented Table 17 (Question 3). Finally, correlations for lagged Toddler Regu latory Strategies with Delay of Gratification are presented in Table 18 (Question 4). Additionally, Pearson correlations for Expression of Negative Emotion and Delay of Gratification (Table 19) with toddler gender and expressive language are reported. Lag ged Maternal Variables and Overall Expression of Negative Emotion. First, correlations for lagged Maternal Regulatory Attempts, lagged Appropriate and lagged Non - Attuned Mind - Related Comments with ENE - O are presented (Table 15) 17 . Similar to correlations fo r current interval Maternal Regulatory Attempts with ENE - O, all correlations for lagged Maternal Regulatory Attempts with ENE - O were small but significant (r = - .32 to .06), except for non - significant correlations for lagged Positive Commands, and for lagg ed Maternal Verbal Orientation to Delay. Unexpectedly, as lagged Verbal Comfort and lagged Physical Comfort increased, ENE - O increased. All other correlations were in expected directions. Strongest negative associations for both current and previous interv al predictors were 17 See A ppendix F for current interval correlation s for Maternal Regulatory Attempts, Appropriate and Non - Attuned Mind - Related Comments with Expression of Negative Emotion (Table F 1). 77 while strongest positive associations were between lagged Negative Emotional Reactions and ressed lower intensity negative emotions as mothers expressed or shared positive affect with their toddlers in previous interval. way or when they became angry them selves, toddlers tended to express higher intensity of negative emotions. Appropriate Mind - Related Comments were not associated with Expression of Negative Emotion while Non - Attuned Mind - Related Comments were positively related to Expression of Negative Em otion as expected; suggesting that higher use of lagged Non - Attuned Mind - Related Comments was associated with higher intensity of Expression of Negative Emotion as expected. Table 15 Correlations for Lagged Maternal Regulatory Attempts, Appropriate and Non - Attuned Mind - Related Comments with Expression of Negative Emotion (n=1437) Variable Name ENE - O Verbal Comfort .14 ** Physical Comfort .06 * Initiated Distraction - .19 ** Joined Distraction - .05 * Verbal Orientation to Delay - 0.04 Positive Comma nds 0.01 Negative Commands .15 ** Punitive Reactions .12 ** Minimizing .18 ** Physical Restraint .28 ** Positive Emotional Reactions - .32 ** Negative Emotional Reactions .20 ** Appropriate Mind Related Comments - 0.02 Non - Attuned Mind Related Comments .0 7 ** Note. ENE - O = Overall Expression of Negative Emotion (average of ENE - I and ENE - P). * p < .05, one - tailed. ** p < .01, one - tailed. p < .10, one - tailed. 78 Lagged Toddlers Regulatory Strategies and Expression of Negative Emotion. Correlations for lagged I ndependent Regulatory Strategies 18 and lagged Dependent Regulatory Strategies with ENE - O are presented (Table 16) 19 . Similar to correlations for current - interval Independent Regulatory Strategies, correlations for lagged Independent Regulatory Strategies wi th ENE - O were small but significant (r = - .31 to .10). Lagged Toddler Initiated Verbal Keys - O. positively associated with ENE - O suggesting that as toddlers vocalized desire states, they tended to express more intense levels of negative emotion 10 seconds later. These associations suggest that verbalization of desire words that are not directed toward mother may facilitate expression of negative emotion. All other correlations were in expected negative directions. Strongest associations were observed for lagged Toddler Initiated Non - Verbal Distraction and ENE - O. As toddlers initiate d engagement in non - verbal distracting behaviors such as play or looking away from the keys, they tended to express milder levels of negative emotion 10 seconds later. Similar to correlations for current - interval dependent regulatory strategies, correlati ons for lagged Dependent Regulatory Strategies with ENE - O were small but significant (r= - .07 to - .19). Lagged Toddler Verbal Distraction - initiated singing while looking at mom), lagged Toddler Verbal Desir e - Bids to Mom (e.g., Self - 18 No correlations were calculated for Toddler Initiated Verbal Self - Control . All observed Toddler Verbal Self - Control was coded under Dependent Regulatory Strategies (i.e., Toddler Verbal Self - Control - Bids to Mom ) indicating that any time toddlers initiated verbal self - ng at mother or in response to mother. 19 See Appendix F for cu rrent interval correlations for Toddlers Regulatory Strategies with Expression of Negative Emotion (Table F2 ). 79 up to mother, hugging mother, leaning on - O. All significant associations were in expected negative direction. Strongest associations were observed for lagged Toddler Joined Non - Verbal Distraction and ENE - O. As toddlers joined their nitiated non - verbal distractions (e.g., Mother points to a direction away from keys and child looks where she is pointing), they expressed lower levels of ENE - O 10 seconds later. Table 16 Correlations for Lagged Toddler Regulatory Strategies with Express ion of Negative Emotion (n = 1437) Variable Name ENE - O Independent Regulatory Strategies (IRS) Initiated Verbal Distraction - .04 * Initiated Non - Verbal Distraction - .31 ** Initiated Verbal Keys - .03 Initiated Verbal Self - Control - Initiated Verbal D esire .10 ** Self - Comfort - .11 ** Dependent Regulatory Strategies (DRS) Verbal Distraction - Bids to Mom - .02 Joined Verbal Distraction - .12 ** Joined Non - Verbal Distraction - .19 ** Verbal Keys - Bids to Mom - .07 ** Verbal Self - Control - Bids to Mom - .03 Verb al Desire - Bids to Mom .02 Physical Comfort Seeking .00 Note. ENE - O = Overall Expression of Negative Emotion (average of ENE - I and ENE - P). * p < .05, one - tailed. ** p < .01, one - tailed. p < .10, one - tailed. Lagged Maternal Variables and Delay of Gratif ication. Next, correlations for lagged Maternal Regulatory Attempts, lagged Appropriate, and lagged Non - Attuned Mind - Related Comments with Delay of Gratification operationalized as amount of Overall Touch are presented (Table 17) 20 . Similar to correlations for current - interval Maternal Regulatory Attempts with Overall Touch, correlations for lagged Maternal Regulatory Attempts with Overall Touch were small but significant (r = - .06 to .14). Lagged Verbal Comfort, lagged Maternal Joined 20 See A ppendix F for current interval correlations for Maternal Regulatory Attempts, Appropriate and Non - Attuned Mind - Related Comments with O verall Touch (Table F3 ). 80 Distraction, lagged Ve rbal Orientation to Delay, lagged Positive Commands, lagged Punitive Reactions, and lagged Minimizing statements were not significantly related to Overall Touch. Unexpectedly, lagged Positive Emotional Reactions was positively associated with Overall Touch touch or attempt to touch the keys increased. All other correlations were in the expected direction. The strongest negative associations for both current and prev ious interval predictors were between lagged Initiated Distraction and Overall Touch, while the strongest positive associations were between lagged Physical Restraint and lagged Negative Emotional Reactions with Overall Touch. As mothers initiated distract mock touch the keys increased as expected. Lagged Appropriate and Lagged Non - Attuned Mind - Related Comments were not correlated with Overall Touch. Table 17 Correlations fo r Lagged Maternal Regulatory Attempts, Appropriate and Non - Attuned Mind - Related Comments with Overall Touch (n=1437) Variable Name Overall Touch Verbal Comfort - 0.02 Physical Comfort - .06 ** Initiated Distraction - .09 ** Joined Distraction - 0.03 Ve rbal Orientation to Delay 0.04 Positive Commands 0.01 Negative Commands .11 ** Punitive Reactions 0.01 Minimizing 0 Physical Restraint .14 ** Positive Emotional Reactions .06 ** Negative Emotional Reactions .13 ** 81 Table 17 (c A ppropriate Mind Related Comments - 0.03 Non - Attuned Mind Related Comments - 0.03 Note. Overall Touch = sum of Touch and Attempted Touch. * p < .05, one - tailed. ** p < .01, one - tailed. p < .10, one - tailed. Lagged Toddler Regulatory Strategies and Delay of Gratification . Correlations for lagged Independent Regulatory Strategies and lagged Dependent Regulatory Strategies with Delay of Gratification operationalized as amount of Overall Touch are presented (Table 16) 21 . Similar to correlations for current - interval Independen t Regulatory Strategies, correlations for lagged independent regulatory strategies with Overall Touch were small but significant (r = - .11 Overall Touch. Unexpect positively associated with Overall Touch suggesting that as toddlers vocalized desire states, they tended to touch or attempted to touch the keys 10 seconds later. These associations sugg est that verbalization of desire words that are not directed toward mother may impede ability to wait. All other correlations were in expected negative directions. Strongest associations were observed for lagged Toddler Self - Comfort and Overall Touch. As e xpected, as toddlers displayed more lagged Self - Comfort (e.g., thumb sucking), they tended to display less Overall Touch. Correlations for lagged dependent regulatory strategies with Overall Touch were also small but significant (r= - .06 to - .04). Lagged Toddler Joined Verbal Distraction, lagged Verbal Keys - Bids to Mom, and lagged Verbal Self - Control - Bids to Mom were not significantly ard mother or in response to mother was positively and marginally associated with Overall Touch. All other correlations were in expected negative directions. 21 See Appendix F for current interval correlations for Toddlers Regulatory Strategies with Overall Touch (Table F4 ). 82 Strongest associations were observed for lagged Toddler Physical Comfort Seeking. As expected, as toddlers displayed more lagged Physical Comfort Seeking (e.g. leaning in to mom, reaching up for hug), they tended to display less Overall Touch. Table 18 Correlations for Lagged Toddler Regulatory Strategies with Overall Touch (n = 1437) Variable Name Overall Touch Independent Regulatory Strategies (IRS) Initiated Verbal Distraction - .03 Initiated Non - Verbal Distraction - .10 ** Initiated Verbal Keys .00 Initiated Verbal Self - Control - Initiated Verbal Desire .05 * Self - Comfort - .11 ** Dependent R egulatory Strategies (DRS) Verbal Distraction - Bids to Mom - .05 * Joined Verbal Distraction - .03 Joined Non - Verbal Distraction - .05 * Verbal Keys - Bids to Mom .00 Verbal Self - Control - Bids to Mom - .01 Verbal Desire - Bids to Mom .04 Physical Comfort See king - .06 ** Note. Overall Touch = sum of Touch and Attempted Touch. * p < .05, one - tailed. ** p < .01, one - tailed. p < .10, one - tailed. Expression of Negative Emotion, Delay of Gratification, Child Gender, and Expressive Language. - O and Ove rall Touch were not significantly correlated 22 . - O, it was not correlated with Overall Touch. On average, boys displayed higher levels of ENE - expressive language was not correlated with ENE - O, it was correlated with Overall Touch. On average, toddlers with reportedly higher scores on productive vocabulary displayed lower amounts of Overall Touch. 22 On average as toddlers displayed more Touch (e.g., touching the keys), ENE - O ( r = - .07, p < .01). However, as toddlers displayed more Attempted Touch (e.g., reach for keys but were restrained), they tended to express higher ENE - O ( r = .06, p < .01). These results suggest that touching the keys is associated with less intense and shorter durations of expressed neg ative emotion, while attempting to touch the keys but being restrained is associated with more intense and longer duration of expressed negative emotion. These results point to interrelations between indicators of self - regulation not examined in the curren t study. 83 Table 19 Intercorrelations for Expression of Negative Emotion with Overall Touch, Toddler Gender, and Vocabulary Production Variable Names 1 2 3 4 1. Overall Expression of Negative Emotion - 2. Overall Touch .03 - 3. Gender .14 ** - .01 - 4. Vocabulary Production .01 - .09 ** - .12 ** - Note. Overall Expression of N egative Emotion = a verage of ENE - I and ENE - P; Overall Touch = sum of Touch and Attempted Touch. 1 = male. * p < .05, one - tailed. ** p < .01, one - tailed. p < .10, one - tailed. Multilevel Analyses This section describes results of all multilevel models util ized to answer study questions. First, Intraclass Correlations (ICCs) 23 were calculated for ENE - O and Overall Touch in the Intercept - Only Models. About 66% of variability in ENE - O, and 20% of variability in Overall Touch, was associated with differences be tween mother - correlations in error terms and violations of assumptions of independence of error underlying hierarchical linear modeling, thus, analyses required grouping variables at more than 1 level (Tabachnick & Fidell , 2007). Using restricted maximum likelihood (REML), separate two - level hierarchical models were estimated for a total of 12 models examining the effects of level - 1 lagged Maternal Regulatory Attempts, a total of 2 models examining the effects of level - 1 lagged Mind - Related Comments (Appropriate and Non - Attuned), and a total of 13 models examining the effects of level - - O and Overall Touch. Models for both outcomes were estimated separately. As previously menti oned, multilevel analyses are well suited to address violations to independence of errors that undermine accuracy and power, especially when there is large number of level - 2 units (Snijders, 2005). To ensure model accuracy, models without random slope vari ance components (IV - 23 See Appendix G for how independence of errors was determined for each outcome. 84 dyads) were tested against models with random slope variance components ( U 1j : deviation of IV - DV slope in dyad j from average slope) for all predicto were used to compare models with fixed slopes against models in which slopes for significant Maternal Regulatory Attempts, Appropriate and Non - Attuned Mind - Related Comments, and allowed to vary among dyads. The deviance test can be used when both models have the same fixed effects and only differ in estimation of random components (Snijders & Bosker, 2012). P - values less than .05 suggest that the full models (allowing slopes to va ry) predicted models better than fixing slopes to be the same across all dyads (Tabachnick & Fidell, 2007). Thus, each predictor was specified and estimated in two separate models for each outcome. Model 1 was specified random intercepts (outcomes varied o ver dyads when all predictors were zero), and fixed slopes (association between level 1 IVs and DV is fixed for all dyads). Model 2 was specified with random intercepts and random slopes in which variance components for slopes were not constrained. Results of the more parsimonious models (Model 1) are reported for all predictors unless the random - intercept random slope model specification (Model 2) improved model fit as indicated by significant p - values associated with chi - squared difference tests . In other words, in cases where freeing the slopes resulted in better model fit results of less parsimonious model (Model 2) were reported. In cases where freeing the slopes did not add significant misfit ( .05), but freeing the slopes resulted in significant main effects (Model 2) where there were not main effects before (Model 1), the results of Model 2 were reported. 24 24 For decision tree refer to Supplementary Summary C . 85 Covariates included lagged maternal verbosity (i.e., total comments), linear and quadratic elapsed time, toddl 25 No interactions were hypothesized either within or between dyads to explain variations in slopes for predictors and outcomes. Therefore, no level - 2 explanatory variables were modeled to explain differences in slopes between IV and DV (e.g., interactions between child gender and maternal verbal comfort to explain differences in slopes between verbal comfort and outcomes). Results of all analyses are reported using final estimation of fixed effects with robust st andard errors 26 , which are more appropriate for data with large level - 2 units (Raudenbush et al., 2011), 134 units in the current study. Results are presented in order of research questions (Table 20). Table 20 Overview of Research Questions: Predictor s X Outcomes Outcome 1 Outcome 2 ENE - O Overall Touch Predictors Positive associations hypothesized Negative associations hypothesized Positive associations hypothesized Negative associations hypothesized MRA/MM Question 1 A Question 1 B Question 3 A Question 3 B TRS - Independent - Question 2 A - Question 4 A TRS - Dependent - Question 2 B - Question 4 B Note. MRA = Maternal Regulatory Attempts, MM=Mind - ENE - O = Overall Expression of Negative Emotion calculated as mean Intensity and Predominance in 10s, Delay of Gratification is indicated by Overall Touch (Sum of Touch and Attempted Touch in 10s). Higher Overall Touch scores indicate lower delay of gratification or ability to wait. Research Que stion 1A: The Random Intercept Model Predicting Expression of Negative Emotion from Lagged 27 Maternal Regulatory Attempts and Mind - Mindedness . Effects of maternal variables in Question 1A were examined and are presented below. A summary of all model results is provided in Table 21 28 . Model results for each significant 25 For details on model specification with covariates see Appendix H. 26 Coefficients are the same for both rob ust and non - robust estimates. Robust standard errors are less biased in presence of violations of assumptions related to distributional normality of error terms. 27 val. 28 Also provided in on page 4 of Supplemental Materials 86 predictor are then presented in the following pages (Tables 22 - 25). Model results for non - significant predictors (Phy - Comf, M - JNT - Dis, M - VO2D, and AMM) are presented in Appendix I (Tables I1 - I4) . 87 Table 21 Question 1A. Do Lagged Maternal Regulatory Attempts and Lagged Mind - Mindedness Negative Emotion (Continuous: 0 - 3) Model 1 - Random Intercept Fixed Slope Model 2 - Random Intercept Random Slope Lagged Maternal Variables in Question 1A Variable Label Does X significantly predict Y? Is directionality as hypothesized? Does X predict Y? Does X - Y slope vary over dyads? Does freeing slope significantly reduce misfit? Is directionality as hypothesized? Verba l Comfort V - Comf Yes No (Pos) Yes No No No (Pos) Physical Comfort Phy - Comf No No (Pos) No No No No (Pos) Initiated Distraction M - INI - Dis Yes Yes (Neg) Yes Yes Yes Yes (Neg) Joined Distraction M - JNT - Dis No Yes (Neg) No No No Yes (Neg) Verbal Or ientation to Delay M - VO2D No Yes (Neg) No Yes Yes Yes (Neg) Positive Commands Pos - Comd No No (Pos) Yes No Yes No (Pos) Positive Emotional Reactions Pos - Aff Yes Yes (Neg) Yes Yes Yes Yes (Neg) Appropriate MRC AMM No Yes (Neg) No No No Yes (Neg) Note . MRC = Mind - Related Comments; 1A predictors were hypothesized to be negatively associated with outcome; Bold font indicates largest slope coefficient. Italics font indicates X became significant predictor in model 2. * = p - value for slope coefficie nts with Non - Robust standard errors were < .05. All results report p - values for models estimated with robust standard errors. p < .10. Non - significant estimates and unexpected directionality are in red. 88 Lagged Maternal Verbal Comfort. A two - le vel hierarchical model examined the effects - O) while controlling for effects of linear elapsed time, quadratic elapsed time, lagged maternal verbosity, and toddler gender. T he model was specified with the following equation: ENE - O ij 00 Gender (Gender j ) V - Comf (V - Comf i - 1j ) + ETL (ET ij ET 2 (ET 2 ij Verb (Verb i - 1j ) + u 0j + e ij Results of final estimated model revealed that on average, levels of ENE - O was sign ificantly associated with lagged maternal Verbal Comfort as hypothesized but were in unexpected direction (Table 22). For every unit increase in lagged maternal Verbal Comfort negative emotion increased by 0.09 on a scale of 0 - was 0.31 units higher for boys after accounting for effects for all predictors. Additionally, the current model was compared against a model (Mo del 2) in which slopes between lagged maternal Verbal Comfort and ENE - O were allowed to vary between dyads. However results of deviance tests ( = chi squared difference) indicated that allowing the relationship between lagged maternal Verbal Comf ort and ENE - O to vary among dyads would not significantly help to explain more variance in ENE - O between toddlers ( = .26, df = 2, p > 0.05). Additionally, the p - value associated with random slope was not significant suggesting that the associat ion between lagged maternal Verbal Comfort and ENE - O does not vary between dyads. In the random - intercept random - slope model (Model 2), the slope coefficient for lagged maternal Verbal Comfort was 0.08 ( p = 0.02). The results of more parsimonious model are presented in Table 22. 89 Table 22 Effects of Lagged Maternal Verbal Comfort on Overall Expression of Negative Emotion Fixed Effects Coefficient S.E Intercept 0.51 *** 0.15 Gender 0.31 * 0.15 Verbal Comfort 0.09 ** 0.03 Linear ET 0.00 0.00 Quadratic E T - 0.00 0.00 Verbosity 0.04 *** 0.01 Random Effects Variance Components SD u 0j 0.72 *** 0.85 e ij 0.32 0.56 Note. Elapsed Time (ET) is in 10s intervals. * p < .05, ** p < .01, *** p < .001, p < .10 Lagged Maternal Initiated Distraction . A two - level hier archical model examined the (ENE - O) while controlling for effects of linear elapsed time, quadratic elapsed time, and toddler gender. The model was specified with the following equation: ENE - O ij 00 Gender (Gender j ) M - INI - Dis ( M - INI - Dis - 1j ) + ETL (ET ij ET 2 (ET 2 ij ) + u 0j + e ij Results of final estimated model revealed that on average, levels of ENE - O were marginally associated with lagged maternal Initiated Distraction as hypothesized (Table 23). For - 3. average overall expression of negative emotion was 0.31 units higher for boys (scale of 0 - 3) after accounting for effects for all predictors. Additionally, the current model was compared against a model (model 2) in which slopes between lagged maternal I nitiated Distraction and ENE - O were allowed to vary between dyads. Results of deviance tests ( = chi squared difference) indicated that allowing the relationship between lagged maternal Initiated Distraction and ENE - O to vary among dyads would si gnificantly improve model fit = 15.98, df = 2, p < 0.001. Indeed, when the association 90 between lagged maternal Initiated Distraction and ENE - O was allowed to vary between dyads, the p - value associated with the random slope coefficient was signif icant (p <.05), suggesting that dyads varied in the association between lagged maternal Initiated Distraction and ENE - O. In the random - intercept random - slope model (Model 2), for every unit increase in lagged maternal Initiated Distraction , ll expression of negative emotion decreased by 0.09 on a scale of 0 - 3. Table 23 Effects of Lagged Maternal Initiated Distraction on Overall Expression of Negative Emotion Model 1 Model 2 Fixed Effects Coefficient S.E Coefficient S.E Intercept 0.63 *** 0.15 0.63 *** 0.15 Gender 0.31 * 0.15 0.30 * 0.15 Initiated Distraction - 0.07 0.04 - 0.09 * Linear ET 0.00 0.00 0.00 0.00 Quadratic ET - 0.00 0.00 - 0.00 0.00 Random Effects Variance Components SD Variance Components SD u 0j 0.71 *** 0.84 0.73 *** 0.86 u M - INI - Dis , j - - 0.12 *** 0.35 e ij 0.32 0.57 0.3 0.55 Deviance 2914.88 2898.91 Note. Elapsed Time (ET) is in 10s intervals. * p < .05, ** p < .01, *** p < .001, p < .10 Lagged Maternal Positive Commands. A two - level hierarchical model examined the effec (ENE - O) while controlling for effects of linear elapsed time, quadratic elapsed time, lagged maternal verbosity, and toddler gender. The model was specified with the follow ing equation: ENE - O ij 00 Gender (Gender j Pos - Comd ( Pos - Comd i - 1j ) + ETL (ET ij ET 2 (ET 2 ij Verb (Verb i - 1j ) + u 0j + e ij - O (Table 24). of negative emotion was 0.31 units higher for boys (scale of 0 - 3) after accounting for effects for all predictors. 91 Additionally, the current model was compared against a model (Model 2) in which slopes between lagged m aternal Positive Commands and ENE - O were allowed to vary between dyads. Results of deviance tests ( = chi squared difference) indicated that allowing the relationship between lagged maternal Positive Commands and ENE - O to vary among dyads would significantly improve model fit = 12.07, df = 2, p < 0.01. However, the p - value associated with random slope was not significant suggesting that the association between lagged maternal Positive Commands and ENE - O does not vary between dyads. Additionally, in the random - intercept random - slope model (Model 2), for every unit increase in lagged ma ternal Positive Commands , scale of 0 - 3 ( p = 0.08) , but not in the expected direction. Table 24 Effects of Lagged Maternal Positive Commands on Overall Expression of Negati ve Emotion Model 1 Model 2 Fixed Effects Coefficient S.E Coefficient S.E Intercept 0.51 *** 0.15 0.53 0.15 Gender 0.31 * 0.15 0.26 0.15 Positive Commands 0.03 0.03 0.03 0.02 Linear ET 0.00 0.00 0.00 0.00 Quadratic ET - 0.00 0.00 - 0.00 0.00 Verbosi ty 0.04 *** 0.01 0.04 *** 0.01 Random Effects Variance Components SD Variance Components SD u 0j 0.73 *** 0.85 0.70 *** 0.84 u Pos - Comd , j - - 0.01 0.09 e ij 0.32 0.56 0.31 0.56 Deviance 2900.18 2888.11 Note. Elapsed Time (ET) is in 10s intervals. *p < . 05, ** p < .01, *** p < .001, p < .10 Lagged Maternal Positive Emotional Reactions. A two - level hierarchical model Negative Emotion (ENE - O) while controlling for effects of linear elapsed time, quadratic elapsed time, and toddler gender. The model was specified with the following equation: 92 ENE - O ij 00 Gender (Gender j ) Pos - Aff (Pos - Aff i - 1j ) + ETL (ET ij ET 2 (ET 2 ij ) + u 0j + e ij Results of final estimated model revealed that on average, levels of ENE - O was significantly associated with lagged maternal Positive Emotional Reactions as hy pothesized in the expected direction (Table 25). For every unit increase in lagged maternal Positive Emotional decreased by 0.19 on a scale of 0 - erage overall expression of negative emotion was 0.31 units higher for boys after accounting for effects for all predictors. Additionally, the current model was compared against a model (Model 2) in which slopes between lagged maternal Positive Emotional Reactions and ENE - O were allowed to vary between dyads. Results of deviance tests ( = chi squared difference) indicated that allowing the relationship between lagged maternal Positive Emotional Reactions and ENE - O to vary among dyads would signif icantly improve model fit = 20.30, df = 2, p < 0.001). Indeed, when the association between lagged maternal Positive Emotional Reactions and ENE - O was allowed to vary between dyads, the p - value associated with the random slope coefficient was s ignificant ( p <0.001), suggesting that dyads varied in the association between lagged maternal Positive Emotional Reactions and ENE - O. In the random - intercept random - slope model (Model 2), for every unit increase in lagged maternal Positive Emotional React ions , expression of negative emotion decreased by 0.25 on a scale of 0 - 3. Table 25 Effects of Lagged Maternal Positive Emotional Reactions on Overall Expression of Negative Emotion Model 1 Model 2 Fixed Effects Coefficient S.E Coeff icient S.E Intercept 0.74 *** 0.15 0.80 *** 0.15 Gender 0.30 * 0.15 0.21 0.13 Positive Emotional Reactions - 0.19 *** 0.05 - 0.25 *** 0.05 Linear ET 0.00 0.00 0.00 0.00 93 Table 25 ( c Quadratic ET - 0.00 0.00 - 0.00 0.00 Random Effects Variance Compone nts SD Variance Components SD u 0j 0.67 *** 0.82 0.75 *** 0.87 u Pos - Aff , j - - 0.08 *** 0.29 e ij 0.32 0.56 0.31 0.56 Deviance 2893.13 Note. Elapsed Time (ET) is in 10s intervals. * p < .05, ** p < .01, *** p < .001, p < .10 Question 1A Summary of Resu lts. In summary, on average, toddler boys expressed higher overall levels of intensity and predominance of negative emotion (i.e., ENE - O) across the two - minute task, after each maternal regulatory attempt ( Q uestion1 A ) from a previous 10s interval was accou negative emotion significantly differed between 10s intervals (within dyad variance) based on the amount of verbal comfort, positive commands (marginal), initiated distractions, and po sitive emotional reactions their mothers displayed in a previous interval. While, contrary to my hypotheses, higher lagged verbal comfort and lagged positive commands were associated with higher levels of overall expression of negative emotion, toddlers e xpressed lower levels of overall negative emotion as mothers initiated distractions and displayed positive emotional reactions in a previous interval as expected. Unexpectedly, the amount of lagged physical comfort (unexpected positive direction), and lagg ed verbal orientation to delay, lagged joint distraction, and lagged appropriate mind - related comments (expected negative directionality) in previous interval were not associated with overall expression of negative emotion in a current interval. Among all significant predictors in Question 1A, lagged maternal positive emotional reactions were most strongly associated with lower levels of overall expression of negative emotion (random - intercept random - slope), while lagged maternal Initiated Distraction was least strongly associated with lower levels of overall expression of negative emotion. 94 Research Question 1B: The Random Intercept Model Predicting Expression of Negative Emotion from Lagged Maternal Regulatory Attempts and Mind - Mindedness . Effects of mate rnal variables in Question 1B were examined and are presented below. A summary of all model results in provided in Table 26 29 . Model results for each predictor are then presented in the following pages (Tables 27 - 30). Model results for non - significant predi ctors (Neg - Comd and N MM) are presented in Appendix I (Tables I5 - I6). Table 26 Question 1B. Do Lagged Maternal Regulatory Attempts and Lagged Mind - Mindedness Negative Emotion (Continuous: 0 - 3) Model 1 - Random In tercept Fixed Slope Model 2 - Random Intercept Random Slope Lagged Maternal Variables in Question 1B Variable Label Does X significantly predict Y? Is directionality as hypothesized? Does X predict Y? Does X - Y slope vary over dyads? Does freeing slope sig nificantly reduce misfit? Is directionality as hypothesized? Negative Commands Neg - Comd No Yes (Pos) No Yes No No (Neg) Punitive Reactions Pun Yes Yes (Pos) No No Yes (Pos) Minimizing Min No Yes (Pos) Yes Yes No Yes (Pos) Physical Rest raint Phy - RST Yes Yes (Pos) Yes Yes Yes Yes (Pos) Negative Emotional Reactions Neg - Aff Yes Yes (Pos) Yes Yes No Yes (Pos) Non - Attuned MRC NMM No Yes (Pos) No Yes No Yes (Pos) Note . MRC = Mind - Related Comments; 1B predictors were hypothesized to b e positively associated with outcome; Bold font indicates largest slope coefficient. Italics font indicates X became significant predictor in model 2. * = p - value for slope coefficients with Non - Robust standard errors were < .05. All results report p - value s for models estimated with robust standard errors. p < .10. Non - significant estimates and unexpected directionality are in red. 29 Also provided in on page 5 of Supplemental Materials 95 Lagged Maternal Punitive Reactions. 30 A two - level hierarchical model examined the effects of lagged maternal Punitive (ENE - O) while controlling for effects of linear elapsed time, quadratic elapsed time, and toddler gender. The model was specified with the following equation: ENE - O ij 00 Gender (Gender j ) M - Pu n ( M - Pun i - 1j ) + ETL (ET ij ET 2 (ET 2 ij ) + u 0j + e ij Results of final estimated model revealed that on average, levels of ENE - O was significantly associated with lagged maternal Punitive Reactions as hypothesized but were in unexpected direction (Tabl e 27). For every unit increase in lagged maternal Punitive Reactions increased by 0.12 on a scale of 0 - ion was 0.31 units higher for boys after accounting for effects for all predictors. Additionally, the current model was compared against a model (Model 2) in which slopes between lagged maternal Punitive Reactions and ENE - O were allowed to vary between dy ads. However, results of deviance tests ( = chi squared difference) indicated that allowing the relationship between lagged maternal Punitive Reactions and ENE - O to vary among dyads would not significantly help to explain more variance in ENE - O b etween toddlers ( = 0.76, df = 2, p > 0.05). Additionally, the p - value associated with random slope was not significant suggesting that the association between lagged maternal Punitive Reactions and ENE - O does not vary between dyads. In the rando m - intercept random - slope model (Model 2), for every unit increase in lagged maternal Punitive Reactions , 30 Although Punitive Reactions contained mostly punitive verbal statements, they also included non - verbal punitive reactions such as slapping Reactions as predictor. 96 marginally increased by 0.09 on a scale of 0 - 3 ( p = 0.05) . The results of more parsimonious model are presented in presented in Table 27. Table 27 Effects of Lagged Punitive Reactions on Overall Expression of Negative Emotion Fixed Effects Coefficient S.E Intercept 0.63 *** 0.15 Gender 0.31 * 0.15 Punitive Reactions 0.12 * 0.05 Linear ET 0.00 0.00 Qua dratic ET - 0.00 0.00 Random Effects Variance Components SD u 0j 0.72 *** 0.85 e ij 0.32 0.57 Note. Elapsed Time (ET) is in 10s intervals. * p < .05, ** p < .01, *** p < .001, p < .10 Lagged Maternal Minimizing. A two - level hierarchical model examined the effects of - O) while controlling for effects of linear elapsed time, quadratic elapsed time, lagged maternal verbosity, and toddler gender. The model was specified with the followi ng equation: ENE - O ij 00 Gender (Gender j ) Min (Min i - 1j ) + ETL (ET ij ET 2 (ET 2 ij Verb (Verb i - 1j ) + u 0j + e ij not associated with ENE - O (Table 28). Toddlers was 0.31 units higher for boys (scale of 0 - 3) after accounting for effects for all predictors associations. Additionally, the current model was compared against a model (Model 2) in which slopes between la gged maternal Minimizing and ENE - O were allowed to vary between dyads. However, results of deviance tests ( = chi squared difference) indicated that allowing the relationship between lagged maternal Minimizing and ENE - O to vary among dyads would not significantly help to explain more variance in ENE - O between toddlers ( = 2.31, df = 2, p > 9 7 0. 05). However, when the association between lagged maternal Minimizing and ENE - O was allowed to vary between dyads, the p - value associated with the random slope coefficient was significant ( p < .05), suggesting that dyads tended to vary in the association b etween lagged maternal Minimizing and ENE - O. Additionally, in the random - intercept random - slope model (Model 2), for every unit increase in lagged maternal Minimizing , negative emotion increased by 0.33 on a scale of 0 - 3, as hypothesized in the expected direction. Table 28 Effects of Lagged Minimizing on Overall Expression of Negative Emotion Model 1 Model 2 Fixed Effects Coefficient S.E Coefficient S.E Intercept 0.51 *** 0.15 .52 *** 0.15 Gender 0.31 * 0.15 0.27 0.15 M inimizing 0.10 0.09 0.33 *** 0.09 Linear ET 0.00 0.00 0.00 0.00 Quadratic ET - 0.00 0.00 - 0.00 0.00 Verbosity 0.04 *** 0.01 0.04 *** 0.01 Random Effects Variance Components SD Variance Components SD u 0j 0.72 *** 0.85 0.72 *** 0.85 u Min ,j - - 0.05 * 0.23 e i j 0.32 0.56 0.31 0.56 Deviance 2898.40 2896.09 Note. Elapsed Time (ET) is in 10s intervals. * p < .05, ** p < .01, *** p < .001, p < .10 Lagged Maternal Physical Restraint. A two - level hierarchical model examined the effects of lagged maternal Physica (ENE - O) while controlling for effects of linear elapsed time, quadratic elapsed time, and toddler gender. The model was specified with the following equation: ENE - O ij 00 Gender (Gender j ) Phy - RST (Phy - RST i - 1j ) + ETL (ET ij ET 2 (ET 2 ij ) + u 0j + e ij Results of final estimated model revealed that on average, levels of ENE - O was significantly associated with lagged maternal Physical Restraint as hypothesize d in the expected direction (Table 29). For every unit increase in lagged maternal Physical Restraint (e.g., 2 = 98 High - firmly holding child in lap, 1 = Low - gently pulls arm back, 0 = absence of any restraining ative emotion increased by 0.18 on a scale of 0 - 3. accounting for effects for all predictors, although this was a marginal association. Additionally, the current model was compared against a model (Model 2) in which slopes between lagged maternal Physical Restraint and ENE - O were allowed to vary between dyads. Results of deviance tests ( = chi squared difference) indicated that allowing the relationship between lagged maternal Physical Restraint and ENE - O to vary among dyads would significantly improve model fit = 55.30, df = 2, p < 0.001. Indeed, when the association between lagged maternal Physical Restraint and ENE - O was allowed to vary betw een dyads, the p - value associated with the random slope coefficient was significant ( p <0.001), suggesting that dyads varied in the association between lagged maternal Physical Restraint and ENE - O. In the random - intercept random - slope model (Model 2), for every unit increase in lagged maternal Physical Restraint , of 0 - 3. Table 29 Effects of Lagged Maternal Physical Restraint on Overall Expression of Negative Emotion Model 1 Mo del 2 Fixed Effects Coefficient S.E Coefficient S.E Intercept 0.45 ** 0.15 0.49 ** 0.15 Gender 0.28 0.14 0.26 0.14 Physical Restraint 0.18 *** 0.04 0.16 *** 0.04 Linear ET 0.00 0.00 0.00 0.00 Quadratic ET - 0.00 0.00 - 0.00 0.00 Random Effects Variance Components SD Variance Components SD u 0j 0.66 0.81 0.65 *** 0.81 u Phy - RST , j - - 0.09 0.30 e ij 0.31 0.60 0.28 0.53 Deviance 2871.39 2816.08 Note. Elapsed Time (ET) is in 10s intervals. * p < .05, ** p < .01, *** p < .001, p < .10 99 Lagged Maternal Neg ative Emotional Reactions. A two - level hierarchical model of Negative Emotion (ENE - O) while controlling for effects of linear elapsed time, quadratic elapsed time, and toddler gender. The model was specified with the following equation: ENE - O ij 00 Gender (Gender j ) Neg - Aff (Neg - Aff i - 1j ) + ETL (ET ij ET 2 (ET 2 ij ) + u 0j + e ij Results of final estimated model revealed that on average, levels of ENE - O were marginally associated with lagged maternal Negative Emotional Reactions as hyp othesized in the expected direction (Table 30). For every unit increase in lagged maternal Negative Emotional emotion tended to increase by 0.22 on a scale of 0 - 3. negative emotion was 0.31 units higher for boys after accounting for effects for all predictors. Additionally, the current model was compared against a model (Model 2) in which slopes between lagged maternal Negativ e Emotional Reactions and ENE - O were allowed to vary between dyads. However, results of deviance tests ( = chi squared difference) indicated that allowing the relationship between lagged maternal Negative Emotional Reactions and ENE - O to vary amo ng dyads would not significantly help to explain more variance in ENE - O between toddlers ( = 3.57, df = 2, p > 0.05). However, when the association between lagged maternal Negative Emotional Reactions and ENE - O was allowed to vary between dyads, the p - value associated with the random slope coefficient was significant (p <.01), suggesting that dyads varied in the association between lagged maternal Negative Emotional Reactions and ENE - O. In the random - intercept random - slope model (Model 2), for eve ry unit increase in lagged maternal Negative Emotional Reactions , 100 0.69 on a scale of 0 - 3 ( p < .001) . The results of more parsimonious model are presented in Table 36. Table 30 Effects of Lagged Negative Emotional Reactions on Overall Expression of Negative Emotion Fixed Effects Coefficient S.E Intercept 0.63 *** 0.15 Gender 0.31 * 0.15 Negative Emotional Reactions 0.22 0.14 Linear ET 0.00 0.00 Quadratic ET - 0.00 0.00 Random Effects Varianc e Components SD u 0j 0.71 *** 0.84 e ij 0.32 0.57 Note. Elapsed Time (ET) is in 10s intervals. * p < .05, ** p < .01, *** p < .001, p < .10 Question 1B Summary of Results . In summary, on average, toddler boys expressed higher overall levels of intensity a nd predominance of negative emotion (i.e., ENE - O) across the two - minute task, after each maternal regulatory attempt (Question 1B) from a previous 10s interval was accounted for in separate multilevel models. on significantly differed between 10s intervals (within dyad variance) based on the amount of punitive reactions, minimizing statements, physical restraint, and negative emotional reactions their mothers displayed in a previous interval. All associations w ere in expected directions (positive). As mothers displayed more punitive reactions, more minimizing statements, higher levels of physical restraint, and negative emotions for longer duration. Unexpectedly, the amount of negative commands and non - attuned mind - related comments in previous interval was not associated with overall expression of negative emotion in a current interval but was in the expected direction. Among all significant predictors in Question 1B, lagged maternal negative emotional reactions were most strongly associated with higher levels of overall expression of negative 101 emotion (random - intercept random - slope), while lagged maternal punitive react ions were least strongly associated with higher levels of overall expression of negative emotion. In fact, lagged maternal negative emotional reactions were more strongly related to higher overall expression of negative emotion ( = 0.69, p <.001) than positive emotional reactions (Question 1A) were related to lower overall expression of negative emotions ( = - 0.25, p <.001), especially after accounting for the significant amount of variation in each of these IV - DV associations between dyads. Interestingly, lagged maternal minimizing statements were moderately associated with higher overall expression of negative emotion only if this association was allowed to vary between dyads (random - intercept random - slope) model. Furthermore , exploratory interactional analyses (not hypothesized) revealed that toddler gender played a moderating role in the association between lagged minimizing statements and overall expression of negative emotion such that the main effect of minimizing on over all expression of negative emotion ( = .52, p < 0.001) was 0.39 units (scale of 0 - 3) lower for toddler boys compared to toddler girls. No statistically significant associations were found for linear or quadratic elapsed time but they were retained in the models (both Questions 1 A and 1 B ) since excluding them did not change significant coefficients and the results of the curve estimations revealed significant linear and quadratic trajectories of change for ENE - O across the task. As evident by the statistic ally significant residuals ( u oj ), indicating significant remaining unexplained variance in average ENE - O between toddlers when all predictors were taken into account, there is room to improve each of the above models. 102 Research Question 2A: The Random Inte rcept Model Predicting Expression of Negative Emotion from Lagged Effects of toddler variables in Question 2A were examined and are presented below. A summary of all model results in provided in Table 31 31 . Model results for each predictor are then presented in the following pages (Tables 32 - 35). Model results for non - significant predictor (T - INI - VDis) are presented in Appendix I (Table I7). Table 31 0 - 3) Model 1 - Random Intercept Fixed Slope Model 2 - Random Intercept Random Slope Lagged Toddler Variables in Question 2A Variable Label Does X significantly predict Y? Is directionality as hypothesized? Does X predict Y? Does X - Y slope vary over dyads? Does freeing slope significantly reduce misfit? Is directionality as hypothesized? Independent Regulatory Strategies Initiated Verbal Distraction T - INI - VDis No Yes (Neg) No Yes No Yes (Neg) Initiated Non - Verbal Distraction T - INI - nonVDis Yes Yes (Neg) Yes Yes Yes Yes (Neg) Initiated Verbal Keys T - INI - Vkeys No No (Pos) Yes Yes Yes No (Pos) Initiated Verbal Self - Control T - INI - Vctrl - - - - - - Initiated Verbal Desire T - INI - Vdes Yes No (Pos) Yes No No No (Pos) Self - Comfort T - SComf Yes Yes (Neg) Yes Yes Yes Yes (Neg) Note . All Toddlers' Regulatory Strategies were predicted to be positively associated with outcome. 2A strategies are toddler init iated and don't involve mother, 2B strategies involve mother or are in response to mother. Bold font indicates largest slope coefficient. Italics font indicates X became significant predictor in model 2. * = p - value for slope coefficients with Non - Robust standard errors were < . 05. All results report p - values for models estimated with robust standard errors. p < .10. Non - significant estimates and unexpected directionality are in red. 31 Also provided in on page 6 of Supplemental Materials 103 Lagged Toddler Initiated Non - Verbal Distraction . A two - level hierarchical model examined the effects of lagged Toddler Initiated Non - of Negative Emotion (ENE - O) while controlling for effects of linear elapsed time, quadratic elapsed time, and toddler gender. The model was specified with the following equation: ENE - O ij 00 Gender (Gender j ) T - INI - nonVdis (T - INI - nonVDis i - 1j ) + ETL (ET ij ET 2 (ET 2 ij ) + u 0j + e ij Results of final estimated model revealed that on average, levels of ENE - O was significantly associated with lagged Toddler Initiated Non - Verbal Di straction as hypothesized in expected negative direction (Table 32). For every unit increase in lagged Toddler Initiated Non - Verbal Distraction (e.g., looks away, walks out of the room, engages in play expression of negative emotion sign ificantly decreased by 0.18 on a scale of 0 - average overall expression of negative emotion was 0.30 units higher for boys after accounting for effects for all predictors. Additionally, the current model was compared against a model (Model 2) in which slopes between lagged Toddler Initiated Non - Verbal Distraction and ENE - O were allowed to vary between dyads. Results of deviance tests ( = chi squared difference) indicated that allowing the relationship between lagged Toddler Initiated Non - Verbal Distraction and ENE - O to vary among dyads would significantly improve model fit = 45.21, df = 2, p < 0.001. Indeed, when the association between lagged Toddler Initiated Non - Verbal Distraction and ENE - O was allowed to vary between dya ds, the p - value associated with the random slope coefficient was significant ( p <.05), suggesting that dyads varied in the association between lagged Toddler Initiated Non - Verbal Distraction and ENE - O. In the random - intercept random - slope model (Model 2), for every unit increase in lagged Toddler Initiated Non - Verbal Distraction , overall expression of negative emotion significantly decreased by 0.25 on a scale of 0 - 3 . 104 Table 32 Effects of Lagged Toddler Initiated Non - Verbal Distraction on Overall Expression of Negative Emotion Model 1 Model 2 Fixed Effects Coefficient S.E Coefficient S.E Intercept 0.67 *** 0.15 0.72 *** 0.15 Gender 0.30 * 0.15 0.21 0.12 Initiated Non - Verbal Distraction - 0.18 *** 0.04 - 0.25 *** 0.04 Linear ET 0.00 0.00 0.00 0.0 0 Quadratic ET - 0.00 0.00 - 0.00 0.00 Random Effects Variance Components SD Variance Components SD u 0j 0.67 *** 0.82 0.78 *** 0.88 u T - INI - nonVdis , j - - 0.11 *** 0.34 e ij 0.32 0.56 0.30 0.55 Deviance 2886.89 2841.68 Note. Elapsed Time (ET) is in 10s i ntervals. * p < .05, ** p < .01, *** p < .001, p < .10 Lagged Toddler Initiated Verbal Keys A two - level hierarchical model examined the (ENE - O) while controlling f or effects of linear elapsed time, quadratic elapsed time, and toddler gender. The model was specified with the following equation: ENE - O ij 00 Gender (Gender j ) T - INI - Vkeys (T - INI - Vkeys i - 1j ) + ETL (ET ij ET 2 (ET 2 ij ) + u 0j + e ij - overall expression of negative emotion was 0.31 units higher for boys (scale of 0 - 3) after accounting for effects for all predictors. Additionally, the current model was compared against a model (Model 2) in which slopes between lagged Toddler Initiated V erbal Keys and ENE - O were allowed to vary between dyads. Results of deviance tests ( = chi squared difference) indicated that allowing the relationship between lagged Toddler Initiated Verbal Keys and ENE - O to vary among dyads would significantly improve model fit ( = 8.98, df = 2, p < 0.05). Indeed, when the associatio n between lagged Toddler Initiated Verbal Keys and ENE - O was allowed to vary 105 between dyads, the p - value associated with the random slope coefficient was significant (p <.05), suggesting that dyads varied in the association between lagged Toddler Initiated Verbal Keys and ENE - O. Additionally, in the random - intercept random - slope model (Model 2), for every unit increase in lagged Toddler Initiated Verbal Keys , emotion marginally increased by 0.24 ( p < .10) on a scale o f 0 - 3, but in unexpected direction. Table 33 Effects of Lagged Toddler Initiated Verbal Keys on Overall Expression of Negative Emotion Model 1 Model 2 Fixed Effects Coefficient S.E Coefficient S.E Intercept 0.61 *** 0.15 0.60 *** 0.15 Gender 0.31 * 0. 15 0.30 * 0.15 Initiated Verbal Keys 0.12 0.08 0.24 0.13 Linear ET 0.00 0.00 0.00 0.00 Quadratic ET - 0.00 0.00 - 0.00 0.00 Random Effects Variance Components SD Variance Components SD u 0j 0.73 *** 0.85 0.73 *** 0.85 u T - INI - Vkeys , j - - 0.24 *** 0.49 e ij 0.32 0.57 0.31 0.56 Deviance 2914.07 2905.10 Note. Elapsed Time (ET) is in 10s intervals. * p < .05, ** p < .01, *** p < .001, p < .10 Lagged Toddler Initiated Verbal Self - Control . No instances of Toddler Initiated Verbal Self - - Control was directed toward or in response to mom and coded as Toddler Initiated Verbal Self - Control Bids to Mom under dependent regulatory strategies. Lagged Toddler Initiated Verbal Desire. A two - level hier archical model examined the (ENE - O) while controlling for effects of linear elapsed time, quadratic elapsed time, and toddler gender. The model was specified with the following equation: ENE - O ij 00 Gender (Gender j ) T - INI - Vdes (T - INI - Vdes i - 1j ) + ETL (ET ij ET 2 (ET 2 ij ) + u 0j + e ij 106 Results of final estimated model revealed that on average, levels of ENE - O was marginally associated with lagged Toddler Initiated Verbal Desire as hypot hesized but in unexpected positive direction (Table 34). For every unit increase in lagged Toddler Initiated marginally increased by 0.11 on a scale of 0 - 3. Tod emotion was 0.31 units higher for boys after accounting for effects for all predictors. Additionally, the current model was compared against a model (Model 2) in which slopes between lagged Toddler Initiated V erbal Desire and ENE - O were allowed to vary between dyads. However, results of deviance tests ( = chi squared difference) indicated that allowing the relationship between lagged Toddler Initiated Verbal Desire and ENE - O to vary among dyads would not significantly help to explain more variance in ENE - O between toddlers ( = 0.016 x, df = 2, p > 0.05). Additionally, the p - value associated with random slope was not significant suggesting that the association between lagged Toddler Initiated Verbal Desire and ENE - O does not vary between dyads. However, in the random - intercept random - slope model (Model 2), for every unit increase in lagged Toddler Initiated Verbal Desire , expression of negative emotion significantly (no longer marginal) increased by 0.15 on a scale of 0 - 3 ( p = 0.02) . Table 34 Effects of Lagged Toddler Initiated Verbal Desire on Overall Expression of Negative Emotion Fixed Effects Coefficient S.E Intercept 0.62 *** 0.15 Gender 0.31 * 0.15 Initiated Verbal De sire 0.11 0.06 Linear ET 0.00 0.00 Quadratic ET - 0.00 0.00 Random Effects Variance Components SD u 0j 0.72 *** 0.85 e ij 0.32 0.57 Note. Elapsed Time (ET) is in 10s intervals. * p < .05, ** p < .01, *** p < .001, p < .10 107 Lagged Toddler Self - Comfort . A two - level hierarchical model examined the effects of lagged Toddler Self - - O) while controlling for effects of linear elapsed time, quadratic elapsed time, and toddler gender. The model was specified w ith the following equation: ENE - O ij 00 Gender (Gender j ) T - Scomf (T - Scomf i - 1j ) + ETL (ET ij ET 2 (ET 2 ij ) + u 0j + e ij Results of final estimated model revealed that on average, levels of ENE - O was significantly associated with lagged Toddler S elf - Comfort as hypothesized in expected negative direction (Table 35). For every unit increase in lagged Toddler Self - Comfort (e.g., thumb decreased by 0.18 on a scale of 0 - was 0.31 units higher for boys after accounting for effects for all predictors. Additionally, the current model was compared against a model (Model 2) in which slopes between lagged Toddler Self - Comfort and ENE - O were allowed to vary between dyads. Results of deviance tests ( = chi squared difference) indicated that allowing the relationship between lagged Toddler Self - Comfort and ENE - O to vary among dyads would significantl y improve model fit = 29.67, df = 2, p < 0.001). Indeed, when the association between lagged Toddler Self - Comfort and ENE - O was allowed to vary between dyads, the p - value associated with the random slope coefficient was significant (p <.05), su ggesting that dyads varied in the association between lagged Toddler Self - Comfort and ENE - O. In the random - intercept random - slope model (Model 2), for every unit increase in lagged Toddler Self - Comfort , of negative emotion sign ificantly decreased by 0.16 on a scale of 0 - 3 . 108 Table 35 Effects of Lagged Toddler Self - Comfort on Overall Expression of Negative Emotion Model 1 Model 2 Fixed Effects Coefficient S.E Coefficient S.E Intercept 0.77 *** 0.15 0.79 *** 0.15 Gender 0. 31 * 0.15 0.28 0.15 Self - Comfort - 0.18 *** 0.04 - 0.16 *** 0.04 Linear ET 0.00 0.00 0.00 0.00 Quadratic ET - 0.00 0.00 - 0.00 0.00 Random Effects Variance Components SD Variance Components SD u 0j 0.72 *** 0.85 0.81 *** 0.90 u T - Scomf , j - - 0.06 *** 0.24 e ij 0.31 0.56 0.29 0.54 Deviance 2877.72 2848.05 Note. Elapsed Time (ET) is in 10s intervals. * p < .05, ** p < .01, *** p < .001, p < .10 Question 2A Summary of Results. In summary, on average, toddler boys expressed higher overall levels of intensi ty and predominance of negative emotion (i.e., ENE - O) across the two - minute task, after each toddler independent regulatory strategy (Question 2A) from a expression of negative emotion significantly differed between 10s intervals (within dyad variance) based on the amount of lagged initiated non - verbal distractions, lagged initiated verbal desire, and lagged self - comforting behaviors. No effects were found for lagged to ddler initiated verbal distraction. Zero instances of independent verbalizations in the form of self - direction were observed. were associated with higher levels of overall expression of negative emotion, toddlers expressed lower levels of overall negative emotion as they initiated non - verbal distractions (e.g., shift focus away from keys for more than 3 seconds) and self - comforting physical behaviors (e.g., rocking back and forth) in previous intervals as expected. Among all significant predictors in Question 2A, more 109 lagged toddler initiated non - verbal distraction was most strongly assoc iated with lower levels of overall expression of negative emotion (random - intercept random slope), while more lagged toddler verbal desire was least strongly associated with higher levels of overall expression of negative emotion. No statistically signifi cant associations were found for linear or quadratic elapsed time but they were retained in the model since excluding them did not change significant coefficients and the results of the curve estimations revealed significant linear and quadratic trajectori es of change for ENE - O across the task. As evident by the statistically significant residuals (u oj ), indicating significant remaining unexplained variance in average ENE - O between dyads when all predictors are taken into account, there is room to improve e ach of the above models. Research Question 2B: The Random Intercept Model Predicting Expression of . Effects of toddler variables in Question 2B were examined and are presented below. A summary of all model results in provided in Table 36 32 . Model results for each predictor are then presented in the following pages (Tables 37 - 40). Model results for non - significant predictors ( T - Vkeys - 2M, T - Vdes - 2M, T - ComfSk ) are presented in Appendix I (T able I8 - I10). 32 Also provided in on page 5 of Supplemental Materials 110 Table 36 0 - 3) Model 1 - Random Intercept Fixed Slope Model 2 - Random Intercept Random Slope Lagged Toddler Vari ables in Question 2B Variable Label Does X significantly predict Y? Is directionality as hypothesized? Does X predict Y? Does X - Y slope vary over dyads? Does freeing slope significantly reduce misfit? Is directionality as hypothesized? Dependent Regulato ry Strategies Verbal Distraction - Bids to Mom T - VDis - 2M Yes No (Pos) Yes Yes Yes No (Pos) Joined Verbal Distraction T - JNT - VDis No Yes (Neg) Yes No Yes Yes (Neg) Joined Non - Verbal Distraction T - JNT - nonVDis Yes Yes (Neg) Yes Yes Yes Yes (Neg) Verba l Keys Bids to Mom T - Vkeys - 2M No Yes (Neg) No Yes Yes Yes (Neg) Verbal Self - Control - Bids to Mom T - Vctrl - 2M No No (Pos) Yes Yes No Yes (Neg) Verbal Desire Bids to Mom T - Vdes - 2M No No (Pos) No Yes No Yes (Neg) Toddler Physical Comfort Seeking T - ComfSk No No (Pos) No Yes Yes No (Pos) Note. All Toddlers' Regulatory Strategies were predicted to be positively associated with outcome. 2A strategies are toddler initiated and don't involve mother, 2B strategies involve mother or are in response to mo ther. Bold font indicates largest slope coefficient. Italics font indicates X became significant predictor in model 2. * = p - value for slope coefficients with Non - Robust standard errors were < .05. All results report p - values for models estimated with robu st standard errors. p < .10. Non - significant estimates and unexpected directionality are in red. 111 Lagged Toddler Verbal Distraction - Bids to Mom . A two - level hierarchical model examined the effects of lagged Toddler Verbal Distraction Bi ds to Mom Expression of Negative Emotion (ENE - O) while controlling for effects of linear elapsed time, quadratic elapsed time, and toddler gender. The model was specified with the following equation: ENE - O ij 00 Gender (Gender j ) T - Vdis - 2M (T - Vdis - 2M i - 1j ) + ETL (ET ij ET 2 (ET 2 ij ) + u 0j + e ij Results of final estimated model revealed that on average, levels of ENE - O was marginally associated with lagged Toddler Verbal Distraction Bids to Mom a s hypothesized but in unexpected positive direction (Table 37). For every unit increase in lagged Toddler Verbal Distraction Bids to Mom (e.g., mom e emotion marginally increased by 0.18 on a scale of 0 - after accounting for effects for all predictors. Additionally, the current model was compared against a model (Model 2) in which slopes between lagged Toddler Verbal Distraction Bids to Mom and ENE - O were allowed to vary between dyads. Results of deviance tests ( = chi squared difference) indicated that allowing the relationship between lagged Toddler Verbal Distraction Bids to Mom and ENE - O to vary among dyads would significantly improve model fit = 9.19, df = 2, p =0.01) Indeed, when the association between lagged Toddler Verbal Distraction Bids to Mom and ENE - O was allowed to vary be tween dyads, the p - value associated with the random slope coefficient was significant (p <.05), suggesting that dyads varied in the association between lagged Toddler Verbal Distraction Bids to Mom and ENE - O. In the random - intercept random - slope model 112 ( Model 2), for every unit increase in lagged Toddler Verbal Distraction Bids to Mom , overall expression of negative emotion significantly increased by 0.21 on a scale of 0 - 3 . Table 37 Effects of Lagged Toddler Verbal Distraction Bids to Mom on Overall Expression of Negative Emotion Model 1 Model 2 Fixed Effects Coefficient S.E Coefficient S.E Intercept 0.63 *** 0.15 0.63 *** 0.15 Gender 0.30 * 0.15 0.31 * 0.15 Verbal Distraction Bids to Mom 0.18 0.06 0.21 ** 0.08 Linear ET 0.00 0.00 0.0 0 0.00 Quadratic ET - 0.00 0.00 - 0.00 0.00 Random Effects Variance Components SD Variance Components SD u 0j 0.73 *** 0.85 0.73 *** 0.85 u T - Vdis - 2M x , j - - 0.7 * 0.26 e ij 0.32 0.56 0.31 0.56 Deviance 2907.88 2898.68 Note. Elapsed Time (ET) is in 10s in tervals. * p < .05, ** p < .01, *** p < .001, p < .10 Lagged Toddler Joined Verbal Distraction. A two - level hierarchical model examined the effects of lagged Toddler Joined Verbal Distraction Emotion (ENE - O) while contro lling for effects of linear elapsed time, quadratic elapsed time, and toddler gender. The model was specified with the following equation: ENE - O ij 00 Gender (Gender j ) T - JNT - VDis (T - JNT - VDis i - 1j) + ETL (ET ij ET 2 (ET 2 ij ) + u 0j + e ij Unexpectedly, lagged Toddler Joined Verbal Distraction (e.g., mother starts counting and child joins in ) was not associated with ENE - O but was in expecte d negative direction (Table 38). of 0 - 3) after accounting for effects for all predictors. Additionally, the current model was compared against a model (Model 2) in which slopes between lagged Toddler Joined Verbal Distraction and ENE - O were allowed to vary between dyads. Results of deviance tests ( = chi squared difference) indicated that allowing the relationship between lagged Toddler Joined 113 Verbal Distraction and ENE - O to vary among dyads would significantly improve model fit ( = 6.49, df = 2, p < 0.05). However, the p - value as sociated with random slope was not significant suggesting that the association between lagged Toddler Joined Verbal Distraction and ENE - O does not vary between dyads. Additionally, in the random - intercept random - slope model (Model 2), for every unit increa se in lagged Toddler Joined Verbal Distraction , expression of negative emotion significantly decreased by 0.19 on a scale of 0 - 3 ( p < 0.001) as hypothesized in the expected direction. Table 38 Effects of Lagged Toddler Joined Verbal Dis traction on Overall Expression of Negative Emotion Model 1 Model 2 Fixed Effects Coefficient S.E Coefficient S.E Intercept 0.63 *** 0.15 0.62 *** 0.15 Gender 0.31 * 0.15 0.32 * 0.15 Joined Verbal Distraction - 0.06 0.05 - 0.19 *** 0.04 Linear ET 0.00 0.00 0.00 0.00 Quadratic ET - 0.00 0.00 - 0.00 0.00 Random Effects Variance Components SD Variance Components SD u 0j 0.72 *** 0.85 0.72 *** 0.85 u T - JNT - VDis , j - - 0.07 0.26 e ij 0.32 0.56 0.32 0.57 Deviance 2916.46 2909.97 Note. Elapsed Time (ET) is in 10 s intervals. * p < .05, ** p < .01, *** p < .001, p < .10 Lagged Toddler Joined Non - Verbal Distraction . A two - level hierarchical model examined the effects of lagged Toddler Joined Non - Verbal Distraction of Negative Emotion (ENE - O) while controlling for effects of linear elapsed time, quadratic elapsed time, and toddler gender. The model was specified with the following equation: ENE - O ij 00 Gender (Gender j ) T - JNT - nonVDis (T - JNT - nonVDis i - 1j) + ETL (ET ij ET 2 (ET 2 ij ) + u 0j + e ij Results of final estimated model revealed that on average, levels of ENE - O was significantly associated with lagged Toddler Joined Non - Verbal Distr action as hypothesized in the expected negative direction (Table 39). For every unit increase in lagged Toddler Joined 114 Non - Verbal Distraction (e.g., mother initiates play and child joins in, mother points to a direction away from keys and child looks where she is pointing emotion significantly decreased by 0.11 on a scale of 0 - of negative emotion was 0.30 units higher for boys after accounting for effects for all predictors. Additionally, the current model was compared against a model (Model 2) in which slopes between lagged Toddler Joined Non - Verbal Distraction and ENE - O were allowed to vary between dyads. Results of deviance tests ( = chi squared difference) indica ted that allowing the relationship between lagged Toddler Joined Non - Verbal Distraction and ENE - O to vary among dyads would significantly improve model fit ( = 7.09, df = 2, p < 0.05). Indeed, when the association between lagged Toddler Joined No n - Verbal Distraction and ENE - O was allowed to vary between dyads, the p - value associated with the random slope coefficient was significant ( p <.01), suggesting that dyads varied in the association between lagged Toddler Joined Non - Verbal Distraction and EN E - O. In the random - intercept random - slope model (Model 2), for every unit increase in lagged Toddler Joined Non - Verbal Distraction , expression of negative emotion significantly decreased by 0.25 on a scale of 0 - 3. Table 39 Effects of La gged Toddler Joined Non - Verbal Distraction on Overall Expression of Negative Emotion Model 1 Model 2 Fixed Effects Coefficient S.E Coefficient S.E Intercept 0.62 *** 0.15 0.66 *** 0.14 Gender 0.30 * 0.15 0.21 0.13 Joined Non - Verbal Distraction - 0.11 ** 0.04 - 0.25 *** 0.04 Linear ET 0.00 0.00 0.00 0.00 Quadratic ET - 0.00 0.00 - 0.00 0.00 Random Effects Variance Components SD Variance Components SD u 0j 0.71 *** 0.84 0.73 *** 0.85 u T - JNT - nonVDis , j - - 0.06 ** 0.24 e ij 0.32 0.57 0.32 0.57 Deviance 2911.65 2904.56 Note. Elapsed Time (ET) is in 10s intervals. * p < .05, ** p < .01, *** p < .001, p < .10 115 Lagged Toddler Verbal Self - Control - Bids to Mom. A two - level hierarchical model examined the effects of lagged Toddler Verbal Self - Control Bids to Mom Expression of Negative Emotion (ENE - O) while controlling for effects of linear elapsed time, quadratic elapsed time, and toddler gender. The model was specified with the following equation: ENE - O ij 00 Gender (Gender j ) T - Vctrl - 2M (T - Vctrl - 2M i - 1j ) + ETL (ET ij ET 2 (ET 2 ij ) + u 0j + e ij Unexpectedly, lagged Toddler Self - Control - Bids to Mom (e.g., while looking at or in response to mom ) was not associated with EN E - average overall expression of negative emotion was 0.31 units higher for boys (scale of 0 - 3) after accounting for effects for all predictors. Additionally, the current model was compared against a model (Model 2) in which slopes b etween lagged Toddler Self - Control - Bids to Mom and ENE - O were allowed to vary between dyads. However, results of deviance tests ( = chi squared difference) indicated that allowing the relationship between lagged Toddler Self - Control - Bids to Mom and ENE - O to vary among dyads would not significantly help to explain more variance in ENE - O between toddlers ( = 2.36, df = 2, p > 0.05). However, when the association between lagged Toddler Self - Control - Bids to Mom and ENE - O was allowed to vary between dyads, the p - value associated with the random slope coefficient was significant (p <.01), suggesting that dyads varied i n the association between lagged Toddler Self - Control - Bids to Mom and ENE - O. Additionally, in the random - intercept random - slope model (Model 2), for every unit increase in lagged Toddler Self - Control - Bids to Mom , emotion significantly decreased by 0.47 on a scale of 0 - 3 ( p = 0.007) as hypothesized in the expected direction. 116 Table 40 Effects of Lagged Toddler Self - Control - Bids to Mom on Overall Expression of Negative Emotion Model 1 Model 2 Fixed Effects Coe fficient S.E Coefficient S.E Intercept 0.63 *** 0.15 0.65 *** 0.15 Gender 0.31 * 0.15 0.28 0.14 Self - Control - Bids to Mom 0.03 0.14 - 0.47 ** 0.17 Linear ET 0.00 0.00 0.00 0.00 Quadratic ET - 0.00 0.00 - 0.00 0.00 Random Effects Variance Components SD Vari ance Components SD u 0j 0.72 *** 0.85 0.72 *** 0.85 u T - Vctrl - 2M , j - - 0.75 ** 0.86 e ij 0.32 0.57 0.32 0.57 Deviance 2915.34 2912.98 Note. Elapsed Time (ET) is in 10s intervals. * p < .05, ** p < .01, *** p < .001, p < .10 Question 2B Summary of Result s . In summary, on average, toddler boys expressed higher overall levels of intensity and predominance of negative emotion (i.e., ENE - O) across the two - minute task, after each toddler dependent regulatory strategy (Question 2B) from a previous 10s interval negative emotion significantly differed between 10s intervals (within dyad variance) based on the amount of verbal distraction - bids to mom, joined verbal distraction, joined n on - verbal distraction, and verbalizations in the form of self - direction directed toward or in response to mother that they displayed in a previous interval. No effects were found for lagged toddler verbalization about keys directed toward mothers , lagged toddler verbalization expressive of desire states directed toward mothers , or lagged toddler comfort seeking. While contrary to my hypotheses, more lagged verbal distraction h higher levels of overall expression of negative emotion, toddlers expressed lower levels of overall negative emotion as they joined mother - initiated verbal (e.g., mother starts counting and toddler joins in counting) and non - verbal distraction (e.g., mot her points to an object other than keys and toddler looks in that 117 direction). Additionally, as hypothesized, toddlers who verbalized statements in the form of self - s intervals, expressed lower levels of intensity and predominance of negative emotions (anger or sadness). In fact, effect of lagged verbalizations in the form of self - direction was the strongest predictor of lowered expression of negative emotion among al l significant dependent regulatory strategies while lagged toddler joined non - verbal distraction had the weakest significant slope. No statistically significant associations were found for linear or quadratic elapsed time but they were retained in the mo del since excluding them did not change significant coefficients and the results of the curve estimations revealed significant linear and quadratic trajectories of change for ENE - O across the task. As evident by the statistically significant residuals (u oj ), indicating significant remaining unexplained variance in average ENE - O between dyads when all predictors are taken into account, there is room to improve each of the above models. Summary of both 2A and 2B: In terms of overall expression of negative emo tion, results provide evidence for effectiveness of lagged i ndependent r egulatory s trategies and lagged d ependent r egulatory s trategies. However, these effects must be considered separately for distracting strategies, verbalizations toward the delay task, and physical strategies. In terms of - verbal distractions were similar between initiated was distraction was direct ed toward mothers (wrong direction positive) or join ed verbal distraction (right direction negative). Among independent verbalizations towards the delay task, toddler initiated talk that describe d the keys (e.g., Pooh) and toddler initiated talk expressive of internal states (e.g., I need it) were related to higher overall expression of negative emotion. However, among dependent verbalizations towards the delay task, although talk about 118 keys and desire states directed toward or in response t o mother were not significant, they were now in t he expected negative direction. Also, verbalizations toward mothers in the form of self - direction were significantly related to lower overall expression of negative emotion. These verbalizations al so had the strongest negative effect on ENE - O among all toddler regulatory strategies (both independent and dependent, both verbal and non - verbal). Additionally, while physically self - comforting behaviors were associated with lower overall expression of ne gative emotion, comfort seeking behaviors such as leaning body against mom did not have significant effects and were in an unexpected positive direction. Research Question 3A: The Random Intercept Model Predicting Overall Touch from Lagged Maternal Regula tory Attempts and Mind - Mindedness. Effects of maternal variables in Question 3A were examined and are presented below. A summary of all model results in provided in Table 41 33 . Model results for each predictor are then presented in the following pages (Tabl es 42 - xx). Model results for non - significant predictors ( V - Comf, M - JNT - Dis, M - VO2D, and AMM ) are presented in Appendix I (Table I11 - I14). 33 Also provided in on page 6 of Supplemental Materials 119 Table 41 Question 3A. Do Lagged Maternal Regulatory Attempts and Lagged Mind - uch (Count) Model 1 - Random Intercept Fixed Slope Model 2 - Random Intercept Random Slope Lagged Maternal Variables in Question 3A Variable Label Does X significantly predict Y? Is directionality as hypothesized? Does X predict Y? Does X - Y slope vary over dyads? Does freeing slope significantly reduce misfit? Is directionality as hypothesized? Verbal Comfort V - Comf No Yes (Neg) No No No Yes (Neg) Physical Comfort Phy - Comf No Yes (Neg) Yes Yes Yes Yes (Neg) Initiated Distraction M - INI - Dis Yes Ye s (Neg) Yes No Yes Yes (Neg) Joined Distraction M - JNT - Dis No Yes (Neg) No No Yes Yes (Neg) Verbal Orientation to Delay M - VO2D No No (Pos) No No No No (Pos) Positive Commands Pos - Comd Yes Yes (Neg) No No No Yes (Neg) Positive Emotional Reactions Pos - Aff Yes* No (Pos) Yes Yes Yes No (Pos) Appropriate MRC AMM No Yes (Neg) No Yes No Yes (Neg) Note . MRC = Mind - Related Comments; 3A predictors were hypothesized to be negatively associated with outcome; Bold font indicates largest slope coefficient. Italics font indicates X became significant predictor in model 2. * = p - value for slope coefficients with Non - Robust standard errors were < .05. All results report p - values for models estimated with robust standard errors. p < .10. Non - significant esti mates and unexpected directionality are in red. 120 Lagged Maternal Physical Comfort. A two - level hierarchical model examined the effects of linear elapsed time , quadratic elapsed time, and toddler gender. The model was specified with the following equation: Overall Touch ij 00 Gender (Gender j ) Phy - Comf (Phy - Comf i - 1j ) + ETL (ET ij ET 2 (ET 2 ij ) + u 0j + e ij Unexpectedly, lagged maternal Physical Comf ort (e.g., giving hugs) was not associated gender were not significantly associated. Additionally, the current model was compared against a model (Model 2) in which sl opes between lagged maternal Physical Comfort and Overall Touch were allowed to vary between dyads. Results of deviance tests ( = chi squared difference) indicated that allowing the relationship between lagged maternal Physical Comfort and Overall Touch to vary among dyads would significantly improve model fit = 7.12, df = 2, p < 0.05. Indeed, when the associat ion between lagged maternal Physical Comfort and Overall Touch was allowed to vary between dyads, the p - value associated with the random slope coefficient was marginally significant ( p <.10), suggesting that dyads tended to vary in the association between lagged maternal Physical Comfort and Overall Touch. Additionally, in the random - intercept random - slope model (Model 2), for every unit increase in lagged maternal Physical Comfort , marginally decreased by 0.05 ( p = 0.06) in an aver age interval (mean per interval = .27, range per interval 0 - 6) as hypothesized in the expected direction. 121 Table 42 Effects of Lagged Maternal Physical Comfort on Overall Touch Model 1 Model 2 Fixed Effects Coefficient S.E Coefficient S.E Interce pt 0.60 *** 0.10 0.59 *** 0.10 Gender - 0.02 0.05 - 0.02 0.05 Physical Comfort - 0.03 0.03 - 0.05 0.03 Linear ET - 0.01 *** 0.01 - 0.01 *** 0.01 Quadratic ET 0.00 ** 0.00 0.00 ** 0.00 Random Effects Variance Components SD Variance Components SD u 0j 0.08 *** 0.28 0.09 *** 0.29 u Phy - comf , j - - 0.01 0.11 e ij 0.28 0.53 0.28 0.53 Deviance 2467.82 2460.70 Note. Elapsed Time (ET) is in 10s intervals. * p < .05, ** p < .01, *** p < .001, p < .10 Lagged Maternal Initiated Distraction . A two - level hierarchical mode l examined the effects of linear elapsed time, quadratic elapsed time, and toddler gender. The model was specified with the following equation: Overall Touch ij 00 Gender (Gender j M - INI - Dis ( M - INI - Dis i - 1j ) + ETL (ET ij ET 2 (ET 2 ij ) + u 0j + e ij Results of final estimated model revealed that on average, Overall Touch was significantly associated with lagged maternal Initiated Distraction as hypoth esized and in expected direction (Table 43). For every unit increase in lagged maternal Initiated Distraction an average interval (mean per interval = .27, rang e per interval 0 - and gender were not significantly associated. Additionally, the current model was compared against a model (Model 2) in which slopes between lagged maternal Initiated Distraction and Overall Touch were allowe d to vary between dyads. Results of deviance tests ( = chi squared difference) indicated that allowing the relationship between lagged maternal Initiated Distraction and Overall Touch to vary among 122 dyads would significantly improve model fit = 12.81, df = 2, p < 0.01. However, the p - value associated with random slope was not significant suggesting that the association between lagged maternal Initiated Distraction and Overall Touch does not vary between dyads. In the random - intercept random - slope model (Model 2), for every unit increase in lagged maternal Initiated Distraction , per interval = .27, range per interval 0 - 6). Table 43 Effects of Lagged Matern al Initiated Distraction on Overall Touch Model 1 Model 2 Fixed Effects Coefficient S.E Coefficient S.E Intercept 0.60 *** 0.11 0.60 *** 0.12 Gender - 0.01 0.06 - 0.03 0.05 Initiated Distraction - 0.06 * 0.03 - 0.08 ** 0.03 Linear ET - 0.01 ** 0.01 - 0.01 ** 0 .00 Quadratic ET 0.00 ** 0.00 0.00 ** 0.00 Random Effects Variance Components SD Variance Components SD u 0j 0.08 *** 0.28 0.09 *** 0.31 u M - INI - Dis , j - - 0.02 0.14 e ij 0.28 0.53 0.28 0.53 Deviance 2465.56 2452.75 Note. Elapsed Time (ET) is in 10s inte rvals. * p < .05, ** p < .01, *** p < .001, p < .10 Lagged Maternal Positive Commands . A two - level hierarchical model examined the controlling for effects of linear elaps ed time, quadratic elapsed time, lagged maternal verbosity, and toddler gender. The model was specified with the following equation: Overall Touch ij 00 Gender (Gender j Pos - Comd ( Pos - Comd i - 1j ) + ETL (ET ij ET 2 (ET 2 ij Verb (Verb i - 1j ) + u 0j + e ij Results of final estimated model revealed that on average, Overall Touch was marginally associated with lagged maternal Positive Commands as hypothesized in expected direction (Table 44). For every unit increase in lagged maternal Positive Command 123 average interval (mean per interval = .27, range per interval 0 - gender were not significantly associated. Addit ionally, the current model was compared against a model (Model 2) in which slopes between lagged maternal Positive Commands and Overall Touch were allowed to vary between dyads. However, results of deviance tests ( = chi squared difference) indic ated that allowing the relationship between lagged maternal Positive Commands and Overall Touch to vary among dyads would not significantly help to explain more variance in Overall Touch between toddlers ( = 4.39, df = 2, p > 0.05). Additionally, the p - value associated with random slope was not significant suggesting that the association between lagged maternal Positive Commands and Overall Touch does not vary between dyads. Effects of lagged maternal Positive Commands on Overall Touch become non - significant in Model 2. Table 44 Effects of Lagged Positive Commands on Overall Touch Fixed Effects Coefficient S.E Intercept 0.57 *** 0.11 Gender - 0.02 0.06 Positive Commands - 0.03 0.02 Linear ET - 0.01 *** 0.00 Quadratic ET 0.00 ** 0.00 Verbosity 0 .02 0.01 Random Effects Variance Components SD u 0j 0.08 *** 0.28 e ij 0.28 0.53 Note. Elapsed Time (ET) is in 10s intervals. * p < .05, ** p < .01, *** p < .001, p < .10 Lagged Maternal Positive Emotional Reactions . A two - level hierarchical model examin Touch while controlling for effects of linear elapsed time, quadratic elapsed time, and toddler gender. The model was specified with the following equation: Overall Touch i j 00 Gender (Gender j ) Pos - Aff (Pos - Aff i - 1j ) + ETL (ET ij ET 2 (ET 2 ij ) + u 0j + e ij 124 Results of final estimated model revealed that on average, levels of Overall Touch was marginally associated with lagged maternal Positive Emotional Reaction s as hypothesized but was in unexpected direction (Table 45). For every unit increase in lagged maternal Positive increase by 0.07 in an average interval (mean per i nterval = .27, range per interval 0 - 6). In the same model, the p - value associated with the slope coefficient estimated with non - robust standard Additionally, the cur rent model was compared against a model (Model 2) in which slopes between lagged maternal Positive Emotional Reactions and Overall Touch were allowed to vary between dyads. Results of deviance tests ( = chi squared difference) indicated that allo wing the relationship between lagged maternal Positive Emotional Reactions and Overall Touch to vary among dyads would significantly improve model fit = 9.58, df = 2, p < 0.01. Indeed, when the association between lagged maternal Positive Emotio nal Reactions and Overall Touch was allowed to vary between dyads, the p - value associated with the random slope coefficient was significant (p <.05), suggesting that dyads varied in the association between lagged maternal Positive Emotional Reactions and O verall Touch . In the random - intercept random - slope model (Model 2), for every unit increase in lagged maternal Positive Emotional Reactions , Overall Touch marginally increased by 0.07 ( p = 0.05) in an average interval (mean per interval = .27, ra nge per interval 0 - 6). Table 45 Effects of Lagged Maternal Positive Emotional Reactions on Overall Touch Model 1 Model 2 Fixed Effects Coefficient S.E Coefficient S.E Intercept 0.56 *** 0.11 0.55 *** 0.11 Gender - 0.01 0.06 0.004 0.05 Positive Emo tional Reactions 0.07 0.04 0.07 0.04 125 Table 45 ( c Linear ET - 0.01 *** 0.00 - 0.01 *** 0.00 Quadratic ET 0.00 ** 0.00 0.00 ** 0.00 Random Effects Variance Components SD Variance Components SD u 0j 0.08 *** 0.28 0.06 *** 0.25 u Pos - Aff , j - - 0.03 * 0.16 e ij 0.28 0.53 0.27 0.52 Deviance 2464.35 2454.76 Note. Elapsed Time (ET) is in 10s intervals. * p < .05, ** p < .01, *** p < .001, p < .10 Question 3A Summary of Results . In summary , (higher scores indicate lower ability to wait) 34 , significantly differed between 10s intervals (within dyad variance) based on the amount of Initiated Distraction their mothers displayed in a previous interval. Expectedly, as m others initiated more distractions in a previous interval, t Overall Touch marginally differed between 10s intervals based on the amount of physical comfort, positive commands and positive emotional reactions their mothers dis played in a previous interval. Expectedly , as mothers offered more physical comfort or used more positive commands in a previous interval, toddlers tended to touch or attempt to touch the keys less often. However, unexpectedly, as mothers displayed positiv e emotional reactions in a previous interval, toddlers tended to touch or attempt to touch the keys more often. Among all significant predictors in Q uestion 3 A , more lagged maternal initiated distractions were most strongly related to less Overall Touch, w hile lagged maternal positive commands were least strongly (and marginally) related to Overall Touch. After accounting for variance within toddlers, while u 0j ), gender 34 higher ability to wait or more ability for delay of gratification. 126 difference s could not significantly explain this amount of between toddler variance in Overall Touch scores . 127 Research Question 3B: The Random Intercept Model Predicting Overall Touch from Lagged Maternal Regulatory Attempts and Mind - Mindedness. Effects of maternal variables in Question 3B are presented below. A summary of all model results in provided in Table 4 6 . Model results for each predictor are then presented in the following pages (Tables 47 - 49). Model results for non - significant predictors (Pun, Min, NMM) ar e presented in Appendix I (Table I15 - I17). Table 46 Question 3B. Do Lagged Maternal Regulatory Attempts and Lagged Mind - Model 1 - Random Intercept Fixed Slope Model 2 - Random Intercept Random Slope L agged Maternal Variables in Question 3B Variable Label Does X significantly predict Y? Is directionality as hypothesized? Does X predict Y? Does X - Y slope vary over dyads? Does freeing slope significantly reduce misfit? Is directionality as hypothesized? Negative Commands Neg - Comd Yes Yes (Pos) No No Yes Yes (Pos) Punitive Reactions Pun No No (Neg) No No No Yes (Pos) Minimizing Min No No (Neg) No Yes No No (Neg) Physical Restraint Phy - RST Yes Yes (Pos) Yes Yes Yes Yes (Pos) Negative Emotiona l Reactions Neg - Aff No* Yes (Pos) No Yes Yes Yes (Pos) Non - Attuned MRC NMM No No (Neg) No No No No (Neg) Note . MRC = Mind - Related Comments; 3B predictors were hypothesized to be positively associated with outcome; Bold font indicates largest slope coef ficient. Italics font indicates X became significant predictor in model 2. * = p - value for slope coefficients with Non - Robust standard errors were < .05. All results report p - p < .10. Non - signifi cant estimates and unexpected directionality are in red 128 Lagged Maternal Negative Commands . A two - level hierarchical model examined the effects of linear elapsed time, quadratic elapsed time, lagged maternal verbosity, and toddler gender. The model was specified with the following equation: Overall Touch ij 00 Gender (Gender j ) Neg - Comd ( Neg - Comd i - 1j ) + ETL (ET ij ET 2 (ET 2 ij Verb (Verb i - 1j ) + u 0j + e ij Results of final estimated model revealed that on average, Overall Touch was marginally associated with lagged maternal Negative Commands as hypothesized in expected direction (Table 47). For every unit increase in lagged maternal Negative Commands per interval = .27, range per interval 0 - significantly associated. Additionally, the curr ent model was compared against a model (Model 2) in which slopes between lagged maternal Negative Commands and Overall Touch were allowed to vary between dyads. Results of deviance tests ( = chi squared difference) indicated that allowing the rel ationship between lagged maternal Negative Commands and Overall Touch to vary among dyads would significantly improve model fit = 6.99, df = 2, p < 0.05. However, the p - value associated with random slope was not significant suggesting that the a ssociation between lagged maternal Negative Commands and Overall Touch does not vary between dyads. Effects of lagged maternal Negative Commands on Overall Touch become non - significant in Model 2. Table 47 Effects of Lagged Maternal Negative Commands on O verall Touch Model 1 Model 2 Fixed Effects Coefficient S.E Coefficient S.E Intercept 0.56 *** 0.11 0.57 *** 0.11 Gender - 0.02 0.06 - 0.03 0.05 Negative Commands 0.03 0.02 0.03 0.02 129 Table 47 ( c Linear ET - 0.01 ** 0.00 - 0.01 *** 0.00 Quadrat ic ET 0.00 ** 0 0.00 ** 0.00 Verbosity 0.002 0.01 0.002 0.01 Random Effects Variance Components SD Variance Components SD u 0j 0.08 *** 0.28 0.07 *** 0.26 u Neg - Comd ,j - - 0.004 0.06 e ij 0.28 0.53 0.27 0.52 Deviance 2471.15 2471.15 Note. Elapsed Time (ET) is in 10s intervals. * p < .05, ** p < .01, *** p < .001, p < .10 Lagged Maternal Physical Restraint . A two - level hierarchical model examined the effects o f linear elapsed time, quadratic elapsed time, and toddler gender. The model was specified with the following equation: Overall Touch ij 00 Gender (Gender j ) Phy - RST (Phy - RST i - 1j ) + ETL (ET ij ET 2 (ET 2 ij ) + u 0j + e ij Results of final estimat ed model revealed that on average, Overall Touch was marginally associated with lagged maternal Physical Restraint as hypothesized in expected direction (Table 48). For every unit increase in lagged maternal Physical Restraint (e.g., 2 = High - firmly holdin g child in lap, 1 = Low - Overall Touch tended to increase by 0.04 in an average interval (mean per interval = .27, range per interval 0 - not significantly associated. Additionally, the current model was compared against a model (Model 2) in which slopes between lagged maternal Physical Restraint and Overall Touch were allowed to vary between dyads. Results of deviance tests ( = c hi squared difference) indicated that allowing the relationship between lagged maternal Physical Restraint and Overall Touch to vary among dyads would significantly improve model fit = 9.40, df = 2, p < 0.01). Indeed, when the association betwe en lagged maternal Physical Restraint and Overall Touch was allowed to vary 130 between dyads, the p - value associated with the random slope coefficient was significant (p <.001), suggesting that dyads varied in the association between lagged maternal Physical Restraint and Overall Touch . In the random - intercept random - slope model (Model 2), for every unit increase in lagged maternal Physical Restraint , by 0.05 ( p = .05) in an average interval ( mean per interval = .2 7, range per interval 0 - 6). Table 48 Effects of Lagged Maternal Physical Restraint on Overall Touch Model 1 Model 2 Fixed Effects Coefficient S.E Coefficient S.E Intercept 0.57 *** 0.11 0.54 *** 0.11 Gender - 0.02 0.05 - 0.04 0.06 Physical Restraint 0.04 0.03 0.05 0.03 Linear ET - 0.01 ** 0.00 - 0.01 ** 0.00 Quadratic ET 0.00 ** 0.00 0.00 ** 0.00 Random Effects Variance Components SD Variance Components SD u 0j 0.07 *** 0.27 0.09 *** 0.29 u Phy - RST , j - - 0.03 0.16 e ij 0.28 0.53 0.27 0.52 Deviance 2465.63 2456.22 Note. Elapsed Time (ET) is in 10s intervals. * p < .05, ** p < .01, *** p < .001, p < .10 Lagged Maternal Negative Emotional Reactions . A two - level hierarchical model examined the effects of lagged maternal Negative Emotional Reactions Touch while controlling for effects of linear elapsed time, quadratic elapsed time, and toddler gender. The model was specified with the following equation: Overall Touch ij 00 Gender (Gender j ) Neg - Aff (Neg - Aff i - 1j ) + ETL (ET ij ET 2 (ET 2 ij ) + u 0j + e ij Unexpectedly, lagged maternal Negative Emotional Reactions (e.g., harsh tone, laughs to mock child) was not associated with Overall Touch but was in the expect ed direction (Table 49) in model with robust standard errors. However, in the same model, the p - value associated with the slope coefficient estimated with non - robust standard errors was < 0.05 suggesting that for every unit increase in lagged maternal Nega 131 increased by 0.30 in an average interval (mean per interval = .27, range per interval 0 - 6). Additionally, the current model was compa red against a model (Model 2) in which slopes between lagged maternal Negative Emotional Reactions and Overall Touch were allowed to vary between dyads. Results of deviance tests ( = chi squared difference) indicated that allowing the relationshi p between lagged maternal Negative Emotional Reactions and Overall Touch to vary among dyads would significantly improve model fit ( = 18.45, df = 2, p < 0.001). Indeed, when the association between lagged maternal Negative Emotional Reactions an d Overall Touch was allowed to vary between dyads, the p - value associated with the random slope coefficient was significant ( p <.05), suggesting that dyads varied in the association between lagged maternal Negative Emotional Reactions and Overall Touch . Ef fects of lagged maternal Negative Emotional Reactions on Overall Touch remain non - significant in Model 2. Table 49 Effects of Lagged Maternal Negative Emotional Reactions on Overall Touch Model 1 Model 2 Fixed Effects Coefficient S.E Coefficient S.E Intercept 0.60 *** 0.11 0.59 *** 0.11 Gender - 0.02 0.05 - 0.01 0.05 Negative Emotional Reactions 0.30 * 0.24 0.17 0.00 Linear ET - 0.01 ** 0.00 - 0.01 ** 0.00 Quadratic ET 0.00 ** 0.00 0.00 ** 0.00 Random Effects Variance Components SD Variance Components SD u 0j 0.07 *** 0.27 0.06 *** 0.25 u Neg - Aff , j - - 0.00 ** 0.60 e ij 0.28 0.53 0.27 0.52 Deviance 2459.45 2441.001 Note. Elapsed Time (ET) is in 10s intervals. * p < .05, ** p < .01, *** p < . 001, p < .10; for Model 1, the p - value associated with the slope c oefficient estimated with non - robust standard errors was < 0.05 for Negative Emotional Reactions Question 3B Summary of Results . In summary , (higher scores indicate lower ability to wait), significantly (non - robust p - valu e) differed between 132 10s intervals (within dyad variance) based on the amount of Negative Emotional Reactions their mothers displayed in a previous interval. As expected, as mothers displayed more Negative Emotional Reactions in a previous interval, toddler s touched and attempted to touch the keys intervals based on the amount of Negative Commands, and Physical Restraint their mothers displayed in a previous interval. As Commands and higher levels of Physical Restraint in previous intervals were marginally predictors in Question 3B, la gged maternal Negative Emotional Reactions were most strongly related ( p - value significant with non - robust standard errors) to more Overall Touch, while lagged maternal Overall Negative Commands was least strongly (and marginally) related to Overall Touch. significantly different between toddlers ( u 0j ), gender differences could not significantly explain this amount of variance. Given that there was not a large amount of between dyad variance in Overall Touch in the intercept only model ( u 0j = 0.08, p < .001 ), and that only 20% of the overall variance was explainable by any level - 2 fixed (e.g. toddler gender) and random effects (e.g., error terms associated with random in tercept and slope), it was not surprising that level - 2 explanatory variables were not significantly related to Overall Touch. In other words, there was very little between dyad variance to explain, suggesting that toddlers were more similar in their abilit y to refrain from touching the keys (i.e., delay gratification or wait) than they were ved as a count variable and was positively skewed (M = 0.27, SD = 0.61, Skewness = 2.94) such that m ajority of toddlers touched the keys 0 times for majority of intervals (94.0%, n = 1477). 133 Regulatory Strategies . Effects of toddler variables in Q uestion 4A were examined and are presented below. A summary of all model results in provided in Table 50 35 . Model results for each predictor are then presented in the following pages (Tables 51 - 54). Model results for non - significant predictors ( T - INI - Vdes ) are presented in Appendix I (Table I18). Table 50 Model 1 - Random Intercept Fixed Slope Model 2 - Random Intercept Random Slope Lagged Toddler Variables in Question 4A Variable Label Does X significantly predict Y? Is directionality as hypothesized? Does X predict Y? Does X - Y slope vary over dyads? Does freeing slope significantly reduce misfit? Is directionality as hypothesized? Independent Regulatory Str ategies Initiated Verbal Distraction T - INI - VDis Yes Yes (Neg) Yes No No Yes (Neg) Initiated Non - Verbal Distraction T - INI - nonVDis Yes Yes (Neg) Yes No No Yes (Neg) Initiated Verbal Keys T - INI - Vkeys Yes Yes (Neg) No No No Yes (Neg) Initiate d Verbal Self - Control T - INI - Vctrl - - - - - - Initiated Verbal Desire T - INI - Vdes No No (Pos) No Yes No No (Pos) Self - Comfort T - SComf Yes Yes (Neg) Yes No Yes Yes (Neg) Note . All Toddlers' Regulatory Strategies were predicted to be positively associ ated with outcome. 4A strategies are toddler initiated and don't involve mother, 4B strategies involve mother or are in response to mother. Bold font indicates largest slope coefficient. Italics font indicates X became significant predictor in model 2. * = p - value for slope coefficients with Non - Robust standard errors were < .05. All results report p - values for models estimated with robust standard errors. p < .10. Non - significant estimates and unexpected directionality are in red. 35 Also provided in on page 7 of Supplemental Materials 134 Lagged Toddler Initiated Verbal Distraction . A two - level hierarchical model examined expressive language. The model was specified with the following equation: Overall Touch ij 00 Vocab (Vocab j T - INI - Vdis (T - INI - VDis i - 1j ) + ETL (ET ij ET 2 (ET 2 ij ) + u 0j + e ij Results of final estimated model revealed that on average, levels of Overall Touch was significantly associated with lagged Toddler Initiated Verbal Distraction as hypothesized in the expected negative direction (Table 51). For every unit increase in lagged Toddler Initiated decreased by 0.23. On average, less Overall Touch was marginally associated with higher scores on expressive language such that for every unit increase in vocabulary production, Overall Touch decreased by 0. 002 in an average interval (mean per interval = .27, range per interval 0 - 6). Additionally, the current model was compared against a model (Model 2) in which slopes between lagged Toddler Initiated Verbal Distraction and Overall Touch were allowed to var y between dyads. However, results of deviance tests ( = chi squared difference) indicated that allowing the relationship between lagged Toddler Initiated Verbal Distraction and Overall Touch to vary among dyads would not significantly help to explain more variance in Overall Touch between toddlers ( = 2.08, df = 2, p > 0.05). Additionally, the p - value associated with random slope was not significant suggesting that the association between lagged Toddler Initiated Verbal Distraction and Overall Touch does not vary between dyads. In the rando m - intercept random - slope model (Model 2), for every unit increase in lagged Toddler Initiated Verbal Distraction , p < 0.001) in an 135 average interval (mean per interval = .27, range per interval 0 - 6). The results of more parsimonious model are presented in Table 51. Table 51 Effects of Lagged Toddler Initiated Verbal Distraction on Overall Touch Fixed Effects Coefficient S.E Intercept 0.69 *** 0.13 Productive Vocab - 0.00 0.001 Initiated Verbal Distraction - 0.23 ** 0.08 Linear ET - 0.01 ** 0.00 Quadratic ET 0.00 ** 0.00 Random Effects Variance Components SD u 0j 0.08 *** 0.28 e ij 0.28 0.53 Note. Elapsed Time (ET) is in 10s intervals. * p < .05, ** p < .01, *** p < .001, p < .10 Lagged Toddler Initiated Non - Verbal Distraction . A two - level hierarchical model examined the effects of lagged Toddler Initiated Non - Touch while controlling for effects of linear elapsed time, quadratic elapse expressive language. The model was specified with the following equation: Overall Touch ij 00 Vocab (Vocab j T - INI - nonVdis (T - INI - nonVDis i - 1j ) + ETL (ET ij ET 2 (ET 2 ij ) + u 0j + e ij Results of final estimated model reveal ed that on average, levels of Overall Touch was significantly associated with lagged Toddler Initiated Non - Verbal Distraction as hypothesized in the expected negative direction (Table 52). For every unit increase in lagged Toddler Initiated Non - Verbal Dist 0.23. On average, less Overall Touch was marginally associated with higher scores on expressive language such that for every unit increase in vocabulary production, Overall Touch decreased by 0.002 in an average interval (mean per interval = .27, range per interval 0 - 6). Additionally, the current model was compared against a model (Model 2) in which slopes between lagged Toddler Initiated Non - Verbal Distraction and Overall Touch were allowed to 136 vary between dyads. However, results of deviance tests ( = chi squared difference) indicated that allowing the relationship between lagged Toddler Initiated Non - Verbal Distraction and Overall Touch to vary among dyads would not significantly help to explain more variance in Overall Touch between toddler s ( = 2.11, df = 2, p > 0.05). Additionally, the p - value associated with random slope was not significant suggesting that the association between lagged Toddler Initiated Non - Verbal Distraction and Overall Touch does not vary between dyads. In th e random - intercept random - slope model (Model 2), for every unit increase in lagged Toddler Initiated Non - Verbal Distraction , decreased by 0.05 ( p = 0.04) in an average interval (mean per interv al = .27, range per interval 0 - 6). The results of more parsimonious model are presented in Table 52. Table 52 Effects of Lagged Toddler Initiated Non - Verbal Distraction on Overall Touch Model 1 Fixed Effects Coefficient S.E Intercept 0.71 *** 0.13 Productive Vocab - 0.00 0.001 Initiated Non - Verbal Distraction - 0.04 0.03 Linear ET - 0.01 ** 0.00 Quadratic ET 0.00 ** 0.00 Random Effects Variance Components SD u 0j 0.08 *** 0.27 e ij 0.28 0.53 Note. Elapsed Time (ET) is in 10s intervals. * p < .05, ** p < .01, *** p < .001, p < .10 Lagged Toddler Initiated Verbal Keys. A two - level hierarchical model examined the effects of linear elapsed time, quadratic model was specified with the following equation: Overall Touch ij 00 Vocab (Vocab j T - INI - Vkeys (T - INI - Vkeys i - 1j ) + ETL (ET ij ET 2 (ET 2 ij ) + u 0j + e ij 137 Results of final estimated model re vealed that on average, levels of Overall Touch was marginally associated with lagged Toddler Initiated Verbal Keys as hypothesized in expected negative direction (Table 53). For every unit increase in lagged Toddler Initiated Verbal Keys marginally decreased by 0.07. On average, less Overall Touch was marginally associated with higher scores on expressive language such that for every unit increase in vocabulary produc tion, Overall Touch decreased by 0.002 in an average interval (mean per interval = .27, range per interval 0 - 6). Additionally, the current model was compared against a model (Model 2) in which slopes between lagged Toddler Initiated Verbal Keys and Over all Touch were allowed to vary between dyads. However, results of deviance tests ( = chi squared difference) indicated that allowing the relationship between lagged Toddler Initiated Verbal Keys and Overall Touch to vary among dyads would not significantly help to explain more variance in Overall Touch between toddlers ( = 0.36, df = 2, p > 0.05). Additionally, the p - value associated with random slope was not significant suggesting that the association between lagged Toddler Initiated Verbal Keys and Overall Touch does not vary between dyads. Effects of lagged Toddler Initiated Verbal Keys on Overall Touch become non - significant in Model 2. Table 53 Effects of Lagged Toddler Initiated Verbal Keys on Overall Touch Fixed Effects Coefficient S.E Intercept 0.70 *** 0.13 Productive Vocab - 0.00 0.001 Initiated Verbal K eys - .07 0.04 Linear ET - 0.01 ** 0.00 Quadratic ET 0.00 ** 0.00 Random Effects Variance Components SD u 0j 0.08 *** 0.28 e ij 0.28 0.53 Note. Elapsed Time (ET) is in 10s intervals. * p < .05, ** p < .01, *** p < .001, p < .10 138 Lagged Toddler Initiated V erbal Self - Control . No instances of Toddler Initiated Verbal Self - - Control was directed toward or in response to mom and coded as Toddler Initiated Verbal Self - Control Bids to Mom under de pendent regulatory strategies. Lagged Toddler Self - Comfort . A two - level hierarchical model examined the effects of lagged toddler Self - elapsed time, quadratic elapsed time, and tod specified with the following equation: Overall Touch ij 00 Vocab (Vocab j T - Scomf (T - Scomf i - 1j ) + ETL (ET ij ET 2 (ET 2 ij ) + u 0j + e ij Results of final estimated model revealed that on average, levels of Overall Touch was significantly associated with lagged Toddler Self - Comfort as hypothesized i n expected negative direction (Table 54). For every unit increase in lagged Toddler Self - Comfort (e.g., e.g., thumb average, less Overall Touch was marginally asso ciated with higher scores on expressive language such that for every unit increase in vocabulary production, Overall Touch decreased by 0.002 in an average interval (mean per interval = .27, range per interval 0 - 6). Additionally, the current model was c ompared against a model (Model 2) in which slopes between lagged Toddler Self - Comfort and Overall Touch were allowed to vary between dyads. Results of deviance tests ( = chi squared difference) indicated that allowing the relationship between lag ged Toddler Self - Comfort and Overall Touch to vary among dyads would significantly improve model fit = 10.58, df = 2, p < 0.01). However, the p - value associated with random slope was not significant suggesting that the association between lagged Toddler Self - Comfort and Overall Touch does not vary between dyads. In the random - intercept random - 139 slope model (Model 2), for every unit increase in lagged Toddler Self - Comfort , Touch significantly decreased by 0.09 ( p < 0.001) in an ave rage interval (mean per interval = .27, range per interval 0 - 6). Expressive vocabulary was no longer marginally associated with Overall Touch in Model 2. Table 54 Effects of Lagged Toddler Self - Comfort on Overall Touch Model 1 Model 2 Fixed Effects Coefficient S.E Coefficient S.E Intercept 0.76 *** 0.14 0.72 *** 0.13 Productive Vocab - 0.00 0.001 - 0.00 0.001 Self - Comfort - 0.08 *** 0.02 - .09 *** 0.02 Linear ET - 0.01 ** 0.00 - 0.01 ** 0.00 Quadratic ET 0.00 ** 0.00 0.00 ** 0.00 Random Effects Variance Co mponents SD Variance Components SD u 0j 0.07 *** 0.27 0.10 *** 0.31 u T - Scomf , j - - 0.01 0.11 e ij 0.28 0.53 0.27 0.53 Deviance 2421.10 2410.42 Note. Elapsed Time (ET) is in 10s intervals. * p < .05, ** p < .01, *** p < .001, p < .10 Question 4A Summar y of Results . In summary, while no gender effects on Overall Touch were observed, on average, toddlers with more expressive vocabulary displayed marginally less amounts of Overall Touch across the two - minute task, after each toddler independent regulatory strategy (Question 4A) from a previous 10s interval was accounted for in (within dyad variance) based on the amount of initiated verbal and non - verbal distract ions and self - comforting behaviors they displayed in a previous interval. Additionally, marginal associations were found for lagged verbalization about keys. No effects were found for lagged toddler independent verbal desire (unexpected positive direction) . Zero instances of independent verbalizations in the form of self - direction were observed. As hypothesized, toddlers touched or attempted to touch the keys less often, when they - verbal d istractions (e.g., look 140 away for more than 3s) in a previous interval. Additionally, as hypothesized, toddlers touched or attempted to touch the keys less often, when they had displayed self - comforting behaviors (e.g., rocking back and forth) in a previou s interval. Although lagged independent verbalizations about desire states were not significantly associated with Overall Touch, as toddlers used more previous int erval, they displayed marginally fewer Overall Touch as expected. Among all significant predictors in Question 4A, lagged toddler initiated verbal distraction was most strongly associated with less Overall Touch, while lagged non - verbal initiated distracti on was least strongly associated with less Overall Touch. Research Question 4B: The Random Intercept Model Predicting Overall Touch . Effects of toddler variables in Question 4B were examined and are pr esented below. A summary of all model results in provided in Table 55 36 . Model results for each predictor are then presented in the following pages (Tables 56 - 57). Model results for non - significant predictors ( T - JNT - VDis , T - JNT - nonVDis, T - Vkeys - 2M, T - Vctrl - 2M, and T - Vdes - 2M are presented in Appendix I (Table I19 - I23). 36 Also provide d in on page 7 of Supplemental Materials 141 Table 55 Model 1 - Random Intercept Fixed Slope Model 2 - Random Intercept Random Slope Lagge d Toddler Variables in Question 4B Variable Label Does X significantly predict Y? Is directionality as hypothesized? Does X predict Y? Does X - Y slope vary over dyads? Does freeing slope significantly reduce misfit? Is directionality as hypothesized? Depe ndent Regulatory Strategies Verbal Distraction - Bids to Mom T - VDis - 2M No Yes (Neg) Yes No No Yes (Neg) Joined Verbal Distraction T - JNT - VDis No Yes (Neg) No No No Yes (Neg) Joined Non - Verbal Distraction T - JNT - nonVDis No Yes (Neg) No No No Yes (Neg) Ve rbal Keys Bids to Mom T - Vkeys - 2M No No (Pos) No No No No (Pos) Verbal Self - Control - Bids to Mom T - Vctrl - 2M No Yes (Neg) No No No Yes (Neg) Verbal Desire Bids to Mom T - Vdes - 2M No No (Pos) No Yes Yes No (Pos) Physical Comfort Seeking T - ComfSk Ye s Yes (Neg) Yes Yes Yes Yes (Neg) Note. All Toddlers' Regulatory Strategies were predicted to be positively associated with outcome. 4A strategies are toddler initiated and don't involve mother, 4B strategies involve mother or are in response to mother. Bold font indicates largest slope coefficient. Italics font indicates X became significant predictor in model 2. * = p - value for slope coefficients with Non - Robust standard errors were < .05. All results report p - values for models estimated with robust st andard errors. p < .10. Non - significant estimates and unexpected directionality are in red. 142 Lagged Toddler Verbal Distraction - Bids to Mom . A two - level hierarchical model examined the effects of lagged Toddler Verbal Distraction - Bids to M expressive language. The model was specified with the following equation: Overall Touch ij 00 Vocab (Vocab j T - Vdis - 2M (T - Vdis - 2M i - 1j ) + ETL (ET ij ET 2 (ET 2 ij ) + u 0j + e ij Unexpectedly, lagged Toddler Verbal Distraction - Bids to Mom (e.g., to mom or initiated singing while looking at mom ) was not associated with Overall Touch but was in expected negati ve direction (Table 56). On average, less Overall Touch was marginally associated with higher scores on expressive language such that for every unit increase in vocabulary production, Overall Touch decreased by 0.002 in an average interval (mean per interv al = .27, range per interval 0 - 6). Additionally, the current model was compared against a model (Model 2) in which slopes between lagged Toddler Verbal Distraction - Bids to Mom and Overall Touch were allowed to vary between dyads. However, results of dev iance tests ( = chi squared difference) indicated that allowing the relationship between lagged Toddler Verbal Distraction - Bids to Mom and Overall Touch to vary among dyads would not significantly help to explain more variance in Overall Touch be tween toddlers ( = 5.60, df = 2, p > 0.05). Additionally, the p - value associated with random slope was not significant suggesting that the association between lagged Toddler Verbal Distraction - Bids to Mom and Overall Touch does not vary between d yads. Additionally, in the random - intercept random - slope model (Model 2), for every unit increase in lagged Toddler Verbal Distraction - Bids to Mom , emotion significantly decreased by 0.07 ( p = 0.02) in an average in terval (mean per interval = .27, range per interval 0 6) as hypothesized in the expected direction. 143 Table 56 Effects of Lagged Toddler Verbal Distraction - Bids to Mom on Overall Touch Model 1 Model 2 Fixed Effects Coefficient S.E Coefficient S.E In tercept 0.70 *** 0.13 0.69 *** 0.13 Productive Vocab - 0.00 0.001 - 0.00 0.001 Verbal Distraction - Bids to Mom - 0.03 0.03 - 0.07 * 0.03 Linear ET - 0.01 ** 0.00 - 0.01 ** 0.00 Quadratic ET 0.00 ** 0.00 0.00 ** 0.00 Random Effects Variance Components SD Varian ce Components SD u 0j 0.08 *** 0.28 0.08 *** 0.28 u T - Vdis - 2M , j - - 0.03 0.16 e ij 0.28 0.53 0.28 0.53 Deviance 2430.19 2424.58 Note. Elapsed Time (ET) is in 10s intervals. * p < .05, ** p < .01, *** p < .001, p < .10 Lagged Toddler Physical Comfort Seeking . A two - level hierarchical model examined the effects of lagged Toddler Physical Comfort Seeking language. The model wa s specified with the following equation: Overall Touch ij 00 Vocab (Vocab j T - ComfSk ( T - ComfSk i - 1j ) + ETL (ET ij ET 2 (ET 2 ij ) + u 0j + e ij Results of final estimated model revealed that on average, levels of Overall Touch was significantly associated with lagged Toddler Physical Comfort Seeking as hypothesized in expected negative direction (Table 57). For every unit increase in lagged Toddler Physical Comfort Seeking (e.g., e.g., body 07. On average, less Overall Touch was marginally associated with higher scores on expressive language such that for every unit increase in vocabulary production, Overall Touch decreased by 0.002 in an average interval (mean per interval = .27, range per i nterval 0 - 6). Additionally, the current model was compared against a model (Model 2) in which slopes between lagged Toddler Physical Comfort Seeking and Overall Touch were allowed to vary 144 between dyads. Results of deviance tests ( = chi squared difference) indicated that allowing the relationship between lagged Toddler Physical Comfort Seeking and Overall Touch to vary among dyads would significantly improve model fit = 14.59, df = 2, p < 0.01). Indeed, when the association between l agged Toddler Physical Comfort Seeking and Overall Touch was allowed to vary between dyads, the p - value associated with the random slope coefficient was significant ( p <.01), suggesting that dyads varied in the association between lagged Toddler Physical C omfort Seeking and Overall Touch . In the random - intercept random - slope model (Model 2), for every unit increase in lagged Toddler Physical Comfort Seeking , expression of negative emotion significantly decreased by 0.07 ( p = 0.03) in an av erage interval (mean per interval = .27, range per interval 0 - 6). Table 57 Effects of Lagged Toddler Physical Comfort Seeking on Overall Touch Model 1 Model 2 Fixed Effects Coefficient S.E Coefficient S.E Intercept 0.71 *** 0.13 0.70 *** 0.13 Produ ctive Vocab - 0.00 0.001 - 0.00 0.001 Physical Comfort Seeking - 0.07 0.04 - 0.07 * 0.03 Linear ET - 0.01 ** 0.00 - 0.01 *** 0.00 Quadratic ET 0.00 ** 0.00 0.00 ** 0.00 Random Effects Variance Components SD Variance Components SD u 0j 0.08 *** 0.28 0.10 *** 0.3 1 u T - ComfSk , j - - 0.02 ** 0.14 e ij 0.28 0.53 0.28 0.53 Deviance 2427.73 2413.14 Note. Elapsed Time (ET) is in 10s intervals. * p < .05, ** p < .01, *** p < .001, p < .10 Question 4B Summary of Results . In summary, while no gender effects on Overall Touch were observed, on average, toddlers with more expressive vocabulary displayed marginally less amounts of Overall Touch across the two - minute task, after each toddler dependent regulatory strategy (Question 4B) from a previous 10s interval was account ed for in 145 (within dyad variance) based on the amount of verbal distractions bids to mom, and amount of physical comfort seeking they displayed in a previous interval. As hypothesized, toddlers touched or attempted to touch the keys less often, when they also looking at mom; random - intercept random slope model). Additionally, as expected, toddlers touched or attempted to touch the keys less often, when they had sought physical comfort from mom (e.g., leaning on mom). Slope coefficients for both lagged initiated verbal distractions with Overall Touch and lagged physical comfort seeking with Overall Touch were the same. Unexpectedly, majority of lagged dependent strategies were not significantly associated with Overall Touch including lagged toddler joined verbal distraction, lagged toddler joined non - verbal distraction, lagged verbalizations about keys, lagged verbalizations of self - control, or lagged verbalizations of desire states. While these associations were not significant, distraction strategies were in the expected negative direction with Overall Touch scores, while verbalizations about key s and desire states were in unexpected positive direction, expect for lagged verbalizations in the form of self - control which was in the negative direction as expected. Summary of both 4A and 4B: In terms of Overall Touch (higher scores reflect lower abil ity to wait), results provide evidence for effectiveness of lagged i ndependent r egulatory s trategies and lagged d ependent r egulatory Strategies. However, these effects must be considered separately for distracting strategies, verbalizations toward the dela y task, and physical strategies. - verbal distractions were only non - verbal strategies. Similarly, e 146 not when directed toward mothers or Among independent verbalizations towards the d elay task, toddler - initiated talk that described the keys (e.g., Pooh) was marginally related to less Overall Touch as expected and toddler - initiated talk expressive of internal states (e.g., I need it) was not associated with Overall Touch and was in unex pected positive direction. However, compared to independent verbalizations toward delay task, among dependent verbalizations, talk about keys directed toward or in response to mother were no longer marginally significant and were in unexpected positive dir ection, verbalization of desire states directed toward or in response to mother remained non - significant and in unexpected positive direction, and verbalizations toward mother in the form of self - direction were in expected negative direction but not signif icant. Additionally, more physically self - comforting behaviors and comfort seeking behaviors were significantly associated with less Overall Touch. 147 CHAPTER 5: Discussion Given the critical role of self - regulation in developmental psychopathology (Cicch etti, Ackerman, & Izard, 1995), t he focus of this study was to examine the contributions of the examined the contributions of m aternal r egulatory a ttempts and m ind - m inde expression of negative emotion and delay of gratification. Questions 3 and 4 examined the i ndependent) and mother - dependent ( d ependent) r egulatory s egative emotion and delay of gratification. Overview Children exposed to early psychosocial stress (e.g., poverty, maternal mental illness) are at greater risk for poor regulatory development and related sequela (Blair & Raver, 2012; Raikes et al., 2007) . This study presents a first step in determination of effective parent - toddler co - regulation under a moment - to - moment lens in children from at risk families . The current results provide descriptive strength of each predictor (e.g., slope coefficient) and warrant future analyses examining interactions and combinations of predictors. Examination of interactions and combinations of predictors will provide more robust comparative evidence for effectiveness of each strategy , particularly when in reality strate gies co - occur. The results of the study provided evidence for within - individual and between - negative emotion and ability to wait as measured over 12 units of 10s intervals nested within 134 mother - toddler d yads. As hypothesized, l agged maternal positive emotional reactions (e.g., laughs with child) explained differences in expression of negative emotion (negative slope), ences in 148 ability to wait (negative slope). Within - individual differences in both expression of negative emotion and ability to wait were most strongly explained (positive slope) by lagged maternal negative emotional reactions (e.g., mother becomes upset, l lagged independent use of non - verbal distractions (e.g., looks away from delay task/object) and lagged mother - directed verbalizations of self - lower expression of negativ distractions (e.g., sings) and lagged mother - dependent physical comfort (e.g., reaching arms up to mother) were most strongly related higher ability to wait or DG. Study results did not provide ev idence for effects of lagged appropriate or non - attuned mind - related comments on expression of negative emotion or ability to wait. Boys expressed negative emotions with higher intensity compared to girls. Toddlers with more expressive language tended to t ouch the attractive keys less often compared to toddlers with lower scores on expressive language. No gender differences were found for ability to wait. Next I discuss results for unique effects of Maternal Regulatory Attempts, Mind - Related Comments and Delay of Gratification. Results are discussed for both emotion expression and ability to wait for each predictor in order to achieve a more cohesive presentation of study finding s. Unique Effects of Maternal Regulatory Attempts on Emotion Regulation Results of multilevel models confirmed study hypotheses for significant effects of Maternal Regulatory Attempts, specifically, lagged positive emotional reactions, lagged negative em otional reactions, lagged initiated distraction, and lagged physical restraint on both expression of negative emotion (emotional regulation) and overall touch (behavioral regulation). Although, findings confirmed study hypotheses for effects of lagged verb al comfort, punitive reactions, and 149 minimizing statements on emotional regulation, no effects were found for verbal comfort, Lagged Maternal Emotional Reactions . As hypothesized lagged posit ive emotional reactions predicted lower intensity of negative emotional expression, but lower ability to wait. As hypothesized, lagged negative emotional reactions were related to higher intensity of emotional expression and lower ability to wait. As previ ously mentioned, among all Maternal Regulatory Attempts, lagged emotional reactions were most strongly related to expression of negative emotion and ability to wait, providing support for the relational mechanisms underlying emotional and behavioral regula tion (Cole, 2014). Positive emotional reactions may be reflective of a larger supportive parenting construct that has been linked to optimal emotion regulation (Bocknek, Brophy - Herb, & Banerjee, 2009) and delay of gratification across toddler years (Brophy - expression by modeling adaptive emotional expressivity and scaffolding positive emotional states (Eisenberg et a l., 1998; Morris et al., 2007). Conversely, negative emotional reactions, such as laughing at or becoming angry with an impatient toddler in distress, may be reflective of - related strategies that regulate affect and organize motivational states and behavior. For instance, maternal negative emotional reactions reflect misalignment of intersubjective states (e.g., continues to display positive affect when toddler is crying), while frightening behaviors (e.g., harshly yells at to ddler for reaching for keys) communicate that the mother, herself, is a source may be disorganizing and frightening and, in turn, ineffective in reducing hei ghtened states of 150 misaligned parent - child interactions are predictive of disorganized attachment patterns characterized by disorganization and dysregulation of emotional and behavioral states, especially heightened emotional states (Lyons - Ruth, Bronfman, & Parsons, 1999; Madigan, Moran, Schuengel, Pederson, & Otten, 2007). Surprisingly, as mothers expressed or shared positive nally increased in subsequent intervals suggesting that perhaps toddlers in the current sample interpreted maternal positive affect as permission to touch the keys. Lagged Maternal Punitive Reactions and Minimizing . Similar to effects of lagged negative em otional reactions, lagged punitive and minimizing strategies reflective of a negative parent - child relationship contributed to higher intensity of negative emotion expression consistent with previous findings (Feldman et al., 2011), but did not exert main effects on ability to wait. Punitive reactions (e.g., lead to heightened states of arousal that may be frightening and disorganizing. Minimizing statements encourage ch may compromise development of physiological regulatory mechanisms by activating the sympathetic nervous system (Gross, 2015). In support of this, previous findings have linked both punitive and minimizing strategies to lower ability to regulate emotiona l distress and delay gratification across early childhood, which in turn, contribute to social and behavioral problems reflective of deficits in regulatory skills (Eisenberg et al., 2010b; Spinrad et al., 2007b). Null effects of punitive and minimizing on Overall Touch may be related to a number of factors. For one, punitive and minimizing strategies were rarely observed and may only be sufficient to explain variation in emotional intensity (recall that there was more within and between - dyad variation for E NE than Overall Touch). Additionally, while some toddlers may adaptively comply with a punitive and minimizing 151 mother, others may be too emotionally distressed to reach for or touch the keys, all resulting in zero observed instances of Overall Touch. This methodological challenge is further discussed in future directions. Lagged Maternal Verbal Comfort . Contrary to my hypotheses, as mothers provided lagged verbal comfort did not appear to affect ability to wait in the current interval. These findings were not expected because verbally comforting statements, reflective of warm and supportive emotion - related parenting practices, promote competence in emotional and b ehavioral regulation (Eisenberg et al., 1998). The positive associations between verbally comforting serval factors. For one, statements that acknowledged the c express or continue to express frustration. Second, statements that reassured the child he or she would get the keys y be facilitating expression of negative emotion. Additionally, it is also likely that toddlers who expressed more intense anger for longer durations elicited more verbally comforting statements from mothers, underscoring the need for future examinations of bidirectional processes not examined in the current study. Furthermore, it may be that comforting statements may not reduce intensity of anger expression but may be more useful to regulate fear or pain in older children. In a study of 3 - 12 year old canc er patients, children who were offered verbally comforting statements, experienced lower pain and distress related to treatment procedures, likely because 152 verbally comforting statements assure the child that he or she is not alone in experiencing their dis forting statements without offering alternative ways to cope with waiting - induced frustration is not enough to regulate heightened arousal in toddlers but may be effective for older children who have gained alternative regulatory skills as they have transi tioned to self - regulation. Lagged Physical Comfort and Physical Restraint . Findings confirmed study but results did not support effects of lagged physical although lagged physical comfort was positively correlated with ENE - O. Similarly, Grolnick et expression of distress du physically soothing behaviors such as hugging or kissing the toddler may facilitate behavioral compliance; they do not necessarily regulate emotional arousal but may facilitate express ion of frustration related to waiting as evidenced by positive correlations between lagged physical comfort and ENE - O. Furthermore, effects of physical comfort differed among dyads suggesting the need to examine differences in relational characteristics th at would moderate the effects of physical comfort. For instance, toddlers may not equally find physical touch soothing based on their attachment status with their mothers. Additionally, higher use of physical restraint was related to lower ability to regul ate both emotion and behavior. Physically restraining behaviors may reflect intrusive parenting practices and as suggested by the current findings may not 153 develop mental period in which emerging self - control necessitates parental support for autonomy (Kopp, 1982). Lagged Positive and Negative Commands. Finally, marginal effects were found for e, only lagged positive contribute to intensity of emotional expression. Similarly, Putnam et al., (2002) found that in a delay of gratification task for toddlers. These findings suggest that the use of positive commands, such as asking a toddler to wait or reminding them about rules of the task without ave to sit down) may hardly be a useful strategy in helping them comply. Also, positive commands do not help toddlers regulate negative emotions while waiting and might even be slightly annoying as evidenced by marginally positive effects on ENE - O. However , the use of negative commands is neither helpful in scaffolding compliance nor does it regulate intensity of anger. In fact, higher use of negative commands was correlated with higher to more intense expressions of anger. Lagged Maternal Distraction and Verbal Orientation to Delay. Results of multilevel analyses did not provide evidence for two Maternal Regulatory Attempts including lagged joined distraction and lagged verbal orientati on to delay. It should be noted that lagged maternal joined distraction occurred with very little frequency per interval, and, the lack of observed effects may be explained by small amount of variation in these strategies per interval. In fact, while toddl ers initiated either verbal or non - verbal distractions not directed toward mothers (range = 0 - 6), and, verbal distractions specifically directed toward mothers (range = 0 - 4), maternal joined 154 distractions only occurred 0 - 2 times per interval suggesting that mothers were not joining toddlers in their initiated distractions all that often. Although study hypotheses were not confirmed in multilevel models for effects of lagged joined distraction, it appears that as mothers s (e.g., toddler starts counting her toes and mother joins in counting toes), toddlers expressed lower intensity of negative emotions and less overall touch as suggested by negative slopes and significant negative correlations. These findings are consisten t with previous studies in which mothers initiated more attention shifting strategies compared to years old (Morris et al., 2011). As hypothesized, lagged mate rnal initiated distractions (e.g., engaging child in non - task related activity) predicted lower anger expression and higher ability to wait suggesting that distractions are useful in immediate regulation of emotional distress and may facilitate compliance in the short term. Similarly, Putnam et al. (2002) found that maternal the effectiveness of distraction as an adaptive regulatory strategy (Grolnick et al., 1998, Spinrad et al., 2004), long - - regulation remain to be understood. In the current study, lagged maternal initiated distraction was most strongly related lated to lower levels of overall expression of negative emotion. Thus, distraction may be an adaptive short term strategy for behavioral regulation, specifically ability to wait (Cole et al., 2011), but may not contribute in the same way or magnitude to em gratification. Similarly, orientation to delay occurred very infrequently per interval and, as evidenced by significant negative correlations, regulatory effects of describing or reframing the 155 task may have been more pronounced for current interval expression of negative emotion discussed further in future directions below. Unique Effects of Mind - Mindedness on Emotion Regulation and Delay of Gratification Contrary to my hypotheses, lagged mind - related comments did not contribute to differences in s may be explained and interpreted in several ways. Null results are discussed separately for appropriate and for non - attune mind - related comments. Appropriate Mind - Related Comments. To date, studies that provide evidence for effects of appropriate mind - r elated comments on developmental outcomes have often assessed mind - mindedness in the first and second years of life , while social, emotional, and behavioral outcomes have been assessed in 3 - 10 year olds . Thus, suggesting that effects of appropriate mind - re lated comments may be expected to manifest as development unfolds over time . For instance, exposure to appropriate mind - related comments in infancy contributes to mechanisms underlying emotional expression and delay of gratification , such as effortful cont rol (Eisenberg et al., 2010a; Spinrad et al., 2007b), in the third and fourth years of life (Bernier et al., 2017; Gagne et al., 2017). Additionally, exposure to appropriate mind - related comments in infancy contribute to lower behavioral problems reflectiv e of deficits in emotional and behavioral regulation (e.g., impulsivity) in 4 - 5 year olds (Meins et al., 20 13), and 10 year olds (Centifanti et al., 2016). In our own work, we have found that maternal representational mind - mindedness and emotion bridging, reflective of a larger mentalization - related parenting construct measured in 18 - 35 month - strategies 6 months later (Senehi et al., 2018). Therefore, it seems reasonab le that effects of appropriate mind - related comments on emotion expression or the ability to wait may not be 156 sustained immediately (as suggested by current findings) . Instead, a ppropriate mind - related comments may be expected to be consequential for regula tory competence over time through maturation of multiple mechanisms that underlie variations in emotion expression and delay of gratification (e.g., effortful control, generation of effective regulatory strategies, ment al state talk/understanding). First , development of effortful control undergoes rapid maturational shifts in the second and third years of life (Eisenberg et al., 2010a; Spinrad et al., 2007a). Also, as children enter the third and fourth years of life, they become markedly more effective (c ompared to 18 - 24 month olds) in generation of regulatory strategies (e.g. distraction) to regulate anger related to waiting (Cole et al., 2011). In fact, subcomponents of self - regulation including emotional, attentional, and behavioral control continue to mature and do not become fully integrated until the fourth and fifth years of life (Bell & Deater - Deckard, 2007; Bridgett et al., 2015; Cole et al., 2011; McClelland & Cameron, 2012). Secondly, as previously mentioned, exposure to mind - related comments fa cilitate identification, understanding, and interpretation of mental states in themselves and (Taumoepeau & Ruffman 2008). In turn, t hese skills are used to generate adaptive regulatory strategies (e.g., bids to caregiver) that help regulate emotion, attention and behavior (Eisenberg et al., 1998; Cole et al., 2010). Similar to maturation of effortful control and self - regulation, mental state understa nding and mental state language develop over time. For instance, children understand desire states s beliefs or emotional states (Wellman & Liu, 2004). Additionally, longitudinal findings (Becker - Razu ri, Hiles - Howa rd, Purvis, & Cross, 2017) 157 second year of life and significantly increase in third and fourth years with consistently more use of desire words (e.g., want) tha n cognitive words (e.g., kno w). Previous studies that have assessed mind - related comments during play or book sharing task have similarly found that mind - related comments occur with low frequency (Brophy - Herb et al., 2015; Meins et al., 2001). In spite of small amounts of variation, mind - related comments still uniquely contribute to child outcomes (Brophy - Herb et al., 2015 ; Meins et al., 2001 ) . Thus, although appropriate mind - related comments may not contribute to expression and ability to wait immediately, accumulation of exposure to mind - related comments may exert effects over time. Non - Attuned Mind - Related Comments. To our knowledge, this is the first study to examine effects of non - attuned mind - emotion and delay of gr atification. It was expected that lagged non - attuned mind - related ile toddler is ) would be related to higher expression of negative emotion and lower ability to wait in current intervals. However, lagged non - attuned mind - related comments did not ession or ability to wait in current intervals. These null findings may be explained by small amount of variance in maternal mind - related comments per interval. But, these findings may also be clarified by significant fixed effects of current interval non - attuned mind - related comments on current - interval expression of negative emotion and delay of gratification, and, significant correlations between non - attuned mind - related comments with maternal and toddler regulatory strategies as discussed in future dire ctions. 158 Gratification Overall toddlers used distractions, verbalizations about the delay task, and physically comforting behaviors to regulate emotional intensity and a bility to wait. On average, toddlers most frequently used non - verbal strategies including self - comforting, physical comfort seeking, as well as initiated and joined non - verbal distraction, while they least frequently used verbal strategies such as verbal d initiated distractions. Additionally, toddlers displayed verbalization strategies about the delay task least frequently either toward self or toward mothers. Presentations of these strategie s confirm previous studies (Cole et al., 2011; Grolnick et al. 1996) that found similar variations in that emphasize the emergence of self - control during the 2 n d year of life as toddlers begin to acquire more autonomy in motor and language skills (Kopp, 1982, Spinrad et al., 2004 ; Premo & Kiel, 201 4 s were more strongly effective for regulation of emotions than they were for regulation of behavior. This finding may be explained by low amounts of between - ability to wait relative to high amount of between - ne regulation and delay of gratification appeared to be of equal magnitudes and directionality (based on slope coefficients), regardless of independently initiate d (not directed toward mothers) or mother - dependent strategy use (mother - initiated or directed toward mother). Notably, similar amounts of variations were observed for both autonomously initiated strategies not directed toward mother (independent regulator y strategies) and initiated and joined strategies that were 159 directed or in response to mother (dependent regulatory strategies). These similarities in effects and variations between independent and dependent regulatory strategies may be explained by Calkin findings, which suggest that the mere presence of a caregiver contributes to adaptive physiological regulation (e.g., suppression of respiratory sinus arrhythmia) in toddlers compared to tasks in which children are left to regulate independently. Given that the mothers were sitting next to the toddlers on the same mat and were not told to refrain from interacting with the toddlers, independent regulatory strategies are only - initiated and not in re sponse to mother. Therefore, it is strategies may differ in tasks where mothers are absent or directly asked not to interact with their toddlers. Distraction. regulatory strategy was associated with lower intensity and predominance of anger expression and higher ability to wait, consistent with previous findings (Cole et al., 2011). As previously discussed, use of distraction appears to be an adaptive short - term strategy for tolerating frustration while waiting . The current findings also suggest that verbal distractions help toddlers wait, but, do not help them tolerate frustration while waiting, unless they are joining their - verbal distractions (either initiated or joined) appear to help toddlers tolerate frustration (e.g., express lower intensity of negative sarily affect their ability to wait to the same extent. These findings point to different pathways (e.g. joint attention) for effects of ve rbal and non - verbal distraction. - Comfort and Comfort Seeking . Additionally, ability to self - comfort (e.g., thumb sucking) was related to lower intensity of overall 160 expression and higher ability to wait as hypothesized. However, toddlers comfort seeking behaviors (e.g., leans on mother) were effective in delaying gratification bu t exerted no effects - income families. This null finding may have been a result of cod ing challenges. Recall that toddlers received a code for comfort seeking if they leaned on their mothers. It is possible that we may have - were not necessarily lean ing for comfort but reflective of bodily movements common when one is attempting to regulate while waiting. Also, effects of comfort seeking may be more pronounced in fear induced tasks where environmental threats such as novelty activate attachment - relate d behaviors such as seeking physical proximity. . Consistent with previous predicted higher intensity of neg ative emotion (Gilliom et al., 2002; Grolnick et al., 1996), but lower overall touch, suggesting that as toddlers focused on the delay task, expression of frustration was harder to suppress, especially when this specific form of focus was enabling them to harder to suppress intensity of emotional expression but had no effects on ability to wait. ion of desire states and expression of negative emotion was allowed to vary between dyads, slopes were in expected negative directionality. These results suggest that as toddlers made more verbalizations expressive of their internal desire states, they ten ded to display higher intensity levels of negative emotion, but this finding was not true when verbalizations were in response to mother or directed 161 toward mother. In addition, using desire talk not directed toward mother was correlated with higher negativ e emotion intensity and lower ability to wait, while these associations disappeared when toddlers engaged their mothers in their expressions of desire states. These findings provide on that children use language to facilitate self - regulation by expressing needs to caregivers. This is further supported - expression of negative emotion, which only o ccurred in response to or directed toward mother. As toddlers verbalized statements reflective of self - toward mother, or responded with that verbalization to their mothers, they suppressed anger expression in - control appeared to be the most effective regulatory strategy among all regulatory strategies for both emotion and behavioral regulation. These findings further suggest that verbalization of desire s tates and self - control may only facilitate regulation in context of a positive parent - child relationship beyond strategies to regulate. Strengths and Limitations It should be noted that findings of the current study present preliminary models in the strategies on overall intensity of negative emotion expression and ability to wait. Given that I did not propose any interactional effects to explain variance in slopes of explanatory variables, each model only focused on the unique contribution of one re gulatory strategy at a time. However, in real time these strategies co - occurred. For example, a mother who initiated a distracting activity may have 162 also expressed positive affect while doing so. Such co - occurrences of socialization strategies warrant futu re analyses where multiple regulatory strategies are tested in a single model and point to future directions in modeling interactions between strategies and accounting for effects of more than one strategy in a single model. One of the strengths of the cu rrent study lies in the multilevel modeling approach utilized to answer study questions. First, the examination of effects of lagged strategies on current - interval ENE - O and Overall Touch adequately address the temporal contingencies embedded in the operat ionalization of emotion regulation (Cole et al., 2011). While majority of studies to date have provided cross - sectional and correlational evidence for links between maternal socialization - regu lation, the results of this study provide evidence for directionality of effects (e.g., ENE and DG follow maternal or toddler strategies). In line with previous work on effects of time - varying maternal behaviors on emotion regulation, significant effects of strategy use on emotion expression in the current study provide Maternal socialization efforts are often measured with interview or self - report methods concurrently or wi th large time gaps between predictors and outcomes. However, results of the currently employed longitudinal approach (10s repeated measurement occasions), highlight immediacy of effects (e.g., within 20s time frame) of observed moment - to - moment socializati on and regulation strategies. Furthermore, results of the currently employed multilevel approach (nesting repeated measurements in mother - toddler dyads), identified the amount of observed variation in overall expression of negative emotion and ability to w ait associated with time - varying influences (e.g., MRAs) above and beyond trait - like factors (e.g., gender). 163 Coding Related Challenges. paradigm, The Delay of Gratification Task, did produce suffici ent distress in toddlers. While around 10% of toddlers did not express any negative emotion at any point during the task, around 60% expressed mild levels of intensity, and 30 % of toddlers expressed moderate to severe levels of intensity at some point dur ing the task. Also, while minority (around 20% ) of toddlers did not touch or make any attempts to touch the keys at any point during the task, majority of toddlers (around 80%) either touched or attempted to touch the keys at some point during the task. One of the strength s of the present study is the way in which v ariables were coded in every 10s interval using extensive observational coding schemes. However, coding related challenges were present. For instance, maternal regulatory attempts and toddler strategies were not always mutually exclusive and could simultaneously co - occur. For instance, mothers could hug the toddler while praising him/her verbally, or mothers could hug the toddler while minimizing his/her expression of negative emo tion. One way to account for these overlaps is to examine effects of multiple regulatory attempts in a single model. Additionally, in some cases mind - want to count your between the effects of non - attuned mind - propriate mind - assess quantity and type of mind - related comments and need to be assessed across several parent - child interaction contexts (e.g., play), to help provide support for interval validity of 164 mind - related comments. Additionally, maternal Verbal Orientation to Delay included differences in the types of talk invo e reframing which has been associated with less anger expression in preschoolers (Morris et al., and verbal statements that demand affection from the child expression of negative emotion was coded using facial, vocal, and postural cues for both anger and sadness combined as negative emotion. The limited visual quality of the video - taped Delay of Gratification Task did not allow us to differentially code toddler anger from sadness, a distinction that has been made in previous research (Morris et al., 2011). However, it was not surpris ing that the majority of the displayed negative expressions reflected anger cues given that anger is elicited when a goal is blocked (Cole, 2014), as the attractive toy was blocked in the Delay of Gratification Task. Future steps should include examination of maternal regulatory strategies as related to expressions of sadness and fear. Additionally, coding Delay of Gratification was limited in the current study such that duration and latency to touch were not measured. Future Directions The findings in the current study point to several avenues for future research including (1) examination of self - regulation differentiating between emotion - related regulation and emotionality (2) examination of moderated effects of multiple lagged Maternal Regulatory Attempt 165 emotion and delay of gratification, (3) examination of significant fixed effects as moderated by - like charact eristics (e.g., gender), specifically, for effects of maternal and toddler strategies that varied over dyads, (4) examination and ability to wait on matern al regulatory strategies and mind - mindedness, (5) assessment of mind - mindedness and other mentalization related parenting behaviors (MRPBs) across multiple parent - child interactional contexts, and, examination of effects of MRPBs on emotion expression and regulatory strategies. First, results of the current study point to several implications regarding the construct of - regulation. While results of Intercept - Only Mode ls suggested that, within toddlers, both ENE - O and Overall Touch varied in similar amounts from other intervals, much more between - toddler variance was observed for ENE - O compared to Overall Touch suggesting that the ability to wait upon request is more eq ually uniform among toddlers, than the ability to equally happy about it . Additionally, the ability to wait is compromised in highly distressed toddlers (Calkins & Johnson, 1998). Results of curve estimations reflected this challenge in measurement of outcomes variables (see Appendix D). As time elapsed, the direction of ENE - O and Overall Touch followed opposite quadratic curves. While ENE - O levels started out low, peaked toward the middle of the task, and returned to lower levels toward the end of the task, amount of Overall Touch started out high, bottomed out toward the midd le of the task and 166 increased toward the end of the task, suggesting that having to wait elicits more intense negative emotions while touching the keys was associated with lower intensity of expression. Furthermore, lack of ability to wait does not necessar ily constitute low effortful control but may be related to less voluntary forms of control such as reactive overcontrol (e.g. fearfulness/shyness) or reactive undercontrol (e.g., impulsivity) (Spinrad, et al., 2007a). Accordingly a major challenge in the p resent study concern s difficulty in differentiating between lack of touch due to intentionally regulated behavior and lack of touch due to dysregulated emotional states. However, this concern was addressed via differentiation between emotion - related self - r and predominance of negative emotion) from its regulation (within - individual differences in ENE - O across intervals) as suggested by current theoretical models of emotion regulation (Eisenberg et al., 2010a, Eisenberg et al., 2005). Furthermore, results of the current study are limited in that I did not control for effects of lagged ENE - O or lagged Overall Touch on current - interval ENE - O and current - interval Overall Touch. Thus, I cannot say that significant maternal and toddlers strategies had effects above and beyond the effects of ENE - O and Overal l Touch in previous intervals. However, t his concern was addressed by controlling for both linear and quadratic elapsed time to account for effects of fatigue and correlated error terms in the outcome given that overall expression of negative emotion and ability to wait in two intervals closer together in time would be more correlated than two intervals farther apart. Additionally , given the bidirectional processes between expression of negative emotion and ability to wait, future in intercepts and slopes of overall expression of negativ e emotion and ability to wait. Future - 167 toddler differences in physiological indicators of adaptive regulation (e.g., suppression of respiratory sinus arrhythmia ). regulatory strategies to further clarify effectiveness of each strategy as they contribute to variations in overall expression and delay of gratification together in single multilevel models. Examination of interactional effects of significant predictors in one model will also help fine - tune interpretation of unique and joint contributions of regulatory strategies. Third, as evident by significant unexplained vari ance between dyads in their average intercept after all predictors were accounted for in each model, there is room to improve the current models. Specifically, future analysis should consider adding explanatory variables for models in which effects of lagg ed predictors varied over dyads (e.g., effects of lagged maternal insta nce, maternal ability to effectively scaffold motivational and emotional states during challenging tasks is compromised in depressed parents (Hoffman, Crnic, and Baker, 2006). Toddlers of depressed mothers are socialized with maladaptive regulation strateg ies (Premo & Kiel, 2015). Also the type of response and frequency with which parents model regulatory strategies is informed by beliefs about negative emotions (Gottman, Fainsilber Katz, & Hooven, 1996; Morris et al., 2007), which are in turn related to to - regulation (Senehi et al., 2018). Therefore, maternal psychosocial functioning such as maternal depression and their beliefs about negative emotions (e.g., approving or disapproving) are likely trait - like characteristics that can be 168 examined to explain slope differences for effects of maternal regulatory attempts such as minimizing a nd negative emotional reactions. Additionally, developmental models of emotion socialization point to examination of cultural variations in parental regulatory strategies as well as cultural variations in parental beliefs about emotional expressivity which inform parental emotion socialization practices (Friedlmeier, Corapci, & Cole, 2011; Halberstadt & Lozada, regulatory attempts (e.g. minimizing) and expression of negative emotion. gratification. For instance, toddlers with more frequent and quicker tendencies to display highly intense and long periods of distress when frustrated in response to delay of gratification are more likely to use aggression, and less likely to voluntarily shift attention from desirable objects (i.e., self - distractio n), or seek maternal support (Calkins & Johnson, 1998). Moreover, bidirectional delay of gratification need to be examined as each may elicit different matern al regulatory attempts and mind - related comments. For instance, Eisenberg et al., (2010c) has found that maternal supportive regulatory strategies are elicited for toddlers with higher effortful control. to effects of maternal mind - related directions. Notably, in the current study, very little variation was observed for both appropriate and non - attuned mind - rel ated comments per interval suggesting the need to examine effects of mind - related comments as assessed in parent - child play contexts which are likely to elicit sufficient variability in mind - related comments (Meins et al., 2001). In fact, maternal referenc es 169 to mental states occur less frequently in negative parent - child interactional contexts (e.g. conflict) compared to positive or neutral contexts (Howe, Rinaldi, & Recchia, 2010). Also, in our own work (Senehi et al., 2018), we have found that maternal re presentational mind - mindedness, is associated with higher effective coping strategies (e.g., self - soothing behaviors when distressed) and effortful control re flective of attentional and behavioral regulatory skills (e.g., attention shifting and inhibitory control). Thus, future directions should consider effects of maternal mind - mindedness, assessed in positive contexts (e.g., play), and other indicators of men talization - gratification. Additionally, maternal mind - mindedness, indicative of a larger mentalization - related ry strategies (Senehi et al., 2018) that support development of optimal emotional and behavioral regulatory competence over time (Brophy - Herb et al., 2015; Centifanti et al., 2016, Meins et al., 2013). Therefore, future studies need to examine direct (e.g. through which mind - related comments contribute to emotional and behavioral regulation over time. No tably, in the current sample, more appropriate mind - related comments, and fewer non - appropriate mind - - verbal and joined verbal distraction, more ver balizations in the - - soothing ability to wait (or both) in multilevel m odels. Also, negative associations were observed between 170 maternal appropriate mind - related comments with maternal negative commands, and, positive associations between maternal non - attuned mind - related comments with maternal minimizing statements and physi cal restraint. Moreover, correlational results revealed that more current - interval and lagged non - attuned mind - related comments were correlated with higher overall expression of negative emotion and greater ability to wait. Therefore, exploratory multilev el analyses were conducted on effects of current - interval non - attuned mind - related comments on study outcomes. Significant fixed effects revealed that for every unit increase in current - interval maternal non - attuned mind - intensity of negative emotional expression significantly increased by 0.21 and overall touch significantly decreased 0.18 units in an average interval. 37 These findings suggest that non - while todd ler clearly wants to play with the keys) may directly (and immediately) suppress with caregiver, but evidently, also activate higher levels of emotional arousal . These findings point to future analyses of current - interval effects of maternal behaviors, specifically, non - attuned mind - related comments. In conclusion, the current study offers empirical support for the effects of maternal regulatory attempts and to - to - moment differences in - regulation with in the parent - child relationship. Additionally, correlational evidence points to associations between 37 To explore bidirectional associations, I also tested effects of current - interval ENE - O and Overall Touch on current - interval non - attuned mind - related comments and found small estimates of .03 that were significant . 171 and ability to wait . These findings provide new directio ns in examination of nuanced contributions of the caregiving context to development of self - regulation as toddlers make the transition from other - regulated to self - regulated emotional and behavioral states. 172 APPENDICES 173 APPENDIX A Ma ternal Regulatory Attempts & Mind - Mindedness Overlap 174 To ensure mutual exclusivity and help with interpretation of effects, I coded all mind - related comments for overlap with any Maternal Regulatory Attempts (Table D1). Meins and Fernyhough (2012) note th at mind - related comments would be considered non - attuned comments if the child was actively engaged in attending to or playing with something else. As described previously, Maternal Distraction was coded when mother shifted attention (distracted) child by holding a conversation about non - task related topics or by pointing out objects in the - attuned mind - related comments could be classified as distractions, they wer e coded as non - attuned comments. Majority of Appropriate Mind - Related Comments did not overlap with any MRA (60.50%) while majority of Non - Attuned Mind - related Comments could also serve as of Appropriate Mind - behaviors while reframing the nature of the waiting t outlined by Meins & Fer nyhough (2015). Majority of mind - related comments identified, - 175 Table A 1 Percentages of Appropriate and Non - Attuned Comments Further Coded for MRA Overlap and Mental State Category Appropriate Mind - Related Comments Non - Attuned Mind - Related Comm ents MRA Overlap No Overlap 72 (60.50%) 23 (22.33%) Initiated Distraction 11 (9.24%) 76 (73.79%) Joined Distraction 22 (18.49%) 1 (0.97%) Cognitive Reframe 6 (5.04%) 0 Physical Comfort 3 (2.52%) 0 Minimizing 1 (0.84%) 0 Positive Command 0 (0. 00%) 1 (0.97%) To Experimenter 2 (1.68%) 0 Mental State Category Emotion 14 (11.76%) 3 (2.91%) Cognition 6 (5.04%) 1 (0.97%) Desire 99 (83.19%) 91 (88.35%) Physical 0 8 (7.77%) Criteria C 0 3 (2.91%) Total 119 103 Note. MRA = Maternal Regulato attending to Keys. Cognitive Reframe was coded as a type of Maternal Verbal Orientation to Delay 176 APPENDIX B Coder Training and Inter - Coder Reliability 177 Coder Training. Coders were instructed to familiarize themselves with the Coding System for Delay of Gratification Task Friedlmeier, Corapci, & Benga, 2015a ), Mind - Mindedness Coding Manual, version 2.2 (Meins & Fernyhough, 2015), Coding Sy stem for Delay of Gratification Task Child Behavior and Regulatory Attempts (Friedlmeier, Corapci, & Benga, 2015b), and Coding System for Delay of Gratification Task - Intensity of Emotion Expression (Friedlmeier, Corapci, & Benga 2015c) before the first t raining meeting. During the first training meeting, the experimenter reviewed the coding manuals with in order to ensure unbiased coding. Therefore, no coder was assigned both child and parent of the same dyad. All videos were transcribed verbatim for all verbal statements made by mother and child by one coder before coding began. Transcription was done in order to ensure accuracy in coding verbal regulatory strat egies made by mother and toddler and mind - mindedness given the guidelines in Meins & Fernyhough (2015). Initial Training Phase. During the initial training phase, coders were instructed to code a frustrating episode each. Coders were instructed to watch t he entire frustrating episode from start to end without coding. This was done to ensure familiarity with the video and more accuracy in coding. Each coder was then asked to watch each episode a second time in 10s intervals during which they coded for the v ariable they had been trained in and assigned. The frustrating episode started when the experimenter placed the keys in front of the parent - child each interval, all videos for all subjects were watched and assigned start and end times for each 10s interval by the first author. Each frustrating episode lasted 2 minutes, t otaling 120 seconds, 178 except for cases in which the experimenter ended the task due to extreme toddler distress that lasted for more than 30 seconds (N = 6). Initial training phase was conducted for 5% - 20% of cases depending on the variable being coded, all disagreements were discussed and finalized during training meetings. Discrepancies between coders were addressed until coders felt that they understood all coding criteria clearly. Baseline Training Phase. During the baseline training phase, 5% - 15% of cas es were coded by two independent coders to achieve inter - coder reliability of .8 or greater for all variables before coders could be considered reliable to code independently. All disagreements were discussed and finalized during weekly meetings. Before in dependent coding took place, a baseline reliability estimate of .8 or higher was achieved for all variables expect for Toddler Initiated Non - Verbal Distraction , toddler shifts focus away from keys for more than 3 seconds (e.g., looks away, walks out of the Toddler Joined Verbal Distraction , toddler joins in alternative verbal activity that mother has initiated (e.g., Toddler Joined Non - Verbal Di straction , child joins in alternative non - verbal activity that mother has initiated (e.g., mom initiates play and child joins in, mom points to a direction away from keys and child looks where ay be explained by the level of coding difficulty of these variables due to poor video quality and difficulty in assessment of eye direction. Independent Coding Phase. After establishing baseline reliability, coders were assigned independent coding. Durin g the independent coding phase, 35% of cases were randomly double coded for all variables in order to ensure consistent and accurate coding. The first author met 179 weekly with each pair of coders to discuss disagreements and assess inter - coder reliability. A ll disagreements were discussed and finalized during coding meetings. Inter - Coder Reliability A total of 4 undergraduate research assistants coded maternal variables, and a separate set of 3 undergraduate research assistants coded toddler variables. All v ariables were assigned to pairs of coders. All pairs were identified during training phases and were selected based on mastery of their understanding of conceptual and operational definitions of variables and higher inter - coder reliability estimates. Duri ng all inter - coder reliability meetings, questions regarding variables were clarified and inter - rater reliability was assessed. 180 APPENDIX C Equations 181 Research Question 1: The Random Intercept Model Predicting Expression of Negative Emotion from Lagg ed Maternal Regulatory Attempts and Mind - Mindedness. The models were specified with the following equations: Maternal Verbal Comfort Level - 1 Equation: ENE - O ij 0j V - Comf,j (V - Comf i - 1j ETL,j (ET ij ET 2 ,j (ET 2 ij Verb,j (Verb i - 1j ) + e ij Level - 2 Equations: 0j 00 Gender (Gender j ) + u 0j v - comf,j = V - Comf + u 1j ETL,j = ETL ETL,j = ETL 2 Verb,j = Verb Mixed Model: E NE - O ij 00 Gender (Gender j ) V - Comf (V - Comf i - 1j ETL (ET ij ET 2 (ET 2 ij Verb (Verb i - 1j ) + u 0j + e ij where 00 indicates the overall intercept, the grand mean of the ENE - O scores across all dyads when all predictors = 0 Gender (Gender j ) = the average slope between ENE - O and Gender times score of Gender in dyad j V - Comf (V - Comf i - 1j ) = the average slope between ENE - O and maternal lagged Verbal Comfort times score of maternal lagged Verbal Comfort in dyad j ETL (ET ij ) = the average slope between ENE - O and linear elapsed time times score of linear elapsed time in interval i in dyad j ET 2 (ET 2 ij ) = the average slope between ENE - O and quadratic elapsed time times score of quadratic elapsed time in interval i in dyad j Verb (Verb i - 1j ) = t he average slope between ENE - O and maternal verbosity times score of maternal verbosity (Total Comments) in interval i in dyad j u 0j = deviation for dyad j from overall intercept, e ij = deviation for interval i from its dyad j Maternal Physical Comfort Level - 1 Equation: ENE - O ij 0j Phy - Comf,j (Phy - Comf i - 1j ETL,j (ET ij ET 2 ,j (ET 2 ij ) +e ij Level - 2 Equations: 0j 00 Gender (Gender j ) + u 0j 182 Phy - Comf,j = Phy - Comf + u 1j ETL,j = ETL ETL,j = ETL 2 Mixed Model ENE - O ij 00 Gender (Gen der j ) Phy - Comf (Phy - Comf i - 1j ETL (ET ij ET 2 (ET 2 ij ) + u 0j + e ij where 00 indicates the overall intercept, the grand mean of the ENE - O scores across all dyads when all predictors = 0 Gender (Gender j ) = the average slope between ENE - O and Ge nder times score of Gender in dyad j Phy - Comf (Phy - Comf i - 1j ) = the average slope between ENE - O and maternal lagged Physical Comfort times score of maternal lagged Physical Comfort in dyad j ETL,j (ET ij ) = the average slope between ENE - O and linear ela psed time times score of linear elapsed time in interval i in dyad j ET 2 ,j (ET 2 ij ) = the average slope between ENE - O and quadratic elapsed time times score of quadratic elapsed time in interval i in dyad j u 0j = deviation for dyad j from overall intercep t, e ij = deviation for interval i from its dyad j Maternal Initiated Distraction Level - 1 Equation: ENE - O ij 0j M - INI - Dis ,j ( M - INI - Dis i - 1j ETL,j (ET ij ET 2 ,j (ET 2 ij ) +e ij Level - 2 Equations: 0j 00 Gender (Gender j ) + u 0 j M - INI - Dis ,j = M - INI - Dis + u 1j ETL,j = ETL ETL,j = ETL 2 Mixed Model ENE - O ij 00 Gender (Gender j ) M - INI - Dis ( M - INI - Dis - 1j ETL (ET ij ET 2 (ET 2 ij ) + u 0j + e ij where 00 indicates the overall intercept, the grand mean o f the ENE - O scores across all dyads when all predictors = 0 Gender (Gender j ) = the average slope between ENE - O and Gender times score of Gender in dyad j M - INI - Dis ( M - INI - Dis i - 1j ) = the average slope between ENE - O and maternal lagged Initiated Distrac tion times score of maternal lagged Initiated Distraction in dyad j ETL,j (ET ij ) = the average slope between ENE - O and linear elapsed time times score of linear elapsed time in interval i in dyad j ET 2 ,j (ET 2 ij ) = the average slope between ENE - O and qu adratic elapsed time times score of quadratic elapsed time in interval i in dyad j u 0j = deviation for dyad j from overall intercept, e ij = deviation for interval i from its dyad j 18 3 Maternal Joint Distraction Level - 1 Equation: ENE - O ij 0j M - JNT - D is ,j ( M - JNT - Dis i - 1j ETL,j (ET ij ET 2 ,j (ET 2 ij ) +e ij Level - 2 Equations: 0j 00 Gender (Gender j ) + u 0j M - JNT - Dis ,j = M - JNT - Dis + u 1j ETL,j = ETL ETL,j = ETL 2 Mixed Model ENE - O ij 00 Gender (Gender j ) M - JN T - Dis ( M - JNT - Dis - 1j ETL (ET ij ET 2 (ET 2 ij ) + u 0j + e ij where 00 indicates the overall intercept, the grand mean of the ENE - O scores across all dyads when all predictors = 0 Gender (Gender j ) = the average slope between ENE - O and Gender times sc ore of Gender in dyad j M - JNT - Dis ( M - JNT - Dis i - 1j ) = the average slope between ENE - O and maternal lagged Joined Distraction times score of maternal lagged Joined Distraction in dyad j ETL,j (ET ij ) = the average slope between ENE - O and linear elapsed tim e times score of linear elapsed time in interval i in dyad j ET 2 ,j (ET 2 ij ) = the average slope between ENE - O and quadratic elapsed time times score of quadratic elapsed time in interval i in dyad j u 0j = deviation for dyad j from overall intercept, e ij = deviation for interval i from its dyad j Maternal Verbal Orientation to Delay Level - 1 Equation: ENE - O ij 0j M - VO2D,j ( M - VO2D i - 1j ETL,j (ET ij ET 2 ,j (ET 2 ij Verb,j (Verb i - 1j ) + e ij Level - 2 Equations: 0j 00 Gender (Ge nder j ) + u 0j v - comf,j = M - VO2D + u 1j ETL,j = ETL ETL,j = ETL 2 Verb,j = Verb Mixed Model: ENE - O ij 00 Gender (Gender j ) M - VO2D ( M - VO2D i - 1j ETL (ET ij ET 2 (ET 2 ij Verb (Verb i - 1j ) + u 0j + e ij 184 where 00 indicates the ov erall intercept, the grand mean of the ENE - O scores across all dyads when all predictors = 0 Gender (Gender j ) = the average slope between ENE - O and Gender times score of Gender in dyad j M - VO2D ( M - VO2D i - 1j ) = the average slope between ENE - O and matern al lagged Verbal Orientation to Delay times score of maternal lagged Verbal Orientation to Delay in dyad j ETL (ET ij ) = the average slope between ENE - O and linear elapsed time times score of linear elapsed time in interval i in dyad j ET 2 (ET 2 ij ) = the average slope between ENE - O and quadratic elapsed time times score of quadratic elapsed time in interval i in dyad j Verb (Verb i - 1j ) = the average slope between ENE - O and maternal verbosity times score of maternal verbosity (Total Comments) in interval i in dyad j u 0j = deviation for dyad j from overall intercept, e ij = deviation for interval i from its dyad j Maternal Positive Commands Level - 1 Equation: ENE - O ij 0j Pos - Comd,j ( Pos - Comd i - 1j ETL,j (ET ij ET 2 ,j (ET 2 ij Verb,j (Verb i - 1j ) + e ij Level - 2 Equations: 0j 00 Gender (Gender j ) + u 0j Pos - Comd,j = Pos - Comd + u 1j ETL,j = ETL ETL,j = ETL 2 Verb,j = Verb Mixed Model: ENE - O ij 00 Gender (Gender j Pos - Comd ( Pos - Comd i - 1j ETL (ET ij ET 2 (ET 2 ij Verb (Verb i - 1j ) + u 0j + e ij where 00 indicates the overall intercept, the grand mean of the ENE - O scores across all dyads when all predictors = 0 Gender (Gender j ) = the average slope between ENE - O and Gender times score of Gender in dyad j Po s - Comd ( Pos - Comd i - 1j ) = the average slope between ENE - O and maternal lagged Positive Commands times score of maternal lagged Positive Commands in dyad j ETL (ET ij ) = the average slope between ENE - O and linear elapsed time times score of linear elapsed ti me in interval i in dyad j ET 2 (ET 2 ij ) = the average slope between ENE - O and quadratic elapsed time times score of quadratic elapsed time in interval i in dyad j Verb (Verb i - 1j ) = the average slope between ENE - O and maternal verbosity times score of mate rnal verbosity (Total Comments) in interval i in dyad j u 0j = deviation for dyad j from overall intercept, e ij = deviation for interval i from its dyad j 185 Maternal Negative Commands Level - 1 Equation: ENE - O ij 0j Neg - Comd,j ( Neg - Comd i - 1j ETL,j (ET ij ET 2 ,j (ET 2 ij ) + Verb,j (Verb i - 1j ) + e ij Level - 2 Equations: 0j 00 Gender (Gender j ) + u 0j Neg - Comd,j = Neg - Comd + u 1j ETL,j = ETL ETL,j = ETL 2 Verb,j = Verb Mixed Model: ENE - O ij 00 Gender (Gender j ) + Neg - Comd ( Neg - Comd i - 1j ETL (ET ij ET 2 (ET 2 ij Verb (Verb i - 1j ) + u 0j + e ij where 00 indicates the overall intercept, the grand mean of the ENE - O scores across all dyads when all predictors = 0 Gender (Gender j ) = the average slope between EN E - O and Gender times score of Gender in dyad j Neg - Comd ( Neg - Comd i - 1j ) = the average slope between ENE - O and maternal lagged Negative Commands times score of maternal lagged Negative Commands in dyad j ETL (ET ij ) = the average slope between ENE - O and l inear elapsed time times score of linear elapsed time in interval i in dyad j ET 2 (ET 2 ij ) = the average slope between ENE - O and quadratic elapsed time times score of quadratic elapsed time in interval i in dyad j Verb (Verb i - 1j ) = the average slope betw een ENE - O and maternal verbosity times score of maternal verbosity (Total Comments) in interval i in dyad j u 0j = deviation for dyad j from overall intercept, e ij = deviation for interval i from its dyad j Maternal Punitive Reactions Level - 1 Equation: ENE - O ij 0j Pun,j ( M - Pun i - 1j ETL,j (ET ij ET 2 ,j (ET 2 ij ) +e ij Level - 2 Equations: 0j 00 Gender (Gender j ) + u 0j Pun,j = Pun + u 1j ETL,j = ETL ETL,j = ETL 2 186 Mixed Model ENE - O ij 00 Gender (Gender j ) M - Pun ( M - Pu n - 1j ETL (ET ij ET 2 (ET 2 ij ) + u 0j + e ij where 00 indicates the overall intercept, the grand mean of the ENE - O scores across all dyads when all predictors = 0 Gender (Gender j ) = the average slope between ENE - O and Gender times score of Gender i n dyad j Pun,j ( M - Pun i - 1j ) = the average slope between ENE - O and maternal lagged Punitive Reactions times score of maternal lagged Punitive Reactions in dyad j ETL,j (ET ij ) = the average slope between ENE - O and linear elapsed time times score of linear elapsed time in interval i in dyad j ET 2 ,j (ET 2 ij ) = the average slope between ENE - O and quadratic elapsed time times score of quadratic elapsed time in interval i in dyad j u 0j = deviation for dyad j from overall intercept, e ij = deviation for interval i from its dyad j Maternal Minimizing Level - 1 Equation: ENE - O ij 0j MIN,j ( Min i - 1j ETL,j (ET ij ET 2 ,j (ET 2 ij ) + Verb,j (Verb i - 1j ) + e ij Level - 2 Equations: 0j 00 Gender (Gender j ) + u 0j MIN,j = Min + u 1j ETL,j = ETL ETL,j = ETL 2 Verb,j = Verb Mixed Model: ENE - O ij 00 Gender (Gender j ) Min (Min i - 1j ETL (ET ij ET 2 (ET 2 ij Verb (Verb i - 1j ) + u 0j + e ij where 00 indicates the overall intercept, the grand mean of the ENE - O scores across al l dyads when all predictors = 0 Gender (Gender j ) = the average slope between ENE - O and Gender times score of Gender in dyad j Min ( Min i - 1j ) = the average slope between ENE - O and maternal lagged Minimizing statements times score of maternal lagged Minim izing statements in dyad j ETL (ET ij ) = the average slope between ENE - O and linear elapsed time times score of linear elapsed time in interval i in dyad j ET 2 (ET 2 ij ) = the average slope between ENE - O and quadratic elapsed time times score of quadratic elapsed time in interval i in dyad j Verb (Verb i - 1j ) = the average slope between ENE - O and maternal verbosity times score of maternal verbosity (Total Comments) in interval i in dyad j 187 u 0j = deviation for dyad j from overall intercept, e ij = deviation for interval i from its dyad j Maternal Physical Restraint Level - 1 Equation: ENE - O ij 0j Phy - RST,j (Phy - RST i - 1j ETL,j (ET ij ET 2 ,j (ET 2 ij ) +e ij Level - 2 Equations: 0j 00 Gender (Gender j ) + u 0j Phy - RST,j = Phy - R ST + u 1j ETL,j = ETL ETL,j = ETL 2 Mixed Model ENE - O ij 00 Gender (Gender j ) Phy - RST (Phy - RST i - 1j ETL (ET ij ET 2 (ET 2 ij ) + u 0j + e ij where 00 indicates the overall intercept, the grand mean of the ENE - O scores across all dyads whe n all predictors = 0 Gender (Gender j ) = the average slope between ENE - O and Gender times score of Gender in dyad j Phy - RST (Phy - RST i - 1j ) = the average slope between ENE - O and maternal lagged Physical Restraint times score of maternal lagged Positive E motional Reactions in dyad j ETL,j (ET ij ) = the average slope between ENE - O and linear elapsed time times score of linear elapsed time in interval i in dyad j ET 2 ,j (ET 2 ij ) = the average slope between ENE - O and quadratic elapsed time times score of qua dratic elapsed time in interval i in dyad j u 0j = deviation for dyad j from overall intercept, e ij = deviation for interval i from its dyad j Maternal Positive Emotional Reactions Level - 1 Equation: ENE - O ij 0j Pos - Aff,j (Pos - Aff i - 1j ETL,j ( ET ij ET 2 ,j (ET 2 ij ) +e ij Level - 2 Equations: 0j 00 Gender (Gender j ) + u 0j Pos - Aff,j = Pos - Aff + u 1j ETL,j = ETL ETL,j = ETL 2 Mixed Model 188 ENE - O ij 00 Gender (Gender j ) Pos - Aff (Pos - Aff i - 1j ETL (ET ij ) + ET 2 (ET 2 ij ) + u 0j + e ij where 00 indicates the overall intercept, the grand mean of the ENE - O scores across all dyads when all predictors = 0 Gender (Gender j ) = the average slope between ENE - O and Gender times score of Gender in dyad j Pos - Aff (Pos - Aff i - 1j ) = the average slope between ENE - O and maternal lagged Positive Emotional Reactions times score of maternal lagged Positive Emotional Reactions in dyad j ETL,j (ET ij ) = the average slope between ENE - O and linear elapsed time times score of line ar elapsed time in interval i in dyad j ET 2 ,j (ET 2 ij ) = the average slope between ENE - O and quadratic elapsed time times score of quadratic elapsed time in interval i in dyad j u 0j = deviation for dyad j from overall intercept, e ij = deviation for interv al i from its dyad j Maternal Negative Emotional Reactions Level - 1 Equation: ENE - O ij 0j Neg - Aff,j (Neg - Aff i - 1j ETL,j (ET ij ET 2 ,j (ET 2 ij ) +e ij Level - 2 Equations: 0j 00 Gender (Gender j ) + u 0j Neg - Aff,j = Ne g - Aff + u 1j ETL,j = ETL ETL,j = ETL 2 Mixed Model ENE - O ij 00 Gender (Gender j ) Neg - Aff (Neg - Aff i - 1j ETL (ET ij ET 2 (ET 2 ij ) + u 0j + e ij where 00 indicates the overall intercept, the grand mean of the ENE - O scores across all dyads when all predictors = 0 Gender (Gender j ) = the average slope between ENE - O and Gender times score of Gender in dyad j Neg - Aff (Neg - Aff i - 1j ) = the average slope between ENE - O and maternal lagged Negative Emotional Reactions times score of maternal lagg ed Negative Emotional Reactions in dyad j ETL,j (ET ij ) = the average slope between ENE - O and linear elapsed time times score of linear elapsed time in interval i in dyad j ET 2 ,j (ET 2 ij ) = the average slope between ENE - O and quadratic elapsed time times score of quadratic elapsed time in interval i in dyad j u 0j = deviation for dyad j from overall intercept, e ij = deviation for interval i from its dyad j Maternal Appropriate Mind - Related Comments Level - 1 Equation: ENE - O ij 0j AMM,j (AMM i - 1j ) + ETL,j (ET ij ET 2 ,j (ET 2 ij Verb,j (Verb i - 1j ) + e ij 189 Level - 2 Equations: 0j 00 Gender (Gender j ) + u 0j AMM,j = AMM + u 1j ETL,j = ETL ETL,j = ETL 2 Verb,j = Verb Mixed Model: ENE - O ij 00 Gender (Gender j ) AMM (AMM i - 1j ETL (ET ij ET 2 (ET 2 ij Verb (Verb i - 1j ) + u 0j + e ij where 00 indicates the overall intercept, the grand mean of the ENE - O scores across all dyads when all predictors = 0 Gender (Gender j ) = the average slope between ENE - O and Gender t imes score of Gender in dyad j AMM (AMM i - 1j ) = the average slope between ENE - O and maternal lagged Appropriate Mind - Related Comments times score of maternal lagged Appropriate Mind - Related Comments in dyad j ETL (ET ij ) = the average slope between ENE - O and linear elapsed time times score of linear elapsed time in interval i in dyad j ET 2 (ET 2 ij ) = the average slope between ENE - O and quadratic elapsed time times score of quadratic elapsed time in interval i in dyad j Verb (Verb i - 1j ) = the average slope between ENE - O and maternal verbosity times score of maternal verbosity (Total Comments) in interval i in dyad j u 0j = deviation for dyad j from overall intercept, e ij = deviation for interval i from its dyad j Maternal Non - Attuned Mind - Related Comments Level - 1 Equation: ENE - O ij 0j NMM,j (NMM i - 1j ETL,j (ET ij ET 2 ,j (ET 2 ij Verb,j (Verb i - 1j ) + e ij Level - 2 Equations: 0j 00 Gender (Gender j ) + u 0j NMM,j = NMM + u 1j ETL,j = ETL ETL,j = ETL 2 Verb,j = Verb Mixed Model: ENE - O ij 00 Gender (Gender j ) NMM (NMM i - 1j ETL (ET ij ET 2 (ET 2 ij Verb (Verb i - 1j ) + u 0j + e ij 190 where 00 indicates the overall intercept, the grand mean of the ENE - O scores across all dyads when all predictors = 0 Gender (Gender j ) = the average slop e between ENE - O and Gender times score of Gender in dyad j NMM (NMM i - 1j ) = the average slope between ENE - O and maternal lagged Non - Attuned Mind - Related Comments times score of maternal lagged Non - Attuned Mind - Related Comments in dyad j ETL (ET ij ) = the a verage slope between ENE - O and linear elapsed time times score of linear elapsed time in interval i in dyad j ET 2 (ET 2 ij ) = the average slope between ENE - O and quadratic elapsed time times score of quadratic elapsed time in interval i in dyad j Verb (Ve rb i - 1j ) = the average slope between ENE - O and maternal verbosity times score of maternal verbosity (Total Comments) in interval i in dyad j u 0j = deviation for dyad j from overall intercept, e ij = deviation for interval i from its dyad j Research Questio n 2: The Random Intercept Model Predicting Expression of Negative The models were specified with the following equations: Toddler Initiated Verbal Distraction Level - 1 Equation: ENE - O ij 0j T - INI - VDis,j (T - INI - VDis i - 1j ETL,j (ET ij ET 2 ,j (ET 2 ij ) + e ij Level - 2 Equations: 0j 00 Gender (Gender j Vocab (Vocab j ) + u 0j T - INI - VDis,j = T - INI - VDis + u 1j ETL,j = ETL ETL,j = ETL 2 Mixed Model: E NE - O ij 00 Gender (Gender j ) Vocab (Vocab j T - INI - Vdis (T - INI - VDis i - 1j ) + ETL (ET ij ET 2 (ET 2 ij ) + u 0j + e ij where 00 indicates the overall intercept, the grand mean of the ENE - O scores across all dyads when all predictors = 0 Gender (Gender j ) = the average slope between ENE - O and Gender times score of Gender in dyad j Vocab (Vocab j ) = the average slope between ENE - O and Vocabulary Production times score of Vocabulary Production in dyad j T - INI - VDis (T - INI - VDis i - 1j ) = the average slope between ENE - O and lagged Toddler Initiated Verbal Distraction times score of lagged Toddler Initiated Verbal Distraction in dyad j 191 ETL (ET ij ) = the average slope between ENE - O and linear elapsed time times score of linear elapsed time in interval i in dyad j ET 2 (ET 2 ij ) = the average slope between ENE - O and quadratic elapsed time times score of quadratic elapsed time in interval i in dyad j u 0j = deviation for dyad j from overall intercept, e ij = deviation for interval i from its dyad j Toddler I nitiated Non - Verbal Distraction Level - 1 Equation: ENE - O ij 0j T - INI - nonVDis,j (T - INI - nonVDis i - 1j ETL,j (ET ij ET 2 ,j (ET 2 ij ) + e ij Level - 2 Equations: 0j 00 Gender (Gender j Vocab (Vocab j ) + u 0j T - INI - nonVDis, j = T - INI - nonVDis + u 1j ETL,j = ETL ETL,j = ETL 2 Mixed Model: ENE - O ij 00 Gender (Gender j ) Vocab (Vocab j T - INI - Vdis (T - INI - nonVDis i - 1j ) + ETL (ET ij ET 2 (ET 2 ij ) + u 0j + e ij where 00 indicates the overall intercept, the grand mean of the ENE - O scores across all dyads when all predictors = 0 Gender (Gender j ) = the average slope between ENE - O and Gender times score of Gender in dyad j Vocab (Vocab j ) = the average slope between ENE - O and Vocabulary Production times score of V ocabulary Production in dyad j T - INI - nonVDis (T - INI - nonVDis i - 1j ) = the average slope between ENE - O and lagged Toddler Initiated Non - Verbal Distraction times score of lagged Toddler Initiated Non - Verbal Distraction in dyad j ETL (ET ij ) = the average slo pe between ENE - O and linear elapsed time times score of linear elapsed time in interval i in dyad j ET 2 (ET 2 ij ) = the average slope between ENE - O and quadratic elapsed time times score of quadratic elapsed time in interval i in dyad j u 0j = deviation for dyad j from overall intercept, e ij = deviation for interval i from its dyad j Toddler Initiated Verbal Keys Level - 1 Equation: ENE - O ij 0j T - INI - Vkeys,j ( T - INI - Vkeys i - 1j ETL,j (ET ij ET 2 ,j (ET 2 ij ) + e ij Level - 2 Equations: 192 0j 00 Gender (Gender j Vocab (Vocab j ) + u 0j T - INI - Vkeys ,j = T - INI - Vkeys + u 1j ETL,j = ETL ETL,j = ETL 2 Mixed Model: ENE - O ij 00 Gender (Gender j ) Vocab (Vocab j T - INI - Vkeys (T - INI - Vkeys i - 1j ) + ETL (ET ij ET 2 (E T 2 ij ) + u 0j + e ij where 00 indicates the overall intercept, the grand mean of the ENE - O scores across all dyads when all predictors = 0 Gender (Gender j ) = the average slope between ENE - O and Gender times score of Gender in dyad j Vocab (Vocab j ) = the average slope between ENE - O and Vocabulary Production times score of Vocabulary Production in dyad j - INI - Vkeys (T - INI - Vkeys i - 1j ) = the average slope between ENE - O and lagged Toddler Initiated Verbal Keys times score of lagged Toddler Initiated Verba l Keys in dyad j ETL (ET ij ) = the average slope between ENE - O and linear elapsed time times score of linear elapsed time in interval i in dyad j ET 2 (ET 2 ij ) = the average slope between ENE - O and quadratic elapsed time times score of quadratic elapsed ti me in interval i in dyad j u 0j = deviation for dyad j from overall intercept, e ij = deviation for interval i from its dyad j Toddler Initiated Verbal Self - Control Level - 1 Equation: ENE - O ij 0j T - INI - Vctrl,j ( T - INI - Vctrl i - 1j ETL,j (ET ij ) + ET 2 ,j (ET 2 ij ) + e ij Level - 2 Equations: 0j 00 Gender (Gender j Vocab (Vocab j ) + u 0j T - INI - Vctrl ,j = T - INI - Vctrl + u 1j ETL,j = ETL ETL,j = ETL 2 Mixed Model: ENE - O ij 00 Gender (Gender j ) Vocab (Vocab j T - INI - Vctrl (T - INI - Vctrl i - 1j ) + ETL (ET ij ET 2 (ET 2 ij ) + u 0j + e ij where 00 indicates the overall intercept, the grand mean of the ENE - O scores across all dyads when all predictors = 0 Gender (Gender j ) = the average slope between ENE - O and Gender times score of Gender in dyad j Vocab (Vocab j ) = the average slope between ENE - O and Vocabulary Production times score of Vocabulary Production in dyad j 193 - INI - Vctrl (T - INI - Vctrl i - 1j ) = the average slope between ENE - O and lagged Toddler Initiated Verb al Self - Control times score of lagged Toddler Initiated Verbal Self - Control in dyad j ETL (ET ij ) = the average slope between ENE - O and linear elapsed time times score of linear elapsed time in interval i in dyad j ET 2 (ET 2 ij ) = the average slope between ENE - O and quadratic elapsed time times score of quadratic elapsed time in interval i in dyad j u 0j = deviation for dyad j from overall intercept, e ij = deviation for interval i from its dyad j Toddler Initiated Verbal Desire Level - 1 Equation: ENE - O ij 0j T - INI - Vdes,j ( T - INI - Vdes i - 1j ETL,j (ET ij ET 2 ,j (ET 2 ij ) + e ij Level - 2 Equations: 0j 00 Gender (Gender j Vocab (Vocab j ) + u 0j T - INI - Vctrl ,j = T - INI - Vctrl + u 1j ETL,j = ETL ETL,j = ETL 2 Mixed Model: ENE - O ij 00 Gender (Gender j ) Vocab (Vocab j T - INI - Vdes (T - INI - Vdes i - 1j ) + ETL (ET ij ET 2 (ET 2 ij ) + u 0j + e ij where 00 indicates the overall intercept, the grand mean of the ENE - O scores across all dyads when all predictors = 0 Gende r (Gender j ) = the average slope between ENE - O and Gender times score of Gender in dyad j Vocab (Vocab j ) = the average slope between ENE - O and Vocabulary Production times score of Vocabulary Production in dyad j - INI - Vdes (T - INI - Vdes i - 1j ) = the averag e slope between ENE - O and lagged Toddler Initiated Verbal Desire times score of lagged Toddler Initiated Verbal Desire in dyad j ETL (ET ij ) = the average slope between ENE - O and linear elapsed time times score of linear elapsed time in interval i in dyad j ET 2 (ET 2 ij ) = the average slope between ENE - O and quadratic elapsed time times score of quadratic elapsed time in interval i in dyad j u 0j = deviation for dyad j from overall intercept e ij = deviation for interval i from its dyad j Toddler Self - Comfo rt Level - 1 Equation: ENE - O ij 0j T - Comf,j ( T - Scomf i - 1j ETL,j (ET ij ET 2 ,j (ET 2 ij ) + e ij 194 Level - 2 Equations: 0j 00 Gender (Gender j Vocab (Vocab j ) + u 0j T - Comf,j = T - Scomf + u 1j ETL,j = ETL ETL,j = ETL 2 Mixed Model: ENE - O ij 00 Gender (Gender j ) Vocab (Vocab j T - Scomf (T - Scomf i - 1j ) + ETL (ET ij ET 2 (ET 2 ij ) + u 0j + e ij where 00 indicates the overall intercept, the grand mean of the ENE - O scores across all dyads when all predictors = 0 Gender (Gender j ) = the average slope between ENE - O and Gender times score of Gender in dyad j Vocab (Vocab j ) = the average slope between ENE - O and Vocabulary Production times score of Vocabulary Production in dyad j T - Scomf (T - Scomf i - 1j ) = the aver age slope between ENE - O and lagged Toddler Self - Comfort times score of lagged Toddler Self - Comfort in dyad j ETL (ET ij ) = the average slope between ENE - O and linear elapsed time times score of linear elapsed time in interval i in dyad j ET 2 (ET 2 ij ) = th e average slope between ENE - O and quadratic elapsed time times score of quadratic elapsed time in interval i in dyad j u 0j = deviation for dyad j from overall intercept, e ij = deviation for interval i from its dyad j Toddler Verbal Distraction - Bids to M om Level - 1 Equation: ENE - O ij 0j T - Vdis - 2M,j ( T - Vdis - 2M i - 1j ETL,j (ET ij ET 2 ,j (ET 2 ij ) + e ij Level - 2 Equations: 0j 00 Gender (Gender j Vocab (Vocab j ) + u 0j T - Vdis - 2M,j = T - Vdis - 2M + u 1j ETL,j = ETL ETL,j = ETL 2 Mixed Model: ENE - O ij 00 Gender (Gender j ) Vocab (Vocab j T - Vdis - 2M (T - Vdis - 2M i - 1j ) + ETL (ET ij ET 2 (ET 2 ij ) + u 0j + e ij where 00 indicates the overall intercept, the grand mean of the ENE - O scores across all dyads when all predictors = 0 195 Gender (Gend er j ) = the average slope between ENE - O and Gender times score of Gender in dyad j Vocab (Vocab j ) = the average slope between ENE - O and Vocabulary Production times score of Vocabulary Production in dyad j T - Vdis - 2M (T - Vdis - 2M i - 1j ) = the average slope b etween ENE - O and lagged Toddler Verbal Distraction - Bids to Mom times score of lagged Toddler Verbal Distraction - Bids to Mom in dyad j ETL (ET ij ) = the average slope between ENE - O and linear elapsed time times score of linear elapsed time in interval i in dyad j ET 2 (ET 2 ij ) = the average slope between ENE - O and quadratic elapsed time times score of quadratic elapsed time in interval i in dyad j u 0j = deviation for dyad j from overall intercept, e ij = deviation for interval i from its dyad j Toddler Join ed Verbal Distraction Level - 1 Equation: ENE - O ij 0j T - JNT - VDis,j ( T - JNT - VDis i - 1j ETL,j (ET ij ET 2 ,j (ET 2 ij ) + e ij Level - 2 Equations: 0j 00 Gender (Gender j Vocab (Vocab j ) + u 0j T - JNT - VDis,j = T - JNT - VDis + u 1j ETL,j = ETL ETL,j = ETL 2 Mixed Model: ENE - O ij 00 Gender (Gender j ) Vocab (Vocab j T - JNT - VDis (T - JNT - VDis i - 1j) + ETL (ET ij ET 2 (ET 2 ij ) + u 0j + e ij where 00 indicates the overall intercept, the grand mean of the ENE - O score s across all dyads when all predictors = 0 Gender (Gender j ) = the average slope between ENE - O and Gender times score of Gender in dyad j Vocab (Vocab j ) = the average slope between ENE - O and Vocabulary Production times score of Vocabulary Production in dyad j T - JNT - VDis (T - JNT - VDis i - 1j ) = the average slope between ENE - O and lagged Toddler Joined Verbal Distraction times score of lagged Joined Verbal Distraction in dyad j ETL (ET ij ) = the average slope between ENE - O and linear elapsed time times score of linear elapsed time in interval i in dyad j ET 2 (ET 2 ij ) = the average slope between ENE - O and quadratic elapsed time times score of quadratic elapsed time in interval i in dyad j u 0j = deviation for dyad j from overall intercept, e ij = deviation for interval i from its dyad j Toddler Joined Non - Verbal Distraction Level - 1 Equation: 196 ENE - O ij 0j T - JNT - nonVDis,j ( T - JNT - nonVDis i - 1j ETL,j (ET ij ET 2 ,j (ET 2 ij ) + e ij Level - 2 Equations: 0j 00 Gender (Gender j Vocab (Vo cab j ) + u 0j T - JNT - nonVDis,j = T - JNT - nonVDis + u 1j ETL,j = ETL ETL,j = ETL 2 Mixed Model: ENE - O ij 00 Gender (Gender j ) Vocab (Vocab j T - JNT - nonVDis (T - JNT - nonVDis i - 1j) + ETL (ET ij ET 2 (ET 2 ij ) + u 0j + e ij where 00 ind icates the overall intercept, the grand mean of the ENE - O scores across all dyads when all predictors = 0 Gender (Gender j ) = the average slope between ENE - O and Gender times score of Gender in dyad j Vocab (Vocab j ) = the average slope between ENE - O and Vocabulary Production times score of Vocabulary Production in dyad j T - JNT - nonVDis (T - JNT - nonVDis i - 1j ) = the average slope between ENE - O and lagged Toddler Joined Non - Verbal Distraction times score of lagged Toddler Joined Non - Verbal Distraction in dya d j ETL (ET ij ) = the average slope between ENE - O and linear elapsed time times score of linear elapsed time in interval i in dyad j ET 2 (ET 2 ij ) = the average slope between ENE - O and quadratic elapsed time times score of quadratic elapsed time in interva l i in dyad j u 0j = deviation for dyad j from overall intercept, e ij = deviation for interval i from its dyad j Toddler Verbal Keys Bids to Mom Level - 1 Equation: ENE - O ij 0j T - Vkeys - 2M ,j ( T - Vkeys - 2M i - 1j ETL,j (ET ij ET 2 ,j (ET 2 ij ) + e ij Level - 2 Equations: 0j 00 Gender (Gender j Vocab (Vocab j ) + u 0j T - Vkeys - 2M ,j = T - Vkeys - 2M + u 1j ETL,j = ETL ETL,j = ETL 2 Mixed Model: ENE - O ij 00 Gender (Gender j ) Vocab (Vocab j T - Vkeys - 2M (T - Vkeys - 2M i - 1j ) + 197 ETL (ET ij ET 2 (ET 2 ij ) + u 0j + e ij where 00 indicates the overall intercept, the grand mean of the ENE - O scores across all dyads when all predictors = 0 Gender (Gender j ) = the average slope between ENE - O and Gender times score of Gender in dyad j Vocab (Vocab j ) = the average slope between ENE - O and Vocabulary Production times score of Vocabulary Production in dyad j T - Vkeys - 2M (T - Vkeys - 2M i - 1j ) = the average slope between ENE - O and lagged Toddler Verbal Keys Bids to Mom times score of l agged Toddler Verbal Keys Bids to Mom in dyad j ETL (ET ij ) = the average slope between ENE - O and linear elapsed time times score of linear elapsed time in interval i in dyad j ET 2 (ET 2 ij ) = the average slope between ENE - O and quadratic elapsed time tim es score of quadratic elapsed time in interval i in dyad j u 0j = deviation for dyad j from overall intercept, e ij = deviation for interval i from its dyad j Verbal Self - Control - Bids to Mom Level - 1 Equation: ENE - O ij 0j T - Vctrl - 2M ,j ( T - Vctrl - 2M i - 1j ETL,j (ET ij ET 2 ,j (ET 2 ij ) + e ij Level - 2 Equations: 0j 00 Gender (Gender j Vocab (Vocab j ) + u 0j T - Vctrl - 2M ,j = T - Vctrl - 2M + u 1j ETL,j = ETL ETL,j = ETL 2 Mixed Model: ENE - O ij 00 Gender (Gender j ) Vocab (Vocab j T - Vctrl - 2M (T - Vctrl - 2M i - 1j ) + ETL (ET ij ET 2 (ET 2 ij ) + u 0j + e ij where 00 indicates the overall intercept, the grand mean of the ENE - O scores across all dyads when all predictors = 0 Gender (Gender j ) = the average slope between ENE - O and Gender times score of Gender in dyad j Vocab (Vocab j ) = the average slope between ENE - O and Vocabulary Production times score of Vocabulary Production in dyad j T - Vctrl - 2M (T - Vctrl - 2M i - 1j ) = the average slope between ENE - O and lagged Toddler Verbal Self - Control Bids to Mom times score of lagged Toddler Verbal Self - Control Bids to Mom in dyad j ETL (ET ij ) = the average slope between ENE - O and linear elapsed time times score of linear elapsed time in interval i in dyad j ET 2 (ET 2 ij ) = the average slope between ENE - O and quadratic elapsed time times score of quadratic elapsed time in interval i in dyad j u 0j = deviation for dyad j from overall intercept, e ij = deviation for interval i from its dyad j 198 Toddler Verbal Desire Bids to Mom L evel - 1 Equation: ENE - O ij 0j T - Vdes - 2M ,j ( T - Vdes - 2M i - 1j ETL,j (ET ij ET 2 ,j (ET 2 ij ) + e ij Level - 2 Equations: 0j 00 Gender (Gender j Vocab (Vocab j ) + u 0j T - Vdes - 2M ,j = T - Vdes - 2M + u 1j ETL,j = ETL ETL,j = ETL 2 Mixed Model: ENE - O ij 00 Gender (Gender j ) Vocab (Vocab j T - Vdes - 2M (T - Vdes - 2M i - 1j ) + ETL (ET ij ET 2 (ET 2 ij ) + u 0j + e ij where 00 indicates the overall intercept, the grand mean of the ENE - O scores across all dyads when all predictors = 0 Gender (Gender j ) = the average slope between ENE - O and Gender times score of Gender in dyad j Vocab (Vocab j ) = the average slope between ENE - O and Vocabulary Production times score of Vocabulary Production in dyad j T - Vdes - 2M (T - Vdes - 2M i - 1j ) = the average slope betw een ENE - O and lagged Toddler Verbal Desire Bids to Mom times score of lagged Toddler Verbal Desire Bids to Mom in dyad j ETL (ET ij ) = the average slope between ENE - O and linear elapsed time times score of linear elapsed time in interval i in dyad j ET 2 (ET 2 ij ) = the average slope between ENE - O and quadratic elapsed time times score of quadratic elapsed time in interval i in dyad j u 0j = deviation for dyad j from overall intercept, e ij = deviation for interval i from its dyad j Toddler Physical Comfort Seeking Level - 1 Equation: ENE - O ij 0j T - ComfSk ,j ( T - ComfSk i - 1j ETL,j (ET ij ET 2 ,j (ET 2 ij ) + e ij Level - 2 Equations: 0j 00 Gender (Gender j Vocab (Vocab j ) + u 0j T - ComfSk ,j = T - ComfSk + u 1j ETL,j = ETL ETL,j = ETL 2 199 Mixed Model: ENE - O ij = 00 Gender (Gender j ) Vocab (Vocab j T - ComfSk ( T - ComfSk i - 1j ) + ETL (ET ij ET 2 (ET 2 ij ) + u 0j + e ij where 00 indicates the overall intercept, the grand mean of the ENE - O scores across all dyads when all predictors = 0 Gender (Gender j ) = th e average slope between ENE - O and Gender times score of Gender in dyad j Vocab (Vocab j ) = the average slope between ENE - O and Vocabulary Production times score of Vocabulary Production in dyad j T - ComfSk ( T - ComfSk i - 1j ) = the average slope between ENE - O and lagged Toddler Physical Comfort Seeking times score of lagged Toddler Physical Comfort Seeking in dyad j ETL (ET ij ) = the average slope between ENE - O and linear elapsed time times score of linear elapsed time in interval i in dyad j ET 2 (ET 2 ij ) = th e average slope between ENE - O and quadratic elapsed time times score of quadratic elapsed time in interval i in dyad j u 0j = deviation for dyad j from overall intercept, e ij = deviation for interval i from its dyad j Research Question 3: The Random Interc ept Model Predicting Overall Touch from Lagged Maternal Regulatory Attempts and Mind - Mindedness. The models were specified with the following equations: Maternal Verbal Comfort Level - 1 Equation: Overall Touch ij 0j V - Comf,j (V - Comf i - 1j ETL,j (ET ij ET 2 ,j (ET 2 ij ) + Verb,j (Verb i - 1j ) + e ij Level - 2 Equations: 0j 00 Gender (Gender j ) + u 0j v - comf,j = V - Comf + u 1j ETL,j = ETL ETL,j = ETL 2 Verb,j = Verb Mixed Model: O verall Touch ij 00 Gender (Gender j ) V - Comf (V - Comf i - 1j ETL (ET ij ET 2 (ET 2 ij Verb (Verb i - 1j ) + u 0j + e ij where 00 indicates the overall intercept, the grand mean of the Overall Touch scores across all dyads when all predictors = 0 200 Ge nder (Gender j ) = the average slope between Overall Touch and Gender times score of Gender in dyad j V - Comf (V - Comf i - 1j ) = the average slope between Overall Touch and maternal lagged Verbal Comfort times score of maternal lagged Verbal Comfort in dyad j ETL (ET ij ) = the average slope between Overall Touch and linear elapsed time times score of linear elapsed time in interval i in dyad j ET 2 (ET 2 ij ) = the average slope between Overall Touch and quadratic elapsed time times score of quadratic elapsed ti me in interval i in dyad j Verb (Verb i - 1j ) = the average slope between Overall Touch and maternal verbosity times score of maternal verbosity (Total Comments) in interval i in dyad j u 0j = deviation for dyad j from overall intercept, e ij = deviation for i nterval i from its dyad j Maternal Physical Comfort Level - 1 Equation: Overall Touch ij 0j Phy - Comf,j (Phy - Comf i - 1j ETL,j (ET ij ET 2 ,j (ET 2 ij ) +e ij Level - 2 Equations: 0j 00 Gender (Gender j ) + u 0j Phy - Comf,j = Phy - Comf + u 1j ETL,j = ETL ETL,j = ETL 2 Mixed Model Overall Touch ij 00 Gender (Gender j ) Phy - Comf (Phy - Comf i - 1j ETL (ET ij ET 2 (ET 2 ij ) + u 0j + e ij where 00 indicates the overall intercept, the grand mean of the Overall Touch s cores across all dyads when all predictors = 0 Gender (Gender j ) = the average slope between Overall Touch and Gender times score of Gender in dyad j Phy - Comf (Phy - Comf i - 1j ) = the average slope between Overall Touch and maternal lagged Physical Comfort times score of maternal lagged Physical Comfort in dyad j ETL,j (ET ij ) = the average slope between Overall Touch and linear elapsed time times score of linear elapsed time in interval i in dyad j ET 2 ,j (ET 2 ij ) = the average slope between Overall Touch and quadratic elapsed time times score of quadratic elapsed time in interval i in dyad j u 0j = deviation for dyad j from overall intercept, e ij = deviation for interval i from its dyad j Maternal Initiated Distraction Level - 1 Equation: 201 Overall Touc h ij 0j M - INI - Dis ,j ( M - INI - Dis i - 1j ETL,j (ET ij ET 2 ,j (ET 2 ij ) +e ij Level - 2 Equations: 0j 00 Gender (Gender j ) + u 0j M - INI - Dis ,j = M - INI - Dis + u 1j ETL,j = ETL ETL,j = ETL 2 Mixed Model Overall Touch ij 00 + Gender (Gender j ) M - INI - Dis ( M - INI - Dis - 1j ETL (ET ij ET 2 (ET 2 ij ) + u 0j + e ij where 00 indicates the overall intercept, the grand mean of the Overall Touch scores across all dyads when all predictors = 0 Gender (Gender j ) = the average slop e between Overall Touch and Gender times score of Gender in dyad j M - INI - Dis ( M - INI - Dis i - 1j ) = the average slope between Overall Touch and maternal lagged Initiated Distraction times score of maternal lagged Initiated Distraction in dyad j ETL,j (ET ij ) = the average slope between Overall Touch and linear elapsed time times score of linear elapsed time in interval i in dyad j ET 2 ,j (ET 2 ij ) = the average slope between Overall Touch and quadratic elapsed time times score of quadratic elapsed time in inte rval i in dyad j u 0j = deviation for dyad j from overall intercept, e ij = deviation for interval i from its dyad j Maternal Joint Distraction Level - 1 Equation: Overall Touch ij 0j M - JNT - Dis ,j ( M - JNT - Dis i - 1j ETL,j (ET ij ET 2 ,j (ET 2 ij ) +e ij Level - 2 Equations: 0j 00 Gender (Gender j ) + u 0j M - JNT - Dis ,j = M - JNT - Dis + u 1j ETL,j = ETL ETL,j = ETL 2 Mixed Model Overall Touch ij 00 Gender (Gender j ) M - JNT - Dis ( M - JNT - Dis - 1j ETL (ET ij ET 2 (ET 2 ij ) + u 0j + e ij where 00 indicates the overall intercept, the grand mean of the Overall Touch scores across all dyads when all predictors = 0 202 Gender (Gender j ) = the average slope between Overall Touch and Gender times score of Gender in dyad j M - JNT - D is ( M - JNT - Dis i - 1j ) = the average slope between Overall Touch and maternal lagged Joined Distraction times score of maternal lagged Joined Distraction in dyad j ETL,j (ET ij ) = the average slope between Overall Touch and linear elapsed time times score of linear elapsed time in interval i in dyad j ET 2 ,j (ET 2 ij ) = the average slope between Overall Touch and quadratic elapsed time times score of quadratic elapsed time in interval i in dyad j u 0j = deviation for dyad j from overall intercept, e ij = deviatio n for interval i from its dyad j Maternal Verbal Orientation to Delay Level - 1 Equation: Overall Touch ij 0j M - VO2D,j ( M - VO2D i - 1j ETL,j (ET ij ET 2 ,j (ET 2 ij Verb,j (Verb i - 1j ) + e ij Level - 2 Equations: 0j 00 Gender (Ge nder j ) + u 0j v - comf,j = M - VO2D + u 1j ETL,j = ETL ETL,j = ETL 2 Verb,j = Verb Mixed Model: Overall Touch ij 00 Gender (Gender j ) M - VO2D ( M - VO2D i - 1j ETL (ET ij ET 2 (ET 2 ij ) + Verb (Verb i - 1j ) + u 0j + e ij where 00 indicate s the overall intercept, the grand mean of the Overall Touch scores across all dyads when all predictors = 0 Gender (Gender j ) = the average slope between Overall Touch and Gender times score of Gender in dyad j M - VO2D ( M - VO2D i - 1j ) = the average slope b etween Overall Touch and maternal lagged Verbal Orientation to Delay times score of maternal lagged Verbal Orientation to Delay in dyad j ETL (ET ij ) = the average slope between Overall Touch and linear elapsed time times score of linear elapsed time in in terval i in dyad j ET 2 (ET 2 ij ) = the average slope between Overall Touch and quadratic elapsed time times score of quadratic elapsed time in interval i in dyad j Verb (Verb i - 1j ) = the average slope between Overall Touch and maternal verbosity times scor e of maternal verbosity (Total Comments) in interval i in dyad j u 0j = deviation for dyad j from overall intercept, e ij = deviation for interval i from its dyad j Maternal Positive Commands 203 Level - 1 Equation: Overall Touch ij 0j Pos - Comd,j ( Pos - C omd i - 1j ETL,j (ET ij ET 2 ,j (ET 2 ij ) + Verb,j (Verb i - 1j ) + e ij Level - 2 Equations: 0j 00 Gender (Gender j ) + u 0j Pos - Comd,j = Pos - Comd + u 1j ETL,j = ETL ETL,j = ETL 2 Verb,j = Verb Mixed Model: Overall Touch ij 0 0 Gender (Gender j Pos - Comd ( Pos - Comd i - 1j ETL (ET ij ET 2 (ET 2 ij ) + Verb (Verb i - 1j ) + u 0j + e ij where 00 indicates the overall intercept, the grand mean of the ENE - O scores across all dyads when all predictors = 0 Gender (Gender j ) = the aver age slope between Overall Touch and Gender times score of Gender in dyad j Pos - Comd ( Pos - Comd i - 1j ) = the average slope between Overall Touch and maternal lagged Positive Commands times score of maternal lagged Positive Commands in dyad j ETL (ET ij ) = th e average slope between Overall Touch and linear elapsed time times score of linear elapsed time in interval i in dyad j ET 2 (ET 2 ij ) = the average slope between Overall Touch and quadratic elapsed time times score of quadratic elapsed time in interval i i n dyad j Verb (Verb i - 1j ) = the average slope between Overall Touch and maternal verbosity times score of maternal verbosity (Total Comments) in interval i in dyad j u 0j = deviation for dyad j from overall intercept, e ij = deviation for interval i from its dyad j Maternal Negative Commands Level - 1 Equation: Overall Touch ij 0j Neg - Comd,j ( Neg - Comd i - 1j ETL,j (ET ij ET 2 ,j (ET 2 ij ) + Verb,j (Verb i - 1j ) + e ij Level - 2 Equations: 0j 00 Gender (Gender j ) + u 0j Neg - Comd,j = Neg - Comd + u 1j ETL,j = ETL ETL,j = ETL 2 204 Verb,j = Verb Mixed Model: Overall Touch ij 00 Gender (Gender j ) Neg - Comd ( Neg - Comd i - 1j ETL (ET ij ET 2 (ET 2 ij ) + Verb (Verb i - 1j ) + u 0j + e ij where 00 indicates the overall intercept, the grand mean of the Overall Touch scores across all dyads when all predictors = 0 Gender (Gender j ) = the average slope between Overall Touch and Gender times score of Gender in dyad j Neg - Comd ( Neg - Comd i - 1j ) = the average slope between Overall Touch and maternal lagged Negative Commands times score of maternal lagged Negative Commands in dyad j ETL (ET ij ) = the average slope between Overall Touch and linear elapsed time times score of linear elapsed time in interval i in dyad j ET 2 (ET 2 ij ) = the a verage slope between Overall Touch and quadratic elapsed time times score of quadratic elapsed time in interval i in dyad j Verb (Verb i - 1j ) = the average slope between Overall Touch and maternal verbosity times score of maternal verbosity (Total Comments) in interval i in dyad j u 0j = deviation for dyad j from overall intercept, e ij = deviation for interval i from its dyad j Maternal Punitive Reactions Level - 1 Equation: Overall Touch ij 0j Pun,j ( M - Pun i - 1j ETL,j (ET ij ET 2 ,j (ET 2 ij ) +e ij Level - 2 Equations: 0j 00 Gender (Gender j ) + u 0j Pun,j = Pun + u 1j ETL,j = ETL ETL,j = ETL 2 Mixed Model Overall Touch ij 00 Gender (Gender j ) M - Pun ( M - Pun - 1j ETL (ET ij ET 2 (ET 2 ij ) + u 0j + e ij where 0 0 indicates the overall intercept, the grand mean of the Overall Touch scores across all dyads when all predictors = 0 Gender (Gender j ) = the average slope between Overall Touch and Gender times score of Gender in dyad j Pun,j ( M - Pun i - 1j ) = the average slope between Overall Touch and maternal lagged Punitive Reactions times score of maternal lagged Punitive Reactions in dyad j ETL,j (ET ij ) = the average slope between Overall Touch and linear elapsed time times score of linear elapsed time in interval i in dyad j 205 ET 2 ,j (ET 2 ij ) = the average slope between Overall Touch and quadratic elapsed time times score of quadratic elapsed time in interval i in dyad j u 0j = deviation for dyad j from overall intercept, e ij = deviation for interval i from its dyad j M aternal Minimizing Level - 1 Equation: Overall Touch ij 0j MIN,j ( Min i - 1j ETL,j (ET ij ET 2 ,j (ET 2 ij ) + Verb,j (Verbi - 1j) + e ij Level - 2 Equations: 0j 00 Gender (Gender j ) + u 0j MIN,j = Min + u 1j ETL,j = ETL ETL,j = ETL 2 Verb,j = Verb Mixed Model: Overall Touch ij 00 Gender (Gender j ) Min (Min i - 1j ETL (ET ij ) + ET 2 (ET 2 ij Verb (Verb i - 1j ) + u 0j + e ij where 00 indicates the overall intercept, the grand mean of the Overall Touch scores acr oss all dyads when all predictors = 0 Gender (Gender j ) = the average slope between Overall Touch and Gender times score of Gender in dyad j Min ( Min i - 1j ) = the average slope between Overall Touch and maternal lagged Minimizing statements times score of maternal lagged Minimizing statements in dyad j ETL (ET ij ) = the average slope between Overall Touch and linear elapsed time times score of linear elapsed time in interval i in dyad j ET 2 (ET 2 ij ) = the average slope between Overall Touch and quadratic e lapsed time times score of quadratic elapsed time in interval i in dyad j Verb (Verb i - 1j ) = the average slope between Overall Touch and maternal verbosity times score of maternal verbosity (Total Comments) in interval i in dyad j u 0j = deviation for dyad j from overall intercept, e ij = deviation for interval i from its dyad j Maternal Physical Restraint Level - 1 Equation: Overall Touch ij 0j Phy - RST,j (Phy - RST i - 1j ETL,j (ET ij ET 2 ,j (ET 2 ij ) +e ij Level - 2 Equations: 206 0j 00 + Gender (Gender j ) + u 0j Phy - RST,j = Phy - RST + u 1j ETL,j = ETL ETL,j = ETL 2 Mixed Model Overall Touch ij 00 Gender (Gender j ) Phy - RST (Phy - RST i - 1j ETL (ET ij ET 2 (ET 2 ij ) + u 0j + e ij where 00 indicates the overall inter cept, the grand mean of the Overall Touch scores across all dyads when all predictors = 0 Gender (Gender j ) = the average slope between Overall Touch and Gender times score of Gender in dyad j Phy - RST (Phy - RST i - 1j ) = the average slope between Overall T ouch and maternal lagged Physical Restraint times score of maternal lagged Positive Emotional Reactions in dyad j ETL,j (ET ij ) = the average slope between Overall Touch and linear elapsed time times score of linear elapsed time in interval i in dyad j ET 2 ,j (ET 2 ij ) = the average slope between Overall Touch and quadratic elapsed time times score of quadratic elapsed time in interval i in dyad j u 0j = deviation for dyad j from overall intercept, e ij = deviation for interval i from its dyad j Maternal Pos itive Emotional Reactions Level - 1 Equation: Overall Touch ij 0j Pos - Aff,j (Pos - Aff i - 1j ETL,j (ET ij ET 2 ,j (ET 2 ij ) +e ij Level - 2 Equations: 0j 00 Gender (Gender j ) + u 0j Pos - Aff,j = Pos - Aff + u 1j ETL,j = ETL ETL,j = ETL 2 Mixed Model Overall Touch ij 00 Gender (Gender j ) Pos - Aff (Pos - Aff i - 1j ETL (ET ij ET 2 (ET 2 ij ) + u 0j + e ij where 00 indicates the overall intercept, the grand mean of the Overall Touch scores across all dyads when all predictors = 0 Gender (Gender j ) = the average slope between O verall Touch and Gender times score of Gender in dyad j 207 Pos - Aff (Pos - Aff i - 1j ) = the average slope between Overall Touch and maternal lagged Positive Emotional Reactions times score of maternal lagged Positive Emotional Reactions in dyad j ETL,j (ET ij ) = the average slope between Overall Touch and linear elapsed time times score of linear elapsed time in interval i in dyad j ET 2 ,j (ET 2 ij ) = the average slope between Overall Touch and quadratic elapsed time times score of quadratic elapsed time in inte rval i in dyad j u 0j = deviation for dyad j from overall intercept, e ij = deviation for interval i from its dyad j Maternal Negative Emotional Reactions Level - 1 Equation: Overall Touch ij 0j Neg - Aff,j (Neg - Aff i - 1j ETL,j (ET ij ET 2 ,j (ET 2 i j ) +e ij Level - 2 Equations: 0j 00 Gender (Gender j ) + u 0j Neg - Aff,j = Neg - Aff + u 1j ETL,j = ETL ETL,j = ETL 2 Mixed Model Overall Touch ij 00 Gender (Gender j ) Neg - Aff (Neg - Aff i - 1j ETL (ET ij ET 2 (ET 2 i j ) + u 0j + e ij where 00 indicates the overall intercept, the grand mean of the Overall Touch scores across all dyads when all predictors = 0 Gender (Gender j ) = the average slope between Overall Touch and Gender times score of Gender in dyad j Neg - Aff (Neg - Aff i - 1j ) = the average slope between Overall Touch and maternal lagged Negative Emotional Reactions times score of maternal lagged Negative Emotional Reactions in dyad j ETL,j (ET ij ) = the average slope between Overall Touch and linear elapsed time times score of linear elapsed time in interval i in dyad j ET 2 ,j (ET 2 ij ) = the average slope between Overall Touch and quadratic elapsed time times score of quadratic elapsed time in interval i in dyad j u 0j = deviation for dyad j from overall intercept , e ij = deviation for interval i from its dyad j Maternal Appropriate Mind - Related Comments Level - 1 Equation: Overall Touch ij 0j AMM,j (AMM i - 1j ETL,j (ET ij ET 2 ,j (ET 2 ij Verb,j (Verb i - 1j ) + e ij Level - 2 Equations: 208 0j 00 Gender (Gender j ) + u 0j AMM,j = AMM + u 1j ETL,j = ETL ETL,j = ETL 2 Verb,j = Verb Mixed Model: Overall Touch ij 00 Gender (Gender j ) AMM (AMM i - 1j ETL (ET ij ET 2 (ET 2 ij Verb (Verb i - 1j ) + u 0j + e ij where 00 indicate s the overall intercept, the grand mean of the Overall Touch scores across all dyads when all predictors = 0 Gender (Gender j ) = the average slope between Overall Touch and Gender times score of Gender in dyad j AMM (AMM i - 1j ) = the average slope between Overall Touch and maternal lagged Appropriate Mind - Related Comments times score of maternal lagged Appropriate Mind - Related Comments in dyad j ETL (ET ij ) = the average slope between Overall Touch and linear elapsed time times score of linear elapsed time in interval i in dyad j ET 2 (ET 2 ij ) = the average slope between Overall Touch and quadratic elapsed time times score of quadratic elapsed time in interval i in dyad j Verb (Verb i - 1j ) = the average slope between Overall Touch and maternal verbosity times score of maternal verbosity (Total Comments) in interval i in dyad j u 0j = deviation for dyad j from overall intercept, e ij = deviation for interval i from its dyad j Maternal Non - Attuned Mind - Related Comments Level - 1 Equation: Overall Touch ij 0j NMM,j (NMM i - 1j ETL,j (ET ij ET 2 ,j (ET 2 ij Verb,j (Verb i - 1j ) + e ij Level - 2 Equations: 0j 00 Gender (Gender j ) + u 0j NMM,j = NMM + u 1j ETL,j = ETL ETL,j = ETL 2 Verb,j = Verb Mixed Model: Overall Touch ij 00 Gender (Gender j ) NMM (NMM i - 1j ) + ETL (ET ij ET 2 (ET 2 ij Verb (Verb i - 1j ) + u 0j + e ij 209 where 00 indicates the overall intercept, the grand mean of the Overall Touch scores across all dyads when all predictors = 0 Gender (Gender j ) = the av erage slope between Overall Touch and Gender times score of Gender in dyad j NMM (NMM i - 1j ) = the average slope between Overall Touch and maternal lagged Non - Attuned Mind - Related Comments times score of maternal lagged Non - Attuned Mind - Related Comments in dyad j ETL (ET ij ) = the average slope between Overall Touch and linear elapsed time times score of linear elapsed time in interval i in dyad j ET 2 (ET 2 ij ) = the average slope between Overall Touch and quadratic elapsed time times score of quadratic elap sed time in interval i in dyad j Verb (Verb i - 1j ) = the average slope between Overall Touch and maternal verbosity times score of maternal verbosity (Total Comments) in interval i in dyad j u 0j = deviation for dyad j from overall intercept, e ij = deviation for interval i from its dyad j Research Question 4: The Random Intercept Model Predicting Overall Touch from The models were specified with the following equations: Toddler Initiated Verbal Distraction Level - 1 Equation: Overall Touch ij 0j T - INI - VDis,j (T - INI - VDis i - 1j ETL,j (ET ij ET 2 ,j (ET 2 ij ) + e ij Level - 2 Equations: 0j 00 Gender (Gender j Vocab (Vocab j ) + u 0j T - INI - VDis,j = T - INI - VDis + u 1j ETL,j = ETL ETL,j = ETL 2 Mixed Model: O verall Touch ij 00 Gender (Gender j ) Vocab (Vocab j T - INI - Vdis (T - INI - VDis i - 1j ) + ETL (ET ij ET 2 (ET 2 ij ) + u 0j + e ij where 00 indicates the overall intercept, the grand mean of the Overall Touch scores across all dyads when all predicto rs = 0 Gender (Gender j ) = the average slope between Overall Touch and Gender times score of Gender in dyad j Vocab (Vocab j ) = the average slope between Overall Touch and Vocabulary Production times score of Vocabulary Production in dyad j 210 T - INI - VDis ( T - INI - VDis i - 1j ) = the average slope between Overall Touch and lagged Toddler Initiated Verbal Distraction times score of lagged Toddler Initiated Verbal Distraction in dyad j ETL (ET ij ) = the average slope between Overall Touch and linear elapsed time ti mes score of linear elapsed time in interval i in dyad j ET 2 (ET 2 ij ) = the average slope between Overall Touch and quadratic elapsed time times score of quadratic elapsed time in interval i in dyad j u 0j = deviation for dyad j from overall intercept, e ij = deviation for interval i from its dyad j Toddler Initiated Non - Verbal Distraction Level - 1 Equation: Overall Touch ij 0j T - INI - nonVDis,j (T - INI - nonVDis i - 1j ETL,j (ET ij ET 2 ,j (ET 2 ij ) + e ij Level - 2 Equations: 0j 00 G ender (Gender j Vocab (Vocab j ) + u 0j T - INI - nonVDis,j = T - INI - nonVDis + u 1j ETL,j = ETL ETL,j = ETL 2 Mixed Model: Overall Touch ij 00 Gender (Gender j ) Vocab (Vocab j T - INI - Vdis (T - INI - nonVDis i - 1j ) + ETL (ET ij ET 2 (E T 2 ij ) + u 0j + e ij where 00 indicates the overall intercept, the grand mean of the Overall Touch scores across all dyads when all predictors = 0 Gender (Gender j ) = the average slope between Overall Touch and Gender times score of Gender in dyad j Vocab (Vocab j ) = the average slope between Overall Touch and Vocabulary Production times score of Vocabulary Production in dyad j T - INI - nonVDis (T - INI - nonVDis i - 1j ) = the average slope between Overall Touch and lagged Toddler Initiated Non - Verbal Distraction times score of lagged Toddler Initiated Non - Verbal Distraction in dyad j ETL (ET ij ) = the average slope between Overall Touch and linear elapsed time times score of linear elapsed time in interval i in dyad j ET 2 (ET 2 ij ) = the average slope between Over all Touch and quadratic elapsed time times score of quadratic elapsed time in interval i in dyad j u 0j = deviation for dyad j from overall intercept, e ij = deviation for interval i from its dyad j Toddler Initiated Verbal Keys Level - 1 Equation: 211 Ove rall Touch ij 0j T - INI - Vkeys,j ( T - INI - Vkeys i - 1j ETL,j (ET ij ET 2 ,j (ET 2 ij ) + e ij Level - 2 Equations: 0j 00 Gender (Gender j Vocab (Vocab j ) + u 0j T - INI - Vkeys ,j = T - INI - Vkeys + u 1j ETL,j = ETL ETL,j = ETL 2 Mixed Model : Overall Touch ij 00 Gender (Gender j ) Vocab (Vocab j T - INI - Vkeys (T - INI - Vkeys i - 1j ) + ETL (ET ij ET 2 (ET 2 ij ) + u 0j + e ij where 00 indicates the overall intercept, the grand mean of the Overall Touch scores across all dyads when all pr edictors = 0 Gender (Gender j ) = the average slope between Overall Touch and Gender times score of Gender in dyad j Vocab (Vocab j ) = the average slope between Overall Touch and Vocabulary Production times score of Vocabulary Production in dyad j - INI - Vkeys (T - INI - Vkeys i - 1j ) = the average slope between Overall Touch and lagged Toddler Initiated Verbal Keys times score of lagged Toddler Initiated Verbal Keys in dyad j ETL (ET ij ) = the average slope between Overall Touch and linear elapsed time times s core of linear elapsed time in interval i in dyad j ET 2 (ET 2 ij ) = the average slope between Overall Touch and quadratic elapsed time times score of quadratic elapsed time in interval i in dyad j u 0j = deviation for dyad j from overall intercept, e ij = de viation for interval i from its dyad j Toddler Initiated Verbal Self - Control Level - 1 Equation: Overall Touch ij 0j T - INI - Vctrl,j ( T - INI - Vctrl i - 1j ETL,j (ET ij ET 2 ,j (ET 2 ij ) + e ij Level - 2 Equations: 0j 00 Gender (Gend er j Vocab (Vocab j ) + u 0j T - INI - Vctrl ,j = T - INI - Vctrl + u 1j ETL,j = ETL ETL,j = ETL 2 Mixed Model: Overall Touch ij 00 Gender (Gender j ) Vocab (Vocab j T - INI - Vctrl (T - INI - Vctrl i - 1j ) + ETL (ET ij ET 2 (ET 2 ij ) + u 0j + e i j 212 where 00 indicates the overall intercept, the grand mean of the Overall Touch scores across all dyads when all predictors = 0 Gender (Gender j ) = the average slope between Overall Touch and Gender times score of Gender in dyad j Vocab (Vocab j ) = the average slope between Overall Touch and Vocabulary Production times score of Vocabulary Production in dyad j - INI - Vctrl (T - INI - Vctrl i - 1j ) = the average slope between Overall Touch and lagged Toddler Initiated Verbal Self - Control times score of lagged T oddler Initiated Verbal Self - Control in dyad j ETL (ET ij ) = the average slope between Overall Touch and linear elapsed time times score of linear elapsed time in interval i in dyad j ET 2 (ET 2 ij ) = the average slope between Overall Touch and quadratic el apsed time times score of quadratic elapsed time in interval i in dyad j u 0j = deviation for dyad j from overall intercept, e ij = deviation for interval i from its dyad j Toddler Initiated Verbal Desire Level - 1 Equation: Overall Touch ij 0j T - INI - Vdes,j ( T - INI - Vdes i - 1j ETL,j (ET ij ET 2 ,j (ET 2 ij ) + e ij Level - 2 Equations: 0j 00 Gender (Gender j Vocab (Vocab j ) + u 0j T - INI - Vctrl ,j = T - INI - Vctrl + u 1j ETL,j = ETL ETL,j = ETL 2 Mixed Model: Overall Touc h ij 00 Gender (Gender j ) Vocab (Vocab j T - INI - Vdes (T - INI - Vdes i - 1j ) + ETL (ET ij ET 2 (ET 2 ij ) + u 0j + e ij where 00 indicates the overall intercept, the grand mean of the Overall Touch scores across all dyads when all predictors = 0 Gen der (Gender j ) = the average slope between Overall Touch and Gender times score of Gender in dyad j Vocab (Vocab j ) = the average slope between Overall Touch and Vocabulary Production times score of Vocabulary Production in dyad j - INI - Vdes (T - INI - Vdes i - 1j ) = the average slope between Overall Touch and lagged Toddler Initiated Verbal Desire times score of lagged Toddler Initiated Verbal Desire in dyad j ETL (ET ij ) = the average slope between Overall Touch and linear elapsed time times score of linear e lapsed time in interval i in dyad j ET 2 (ET 2 ij ) = the average slope between Overall Touch and quadratic elapsed time times score of quadratic elapsed time in interval i in dyad j 213 u 0j = deviation for dyad j from overall intercept, e ij = deviation for inte rval i from its dyad j Toddler Self - Comfort Level - 1 Equation: Overall Touch ij 0j T - Comf,j ( T - Scomf i - 1j ETL,j (ET ij ET 2 ,j (ET 2 ij ) + e ij Level - 2 Equations: 0j 00 Gender (Gender j Vocab (Vocab j ) + u 0j T - Co mf,j = T - Scomf + u 1j ETL,j = ETL ETL,j = ETL 2 Mixed Model: Overall Touch ij 00 Gender (Gender j ) Vocab (Vocab j T - Scomf (T - Scomf i - 1j ) + ETL (ET ij ET 2 (ET 2 ij ) + u 0j + e ij where 00 indicates the overall intercept, the grand mean of the Overall Touch scores across all dyads when all predictors = 0 Gender (Gender j ) = the average slope between Overall Touch and Gender times score of Gender in dyad j Vocab (Vocab j ) = the average slope between Overall Touch and Vocabulary Production times score of Vocabulary Production in dyad j T - Scomf (T - Scomf i - 1j ) = the average slope between Overall Touch and lagged Toddler Self - Comfort times score of lagged Toddler Self - Comfort in dyad j ETL (ET ij ) = the average slope between Overall Touch an d linear elapsed time times score of linear elapsed time in interval i in dyad j ET 2 (ET 2 ij ) = the average slope between Overall Touch and quadratic elapsed time times score of quadratic elapsed time in interval i in dyad j u 0j = deviation for dyad j fro m overall intercept, e ij = deviation for interval i from its dyad j Toddler Verbal Distraction - Bids to Mom Level - 1 Equation: Overall Touch ij 0j T - Vdis - 2M,j ( T - Vdis - 2M i - 1j ETL,j (ET ij ET 2 ,j (ET 2 ij ) + e ij Level - 2 Equations: Intercept: 0j 00 Gender (Gender j Vocab (Vocab j ) + u 0j T - Vdis - 2M,j = T - Vdis - 2M + u 1j ETL,j = ETL 214 ETL,j = ETL 2 Mixed Model: Overall Touch ij 00 Gender (Gender j ) Vocab (Vocab j T - Vdis - 2M (T - Vdis - 2M i - 1j ) + ETL (ET ij ET 2 (ET 2 ij ) + u 0j + e ij where 00 indicates the overall intercept, the grand mean of the Overall Touch scores across all dyads when all predictors = 0 Gender (Gender j ) = the average slope between Overall Touch and Gender times score of Gender in dyad j Voca b (Vocab j ) = the average slope between Overall Touch and Vocabulary Production times score of Vocabulary Production in dyad j T - Vdis - 2M (T - Vdis - 2M i - 1j ) = the average slope between Overall Touch and lagged Toddler Verbal Distraction - Bids to Mom times sc ore of lagged Toddler Verbal Distraction - Bids to Mom in dyad j ETL (ET ij ) = the average slope between Overall Touch and linear elapsed time times score of linear elapsed time in interval i in dyad j ET 2 (ET 2 ij ) = the average slope between Overall Touch and quadratic elapsed time times score of quadratic elapsed time in interval i in dyad j u 0j = deviation for dyad j from overall intercept, e ij = deviation for interval i from its dyad j Toddler Joined Verbal Distraction Level - 1 Equation: Overall Tou ch ij 0j T - JNT - VDis,j ( T - JNT - VDis i - 1j ETL,j (ET ij ET 2 ,j (ET 2 ij ) + e ij Level - 2 Equations: 0j 00 Gender (Gender j Vocab (Vocab j ) + u 0j T - JNT - VDis,j = T - JNT - VDis + u 1j ETL,j = ETL ETL,j = ETL 2 Mixed Model: Overall Touch ij 00 Gender (Gender j ) Vocab (Vocab j T - JNT - VDis (T - JNT - VDis i - 1j) + ETL (ET ij ET 2 (ET 2 ij ) + u 0j + e ij where 00 indicates the overall intercept, the grand mean of the Overall Touch scores across all dyads when all predict ors = 0 Gender (Gender j ) = the average slope between Overall Touch and Gender times score of Gender in dyad j Vocab (Vocab j ) = the average slope between Overall Touch and Vocabulary Production times score of Vocabulary Production in dyad j 215 T - JNT - VDis (T - JNT - VDis i - 1j ) = the average slope between Overall Touch and lagged Toddler Joined Verbal Distraction times score of lagged Joined Verbal Distraction in dyad j ETL (ET ij ) = the average slope between Overall Touch and linear elapsed time times score of linear elapsed time in interval i in dyad j ET 2 (ET 2 ij ) = the average slope between Overall Touch and quadratic elapsed time times score of quadratic elapsed time in interval i in dyad j u 0j = deviation for dyad j from overall intercept, e ij = deviation for interval i from its dyad j Toddler Joined Non - Verbal Distraction Level - 1 Equation: Overall Touch ij 0j T - JNT - nonVDis,j ( T - JNT - nonVDis i - 1j ETL,j (ET ij ET 2 ,j (ET 2 ij ) + e ij Level - 2 Equations: 0j 00 Gender (Gender j ) Vocab (Vocab j ) + u 0j T - JNT - nonVDis,j = T - JNT - nonVDis + u 1j ETL,j = ETL ETL,j = ETL 2 Mixed Model: Overall Touch ij 00 Gender (Gender j ) Vocab (Vocab j T - JNT - nonVDis (T - JNT - nonVDis i - 1j) + ETL (ET ij ET 2 (ET 2 ij ) + u 0j + e ij where 00 indicates the overall intercept, the grand mean of the Overall Touch scores across all dyads when all predictors = 0 Gender (Gender j ) = the average slope between Overall Touch and Gender times score of Gender in dyad j Vocab (Vocab j ) = the average slope between Overall Touch and Vocabulary Production times score of Vocabulary Production in dyad j T - JNT - nonVDis (T - JNT - nonVDis i - 1j ) = the average slope between Overall Touch and lagged Toddler Joined Non - Verbal Distraction times score of lagged Toddler Joined Non - Verbal Distraction in dyad j ETL (ET ij ) = the average slope between Overall Touch and linear elapsed time times score of linear elapsed time in interval i in dyad j ET 2 (ET 2 ij ) = the average slope between Overall Touch and quad ratic elapsed time times score of quadratic elapsed time in interval i in dyad j u 0j = deviation for dyad j from overall intercept, e ij = deviation for interval i from its dyad j Toddler Verbal Keys Bids to Mom Level - 1 Equation: 216 Overall Touch ij 0 j T - Vkeys - 2M ,j ( T - Vkeys - 2M i - 1j ETL,j (ET ij ET 2 ,j (ET 2 ij ) + e ij Level - 2 Equations: 0j 00 Gender (Gender j Vocab (Vocab j ) + u 0j T - Vkeys - 2M ,j = T - Vkeys - 2M + u 1j ETL,j = ETL ETL,j = ETL 2 Mixed Model: Overal l Touch ij 00 Gender (Gender j ) Vocab (Vocab j T - Vkeys - 2M (T - Vkeys - 2M i - 1j ) + ETL (ET ij ET 2 (ET 2 ij ) + u 0j + e ij where 00 indicates the overall intercept, the grand mean of the Overall Touch scores across all dyads when all predictors = 0 Gender (Gender j ) = the average slope between Overall Touch and Gender times score of Gender in dyad j Vocab (Vocab j ) = the average slope between Overall Touch and Vocabulary Production times score of Vocabulary Production in dyad j T - Vkeys - 2M (T - Vkeys - 2M i - 1j ) = the average slope between ENE - O and lagged Toddler Verbal Keys Bids to Mom times score of lagged Toddler Verbal Keys Bids to Mom in dyad j ETL (ET ij ) = the average slope between Overall Touch and linear elapsed time times score of linear e lapsed time in interval i in dyad j ET 2 (ET 2 ij ) = the average slope between Overall Touch and quadratic elapsed time times score of quadratic elapsed time in interval i in dyad j u 0j = deviation for dyad j from overall intercept, e ij = deviation for inter val i from its dyad j Verbal Self - Control - Bids to Mom Level - 1 Equation: Overall Touch ij 0j T - Vctrl - 2M ,j ( T - Vctrl - 2M i - 1j ETL,j (ET ij ET 2 ,j (ET 2 ij ) + e ij Level - 2 Equations: 0j 00 Gender (Gender j Vocab (Vocab j ) + u 0 j T - Vctrl - 2M ,j = T - Vctrl - 2M + u 1j ETL,j = ETL ETL,j = ETL 2 Mixed Model: Overall Touch ij 00 Gender (Gender j ) Vocab (Vocab j T - Vctrl - 2M (T - Vctrl - 2M i - 1j ) + ETL (ET ij ET 2 (ET 2 ij ) + u 0j + e ij 217 where 00 indicates the over all intercept, the grand mean of the Overall Touch scores across all dyads when all predictors = 0 Gender (Gender j ) = the average slope between Overall Touch and Gender times score of Gender in dyad j Vocab (Vocab j ) = the average slope between Overall T ouch and Vocabulary Production times score of Vocabulary Production in dyad j T - Vctrl - 2M (T - Vctrl - 2M i - 1j ) = the average slope between Overall Touch and lagged Toddler Verbal Self - Control Bids to Mom times score of lagged Toddler Verbal Self - Control B ids to Mom in dyad j ETL (ET ij ) = the average slope between Overall Touch and linear elapsed time times score of linear elapsed time in interval i in dyad j ET 2 (ET 2 ij ) = the average slope between Overall Touch and quadratic elapsed time times score of q uadratic elapsed time in interval i in dyad j u 0j = deviation for dyad j from overall intercept, e ij = deviation for interval i from its dyad j Toddler Verbal Desire Bids to Mom Level - 1 Equation: Overall Touch ij 0j T - Vdes - 2M ,j ( T - Vdes - 2M i - 1j ETL,j (ET ij ET 2 ,j (ET 2 ij ) + e ij Level - 2 Equations: 0j 00 Gender (Gender j Vocab (Vocab j ) + u 0j T - Vdes - 2M ,j = T - Vdes - 2M + u 1j ETL,j = ETL ETL,j = ETL 2 Mixed Model: Overall Touch ij 00 Gender (Gender j ) Vocab (Vocab j T - Vdes - 2M (T - Vdes - 2M i - 1j ) + ETL (ET ij ET 2 (ET 2 ij ) + u 0j + e ij where 00 indicates the overall intercept, the grand mean of the Overall Touch scores across all dyads when all predictors = 0 Gender (Gender j ) = the average slope between Overall Touch and Gender times score of Gender in dyad j Vocab (Vocab j ) = the average slope between Overall Touch and Vocabulary Production times score of Vocabulary Production in dyad j T - Vdes - 2M (T - Vdes - 2M i - 1j ) = the average slope between Ov erall Touch and lagged Toddler Verbal Desire Bids to Mom times score of lagged Toddler Verbal Desire Bids to Mom in dyad j ETL (ET ij ) = the average slope between Overall Touch and linear elapsed time times score of linear elapsed time in interval i in d yad j ET 2 (ET 2 ij ) = the average slope between Overall Touch and quadratic elapsed time times score of quadratic elapsed time in interval i in dyad j u 0j = deviation for dyad j from overall intercept, e ij = deviation for interval i from its dyad j 218 Toddler Physical Comfort Seeking Level - 1 Equation: Overall Touch ij 0j T - ComfSk ,j ( T - ComfSk i - 1j ETL,j (ET ij ET 2 ,j (ET 2 ij ) + e ij Level - 2 Equations: 0j 00 Gender (Gender j Vocab (Vocab j ) + u 0j T - ComfSk ,j = T - ComfSk + u 1j ETL,j = ETL ETL,j = ETL 2 Mixed Model: Overall Touch ij 00 Gender (Gender j ) Vocab (Vocab j T - ComfSk ( T - ComfSk i - 1j ) + ETL (ET ij ET 2 (ET 2 ij ) + u 0j + e ij where 00 indicates the overall intercept, the grand mean of the Overal l Touch scores across all dyads when all predictors = 0 Gender (Gender j ) = the average slope between Overall Touch and Gender times score of Gender in dyad j Vocab (Vocab j ) = the average slope between Overall Touch and Vocabulary Production times score of Vocabulary Production in dyad j T - ComfSk ( T - ComfSk i - 1j ) = the average slope between Overall Touch and lagged Toddler Physical Comfort Seeking times score of lagged Toddler Physical Comfort Seeking in dyad j ETL (ET ij ) = the average slope between Over all Touch and linear elapsed time times score of linear elapsed time in interval i in dyad j ET 2 (ET 2 ij ) = the average slope between Overall Touch and quadratic elapsed time times score of quadratic elapsed time in interval i in dyad j u 0j = deviation for dyad j from overall intercept, e ij = deviation for interval i from its dyad j 219 APPENDIX D Patterns of Study Variables a cross 12 Intervals 220 Toddler Expression of Negative Emotion and Delay of Gratification Patterns Across 12 Intervals. De emotion and delay of gratification to identify and describe patterns of change across all 12 measurement occasions. Overall Expression of Negative Emotion (ENE - O) scores increased over the 12 intervals, while Overall Touch scores decreased over the 12 intervals. Additionally, one - way analysis of variance (ANOVA) was conducted to test mean differences of ENE - O and Overall Touch as a function of elapsed intervals (i.e., Elapsed Time) (Tabl e D1). Means in ENE - O were not significantly different across 12 intervals but means in Overall Touch F (11, 1559) = 6.32, p = 0.000, 2 = . 04 were significantly different across 12 intervals (Table D1). To further examine patterns of ENE - O and Overall Touch across the 12 intervals, to address the intercept - only models descriptively, and examine the linearity of effects of time on outco mes when all predictors were zero, I conducted Curve Estimation to estimate regression statistics between Elapsed Time and all dependent variables using linear and quadratic curves (Figure D1). Analysis of variance (ANOVA) was used to test linear and quadr atic curves for goodness of fit for all dependent variables. Results revealed that the best fit linear models were not statistically significant for any of the outcome variables while the best fit quadratic models were significant for all dependent variabl es. These results revealed that in addition to a linear relationship, the relationship between Elapsed Time and ENE - O [ F (2, 1568) = 4.05, p = 0.02, r 2 Negative Emotion followed an inverse U - shaped pattern by starting off with lower levels of intensity expression in the first half of the task, reached peak intensity towards the middle of the task, and returned to lower levels towards the end of the task. Additiona lly, compared to a linear model, a quadratic model was used to describe the pattern of observed scores for Overall Touch 221 [ F (2, 1568) = 31.53, p = 0.000, r 2 = .04] . As Elapsed Time increased, on average toddlers followed a U - Shaped curve in which they star ted off with higher amounts of Touch, Attempted Touch, and Overall Touch, followed by a decrease in Touch, Attempted Touch, and Overall Touch in the middle of the task, and a return to higher amounts of Touch, Attempted Touch, and Overall Touch towards the end of the task (Figure D1). Thus, Elapsed Time (Linear) and Elapsed Time 2 (Quadratic) were used as time - related predictor variables in all multilevel analyses. 222 Table D1 Analysis of Variance (ANOVA) Between Toddler Expression of Negative Emotion and Ove rall Touch with Elapsed Time Across 12 Intervals Variable Name ENE - O Overall Touch M (SD) M (SD) Interval 1 0.51 (0.74) 0.58 (0.65) Interval 2 0.81 (0.98) 0.41 (0.76) Interval 3 0.82 (1.00) 0.35 (0.70) Interval 4 0.88 (1.02) 0.34 (0.77) Interval 5 0.85 (1.04) 0.19 (0.44) Interval 6 0.74 (1.04) 0.26 (0.59) Interval 7 0.87 (1.06) 0.17 (0.67) Interval 8 0.79 (1.04) 0.18 (0.54) Interval 9 0.78 (1.00) 0.12 (0.35) Interval 10 0.75 (0.98) 0.18 (0.43) Interval 11 0.69 (0.98) 0.24 (0.65) Interval 12 0 .69 (0.97) 0.19 (0.47) Variable Name ENE - O Overall Touch df 11.00 11.00 F 1.39 6.32 p 0.17 0.00 2 0.01 0.04 Note. ENE - O = Overall Intensity of Expression of Negative Emotion (Average of ENE - I and ENE - P); Overall Touch = sum of Touch and Attempted Touch. 223 224 Additionally, descriptive analyses were conducted to describe direction of change in pred ictor variables including Maternal Regulatory Attempts, Appropriate, and Non - Attuned Mind - Maternal Regulatory Attempts and Mind - Mindedness Patterns across 12 Intervals . Descriptive analyses were conducted for Maternal Regulatory Attempts, Appropriate Mind Related Comments, and Non - Attuned Mind Related Comments to identify and describe patterns of change across all 12 measurement occasions. Additionally, one - way analysis of variance was conducted for observed Maternal Regulatory Attempts and Mind - Mindedness as a function of elapsed intervals (i.e., Elapsed Time). Overall, Appropriate and Non - Attuned Mind Related Comments and all Maternal Regulatory Attempts increased over time except for Positive Commands, Negative Commands, Physical Restraint, and Positive Emotional Expressions, Positive Verbal Control and Negative Control which decreased over the 12 intervals. For parsimony, only significant main effects of Elapsed Time are reported (Table D 2). Specifically, means for maternal Physical Comfort, Maternal Initiated Distraction , Negative Commands, and Physical Restraint differed as a function of Elapsed Time. 225 Table D2 Analysis of Variance (ANOVA) Between Maternal Regulatory Strategies and Elapsed Time Across 12 Intervals Variable Name Phy - Comf M (SD) M - INI - Dis M (SD) Neg - Comd M (SD) Phy - RST M (SD) Interval 1 0.04 (0.21) 0.11 (0.32) 1.07 (1.53) 0.83 (0.80) Interval 2 0.13 (0.40) 0.21 (0.43) 0.98 (1.66) 0.84 (0.89) Interval 3 0.17 (0 .45) 0.25 (0.45) 0.66 (1.24) 0.69 (0.85) Interval 4 0.18 (0.50) 0.30 (0.51) 0.57 (1.33) 0.61 (0.82) Interval 5 0.25 (0.57) 0.33 (0.50) 0.66 (1.48) 0.63 (0.85) Interval 6 0.22 (0.53) 0.36 (0.59) 0.49 (1.14) 0.60 (0.81) Interval 7 0.17 (0.43) 0.38 (0.50) 0.50 (1.04) 0.67 (0.85) Interval 8 0.22 (0.53) 0.32 (0.53) 0.59 (1.29) 0.61 (0.84) Interval 9 0.16 (0.44) 0.31 (0.69) 0.32 (0.70) 0.49 (0.79) Interval 10 0.22 (0.52) 0.40 (0.59) 0.54 (1.34) 0.61 (0.86) Interval 11 0.21 (0.53) 0.34 (0.65) 0.35 (0.93) 0 .61 (0.85) Interval 12 0.21 (0.48) 0.29 (0.51) 0.31 (0.81) 0.52 (0.82) Variable Name Phy - Comf M - INI - Dis Neg - Comd Phy - RST df 11 11 11 11 F 1.8 3 4.79 2.03 p 0.05 0.000 0.000 0.02 2 0.01 0.02 0.03 0.01 Note. Phy - Comf = Physical Comfort, M - INI - Dis = Maternal Initiated Distraction, Neg - Comd = Negative Commands, Phy - RST = Physical Restraint, Scales: M - Dis = Maternal Distraction (sum of M - INI - Dis and M - JNT - Dis), Neg - Ctrl = Negative Control (sum of Neg - Comd, Pun, and Min). 226 Toddler Regulatory Strategies Patterns across 12 Intervals. Descriptive analyses to identify and describe patterns of change across all 12 measurement occasions. Additionally, one - way analysis of variance was conducted for observed Initiated Verbal and Initiated Non - Verbal Distraction which increased over time. of Dependent Regulatory Strategies increased over the 12 intervals, except for Comfort Seeking which decreased over time. For parsimony, only significant main effects of Elapsed Time are reported (Table D3). Amongst Independent Regulatory St rategies, means for Toddler Initiated Non - Verbal Distraction and Toddler Initiated Verbal Keys differed as a function of Elapsed Time . Amongst Dependent Regulatory Strategies, means for Toddler Verbal Distraction Bids to Mom and Toddler Joined Non - Verbal Distraction differed as a function of Elapsed Time. 227 Table D3 Analysis of Variance (ANOVA) Between Toddler Regulatory Strategies and Elapsed Time Across 12 Intervals Independent Regulatory Strategies (IRS) Dependent Regulatory Strategies (DRS) Varia ble Name T - INI - nonVDis T - INI - Vkeys T - VDis - 2M T - JNT - nonVDis M (SD) M (SD) M (SD) M (SD) Interval 1 0.26 (0.47) 0.11 (0.45) 0.01 (0.09) 0.04 (0.19) Interval 2 0.40 (0.55) 0.10 (0.50) 0.01 (0.09) 0.06 (0.27) Interval 3 0.41 (0.58) 0.04 (0.24) 0.01 (0.12) 0.12 (0.35) Interval 4 0.46 (0.58) 0.00 (0.00) 0.03 (0.17) 0.13 (0.33) Interval 5 0.54 (0.70) 0.03 (0.21) 0.02 (0.15) 0.19 (0.48) Interval 6 0.45 (0.61) 0.03 (0.24) 0.05 (0.22) 0.23 (0.56) Interval 7 0.48 (0.66) 0.02 (0.15) 0.06 (0.27) 0.16 (0.39) In terval 8 0.45 (0.54) 0.02 (0.12) 0.10 (0.46) 0.14 (0.37) Interval 9 0.48 (0.58) 0.02 (0.15) 0.06 (0.27) 0.19 (0.60) Interval 10 0.50 (0.59) 0.02 (0.15) 0.10 (0.43) 0.13 (0.42) Interval 11 0.53 (0.63) 0.02 (0.12) 0.14 (0.50) 0.12 (0.45) Interval 12 0.45 (0.56) 0.01 (0.09) 0.11 (0.41) 0.13 (0.34) Variable Name T - INI - nonVDis T - INI - Vkeys T - VDis - 2M T - JNT - nonVDis df 11 11 11 11 F 1.98 2.55 3.00 2.26 p 0.03 0.00 0.00 0.01 2 0.01 0.02 0.02 0.02 Note . T - INI - nonVDis = Toddler Initiated Non - Verbal Distraction, T - INI - Vkeys = Toddler Initiated Verbal Keys, T - VDis - 2M = Toddler Verbal Distraction - Bids to Mom, T - JNT - nonVDis = Toddler Joined Non - Verbal Distraction 228 APPENDIX E Descriptive Statistics across Task (120s) 229 Descriptive statistics were conducted for all study variables in (1) overall scores across the entire 2 - minute delay of gratification task (all 120 seconds) for each parent - child dyad (sums and averages for all 1 2 intervals) (Tables E1 - E3). Differences in means of observed toddler independent regulatory strategies and dependent regulatory strategies were analyzed across task (Table E4) and per interval (Table E5) using paired sample t - tests. Toddler Expression of Negative Emotion and Delay of Gratification across Task (120 seconds). Descriptive statistics for Overall Expression of Negative Emotion (ENE - O), and Overall Touch are reported (Table E1). On average, on average, toddlers expressed a mild level of inten sity (ENE - O: M = 0.75, SD = 0.78) on a scale of 0 - 3 (0 indicating no cues for negative emotion, 1 = mild, 2 = moderate, 3 = severe level of intensity of negative emotion). Overall Touch scores suggest that on average toddlers either touched or attempted to touch the keys about 3 times during the entire 120 seconds ( M = 3.18, SD = 3.66). Table E1 Descriptive Statistics for Toddler Expression of Negative Emotion and Overall Touch Across Task (All 120 Seconds) Outcome Variables Min Max Mean SD Skewness (SE) ENE - O 0.00 2.88 0.75 0.78 0.94 (.21) Overall Touch 0.00 21.00 3.18 3.66 1.77 (.21) Note . ENE - O = Overall Expression of Negative Emotion (average of ENE - I and ENE - P); Overall Touch = sum of Touch and Attempted Touch Maternal Regulatory Attempts an d Mind - Mindedness across Task (120 seconds). In general, Maternal Regulatory Attempts, Appropriate and Non - Attuned Mind - Related Comments were used infrequently and on average occurred less than 7 times across the 2 - mintute task. Maximum variation in freq uency of strategies used was observed for Negative 230 Maternal Joined Distraction, Minimizing, Appropriate and Non - Attuned Mind - Related Comments during the entire 120 seconds task. The most frequently used regulatory attempts attempts w more physical comfort than verbal comfort. Mothers initiated distractions more often than joined minimizing statements but less than any amount of physical restraint. Maternal display of negative emotional reactions was used less often th an positive emotional reactions. Overall means for maternal mind related comments for all 12 intervals suggest that on average mothers made more Appropriate Mind - Related Comments compared to Non - Attuned Mind - Related Comments (Table E2). Table E2 Descript ive Statistics of for Maternal Regulatory Attempts, Appropriate and Non - Attuned Mind - Related Comments across Task (All 120 Seconds) Variable Name Min Max Mean SD Skewness (SE) Verbal Comfort 0 12 1.69 2.63 2.20 (.21) Physical Comfort 0 24 2.12 3.43 3 .06 (.21) Initiated Distraction 0 22 3.51 4.11 1.47 (.21) Joined Distraction 0 7 0.78 1.44 2.21 (.21) Verbal Orientation to Delay 0 38 4.24 7.02 2.45 (.21) Positive Commands 0 22 4.93 5.29 1.15 (.21) Negative Commands 0 65 6.93 8.02 3.35 (.21) Pu nitive Reactions 0 14 0.57 1.61 5.20 (.21) Minimizing 0 6 0.34 0.99 3.74 (.21) Physical Restraint 0 2 0.63 0.60 0.74 (.21) Positive Emotional Reactions 0 1 0.39 0.30 0.41 (.21) Negative Emotional Reactions 0 0.67 0.02 0.08 5.71 (.21) Appropriate Mind Related Comments 0 10 0.89 1.56 2.84 (.21) Non - Attuned Mind Related Comments 0 11 0.77 1.59 3.23 (.21) 231 Toddler Regulatory Strategies across Task (120 seconds). Overall means for the frequency of toddler regulatory strategies are reported (Table E3). I ndependent Regulatory Strategies across Task (120 seconds). In general, Independent Regulatory Strategies were used infrequently and on average occurred less than 6 times during the entire 2 - minute task. Maximum variation in frequency of strategies used w as observed for toddler initiated Non - Verbal Distraction (e.g., looks away for more that 3s, walks out of the room) and physically self - comforting behaviors (e.g., thumb sucking, rocking back and forth) during the entire two - minute task. Minimum variation in frequency of strategies used was displayed independent strategy by toddlers was Self - Comfort (e.g., thumb sucking, rocking back and forth), while the leas t frequently displayed independent strategy was Initiated Verbal Non - Verbal Distraction (e.g. looks away for more than 3 secon ds) during the entire 2 - minute task. Toddlers vocalized more words or phrases describing the keys or reframing the task (e.g., alizations in the form of self - direction or in reference to the rules of below). Dependent Regulatory Strategies across Task (120 seconds). In general, Dependent Regu latory Strategies were used infrequency and on average occurred less than 3 times during the entire 2 - mintue task. Maximum variation in frequency of strategies used was observed for toddler Joint Non - Verbal Distraction (e.g., looks to where mother points, 232 initiated distracting activity). Minimum variation in frequency of strategies used was observed for toddler verbalizations in the form of self - direction or in reference to the rules of the task (e.g. other during the entire 2 - minute task. The most frequently displayed strategy directed toward or in response to mother was Physical Comfort Seeking (e.g., strategy directed toward or in response to mother were verbalizations in the form of self - - minute task. On average, toddlers displayed more Joined Non - Verbal Distracti on than Joined Verbal Distraction. On average, toddlers vocalized more words or phrases describing the keys or erbalizations in the form of self - direction or in Table E3 Descriptive Statistics for Toddler Regulatory Strategies Across Task (All 120 Seconds) Variable Name Min Max Mean SD Skewness (SE ) Independent Regulatory Strategies (IRS) Initiated Verbal Distraction 0.00 6.00 0.12 0.60 7.74 (.21) Initiated Non - Verbal Distraction 0.00 20.00 5.28 4.51 0.59 (.21) Initiated Verbal Keys 0.00 9.00 0.42 1.25 4.32 (.21) Initiated Verbal Self - Control 0.00 0.00 0.00 0.00 . Initiated Verbal Desire 0.00 8.00 0.18 0.91 6.67 (.21) Self - Comfort 0.00 22.00 5.89 4.94 0.79 (.21) Dependent Regulatory Strategies (DRS) Verbal Distraction - Bids to Mom 0.00 10.00 0.69 1.59 3.06 (.21) Joined Verbal Distraction 0 .00 11.00 0.68 1.90 3.51 (.21) Joined Non - Verbal Distraction 0.00 18.00 1.59 2.98 2.72 (.21) Verbal Keys - Bids to Mom 0.00 16.00 0.99 2.36 3.92 (.21) Verbal Self - Control - Bids to Mom 0.00 7.00 0.11 0.72 7.98 (.21) Verbal Desire - Bids to Mom 0.00 8.00 0. 37 1.03 4.30 (.21) Physical Comfort Seeking 0.00 14.00 2.80 3.47 1.35 (.21) 233 For exploratory purposes, differences in means of observed toddler independent regulatory strategies and dependent regulatory strategies were analyzed across task (Table E4) a nd per interval (Table E5) using paired sample t - tests. Table E4 Dependent Regulatory Strategies Across Task (All 120 Seconds) Independent Regulatory Strategies De pendent Regulatory Strategies Toddler Initiated Verbal Distraction < *** Toddler Verbal Distraction - Bids to Mom Toddler Initiated Verbal Distraction < *** Toddler Joined Verbal Distraction Toddler Initiated Non - Verbal Distraction > *** Toddler Joined Non - Verbal Distraction Toddler Initiated Verbal Keys < ** Toddler Verbal Keys Bids to Mom Toddler Initiated Verbal Self - Control < Toddler Verbal Self - Control - Bids to Mom Toddler Initiated Verbal Desire < Toddler Verbal Desire - Bids to Mom To ddler Self - Comfort > *** Toddler Physical Comfort Seeking Total Toddler Independent Regulatory Strategies > *** Total Toddler Dependent Regulatory Strategies Note. Paired Samples T - Tests * p < .05, two - tailed. ** p < .01, two - tailed. *** p <.001, two - tailed . p < .10, two - tailed. Table E5 Comparison between means for frequency of toddlers Independent Regulatory Strategies and Dependent Regulatory Strategies per 10 - Second Interval Independent Regulatory Strategies Dependent Regulatory Strategies Toddler Initiated Verbal Distraction < *** Toddler Verbal Distraction - Bids to Mom Toddler Initiated Verbal Distraction < *** Toddler Joined Verbal Distraction Toddler Initiated Non - Verbal Distraction > *** Toddler Joined Non - Verbal Distraction Toddler Initi ated Verbal Keys < *** Toddler Verbal Keys Bids to Mom Toddler Initiated Verbal Self - Control < ** Toddler Verbal Self - Control - Bids to Mom Toddler Initiated Verbal Desire < ** Toddler Verbal Desire - Bids to Mom Toddler Self - Comfort > *** Toddler P hysical Comfort Seeking Total Toddler Independent Regulatory Strategies > *** Total Toddler Dependent Regulatory Strategies Note. Paired Samples T - Tests * p < .05, two - tailed. ** p < .01, two - tailed. *** p <.001, two - tailed. p < .10, two - tailed. 234 APPENDIX F Correlations of Outcomes with Current - Interval Predictors 235 Current - Interval Maternal Variables and Overall Expression of Negative Emotion First, correlations for current interval Maternal Regulatory Attempts, current - interval Appropriate and curren t - interval Non - Attuned Mind - Related Comments with ENE - O are presented (Table F1). All correlations for current - interval Maternal Regulatory Attempts with ENE - O were small but significant (r = - .35 to .06), except for non - significant correlations for curren t - interval Positive Commands. Unexpectedly, as current - interval Verbal Comfort and current - interval Physical Comfort increased, ENE - O increased. All other correlations were in expected directions. Strongest negative associations for current interval predi ctors were between maternal positive associations were between maternal Physical Restraint followed by Negative Emotional ive Emotion. Toddlers expressed lower intensity negative emotions as mothers expressed or shared positive affect with their toddlers in current behavior increased a mocking way or when they became angry themselves, toddlers tended to express higher intensity of negative emotions. Appropriate Mind - Related Comments were not associated with Exp ression of Negative Emotion while Non - Attuned Mind - Related Comments were positively related to Expression of Negative Emotion as expected; suggesting that higher use of lagged Non - Attuned Mind - Related Comments was associated with higher intensity of Expres sion of Negative Emotion as expected. 236 Table F1 Correlations for Current Interval Maternal Regulatory Attempts, Appropriate and Non - Attuned Mind - Related Comments with Expression of Negative Emotion (n=1571) Variable Name ENE - O Verbal Comfort .12 ** Physical Comfort .09 ** Initiated Distraction - .17 ** Joined Distraction - .06 * Verbal Orientation to Delay - .11 ** Positive Commands .03 Negative Commands .15 ** Punitive Reactions .15 ** Minimizing .22 ** Physical Restraint .30 ** Positive Emotion al Reactions - .35 ** Negative Emotional Reactions .19 ** Appropriate Mind Related Comments - .02 Non - Attuned Mind Related Comments .10 ** Note. ENE - O = Overall Expression of Negative Emotion (average of ENE - I and ENE - P). * p < .05, one - tailed. ** p < .01, on e - tailed. p < .10, one - tailed. Current - Interval Toddlers Regulatory Strategies and Expression of Negative Emotion. Correlations for lagged Independent Regulatory Strategies and lagged Dependent Regulatory Strategies with ENE - O are presented (Table F2). Correlations for current - interval Independent Regulatory Strategies with ENE - O were small but significant (r = - .33 to .06). Unexpectedly, current - positively associated with ENE - O suggesting that as toddlers vocalized desire states, they tended to express more intense levels of negative emotion. These associations suggest that verbalization of desire words may facilitate expression of negative emotion. All other correlations were in expected negative directions. Strongest associations were observed for current - interval Toddler Initiated Non - Verba l Distraction and ENE - O. As toddlers initiated engagement in non - verbal distracting behaviors such as play or looking away from the keys, they tended to express milder levels of negative emotions. 237 Correlations for current - interval Dependent Regulatory Str ategies with ENE - O were small but significant (r= - .05 to - .18). Current - interval Toddler Verbal Distraction - Bids to Mom - interval Toddler Verbal Desire - Bids to Mom (e.g. in response to mom), current - interval Toddler Physical Comfort Seeking (e.g., reaching arms up - O. All signi ficant associations were in expected negative direction. Strongest associations were observed for current - interval Toddler Joined Non - Verbal Distraction and ENE - O. As toddlers - verbal distractions (e.g., Mother points to a direction away from keys and child looks where she is pointing), they expressed lower levels of ENE - O. Table F2 Correlations for Current Interval Toddler Regulatory Strategies with Expression of Negative Emotion (n = 1571) Variable Name ENE - O Inde pendent Regulatory Strategies (IRS) Initiated Verbal Distraction - .05 * Initiated Non - Verbal Distraction - .33 ** Initiated Verbal Keys - .06 * Initiated Verbal Self - Control - Initiated Verbal Desire .06 ** Self - Comfort - .09 ** Dependent Regulatory Strateg ies (DRS) Verbal Distraction - Bids to Mom - .03 Joined Verbal Distraction - .13 ** Joined Non - Verbal Distraction - .18 ** Verbal Keys - Bids to Mom - .09 ** Verbal Self - Control - Bids to Mom - .05 * Verbal Desire - Bids to Mom .02 Physical Comfort Seeking .00 No te. ENE - O = Overall Expression of Negative Emotion (average of ENE - I and ENE - P); * p < .05, one - tailed. ** p < .01, one - tailed. p < .10, one - tailed. 238 Current - Interval Maternal Variables and Delay of Gratification Next, correlations for current - interval Maternal Regulatory Attempts, current - interval Appropriate, and current - interval Non - Attuned Mind - Related Comments with Delay of Gratification operationalized as amount of Overall Touch are presented (Table F3). All correlations for current interval Matern al Regulatory Attempts with Overall Touch, were small but significant (r = - .06 to .32). Current - interval Verbal Comfort, Verbal Orientation to Delay, Punitive Reactions, and Minimizing statements were not significantly related to Overall Touch. Unexpected ly, current - interval Positive Commands and Positive Emotional Reactions were r attempt to touch the keys increased. All other correlations were in the expected direction. Strongest negative associations for current - interval predictors were between maternal Initiated Distraction and Overall Touch, while strongest positive associati ons were between Negative mpt to touch the keys increased as expected. Current - interval Appropriate and Non - Attuned Mind - Related Comments were not correlated with Overall Touch . 38 38 Appropriate mind - related comments and Touch were negatively associated ( r = - .05, p < .10), suggesting that as mothers displayed more appropriate mind - expected. Unexpect edly, higher frequency of Non - Attuned Mind - Related Comments was related to lower frequency of Touch ( r = - .05, p < .05), suggesting that as mothers displayed more Non - Attuned Mind - Related Comments, his association may be explained by a large amount of overlap between Non - Attuned Mind - Related Comments and Maternal Initiated Distraction (76.8 %). 239 Table F3 Correlations for Current Interval Maternal Regulatory Attempts, Appropriate and Non - Attuned Mind - Related Comments with Touch (n=1571) Variable Name Overall Touch Verbal Comfort - .03 Physical Comfort - .12 ** Initiated Distraction - .12 ** Joined Distraction - .06 ** Verbal Orientation to Delay .04 Positive Commands .16 ** Negative Command s .32 ** Punitive Reactions .02 Minimizing .01 Physical Restraint .25 ** Positive Emotional Reactions .06 ** Negative Emotional Reactions .12 ** Appropriate Mind Related Comments - .04 Non - Attuned Mind Related Comments - .03 Note. Overall Touch = sum of Touch and Attempted Touch. * p < .05, one - tailed. ** p < .01, one - tailed. p < .10, one - tailed. Current Interval Toddler Regulatory Strategies and Delay of Gratification. Correlations for lagged Independent Regulatory Strategies and lagged Dependent Regul atory Strategies with Delay of Gratification operationalized as amount of Overall Touch are presented (Table F4) . Correlations for current - interval Independent Regulatory Strategies with Overall Touch were small but significant (r = - .21 to .06). Current - i nterval Toddler Initiated Verbal - interval Overall Touch suggesting that as toddlers vocalized desire states, they tended to touch or attempted to touch the keys. These associations suggest that verbalization of desire words that are not directed toward mother may impede ability to wait. All other correlations were in expected negative directions. Strongest associations were observed for current - interval Toddler Initiated Non - Verbal Distraction (e.g., looking away for more than 3s). As expected, as toddlers initiated more non - verbal distractions, they tended to display less Ove rall Touch. 240 Correlations for current - interval Dependent Regulatory Strategies with Overall Touch were also small but significant (r= - .11 to .07). Current - interval Toddler Initiated Verbal Keys significantly related to Overall Touch. Unexpectedly, current - toward mother or in response to mother, was positively associated with Overall Touch. Also unexpectedly, current - inte rval Toddler Initiated Verbal Self - was positively but marginally associated with Overall Touch. These results suggest that as toddlers vocalized desire states or task - related rules in current interval, they tended to to uch or attempted to touch the keys. All other correlations were in expected negative directions. Strongest associations were observed for current - interval Toddler Physical Comfort Seeking. As expected, as toddlers displayed more Physical Comfort Seeking (e .g. leaning in to mom, reaching up for hug), they tended to display less Overall Touch. Table F4 Correlations for Current Interval Toddler Regulatory Strategies with Expression of Negative Emotion and Overall Touch (n = 1571) Variable Name Overall Touc h Independent Regulatory Strategies (IRS) Initiated Verbal Distraction - 0.02 Initiated Non - Verbal Distraction - .21 ** Initiated Verbal Keys .09 ** Initiated Verbal Self - Control - Initiated Verbal Desire .06 ** Self - Comfort - .09 ** Dependent Regulatory Strategies (DRS) Verbal Distraction - Bids to Mom - .07 ** Joined Verbal Distraction - .06 ** Joined Non - Verbal Distraction - .09 ** Verbal Keys - Bids to Mom .01 Verbal Self - Control - Bids to Mom .04 Verbal Desire - Bids to Mom .07 ** Physical Comfort Seeking - .11 ** Note. Overall Touch = sum of Touch and Attempted Touch. * p < .05, one - tailed. ** p < .01, one - tailed. p < .10, one - tailed. 241 APPENDIX G Independence of Error s and Intraclass Correlations 242 Independence of Errors. Intraclass correlations (ICCs) were calculated from the two - level base model (null or unconditional model) to determine amount of variability in Expression - O and Overall Touch differ across the tw o - minute task? Operationalized in multilevel modeling framework: Are there intraindividual differences in means of ENE - O within toddlers across 12 intervals? Level - 1 Equation: The Intercept - Only Model (Outcome 1) ENE - O ij = 0j +e ij where i indicates th e time interval within a dyad and j indicates a mother - child dyad, 0j indicates the intercept (mean) for ENE - O in dyad j (varies over j dyads), e ij indicates random errors of prediction for level - 1 equation (deviation for interval i from its dyad j ). Level - 2 Equation: The Intercept - Only Model (Outcome 1) 0j = 00 + u 0j where 0j, an intercept (mean) for group j is the sum of the overall intercept, the grand mean of the ENE - O scores across all dyads when all predictors = 0, 00 , and the random erro r associated with the group intercept, u 0j (deviation for group j from overall intercept) Operationalized in multilevel modeling framework: Are there intraindividual differences in means of Overall Touch within toddlers across 12 intervals? Level - 1 Equa tion: The Intercept - Only Model (Outcome 2) OverallTouch ij = 0j +e ij where i indicates the time interval within a dyad and j indicates a mother - child dyad, 0j indicates the intercept (mean) for Overall Touch in dyad j (varies over j dyads), e ij indicates random errors of prediction for level - 1 equation (devia tion for interval i from its dyad j ). Level - 2 Equation: The Intercept - Only Model (Outcome 2) 0j = 00 + u 0j where 0j, an intercept (mean) for group j is the sum of the overall intercept, the grand mean of the Overall Touch scores across all dyads w hen all predictors = 0, 00 , and the random error associated with the group intercept, u 0j (deviation for group j from overall intercept) 243 Intraclass Correlations (ICC). ICCs were calculated as the amount of variance among toddlers (between groups) divide d by the sum of variance among toddlers and within dyad (i.e., between variability in ENE - I, 65% of variability in ENE - P, 17 % of variability in Touch, 17 % of variability i n Attempted Touch. There was 64% of variability in ENE - I, 65% of variability in ENE - P, 17 % of variability in Touch, 17 % of variability in Attempted Touch. Only ENE - O and Overall Touch were specified as outcomes in random coefficient models. Table G1 Int raclass correlations for Intercept - Only Models for Expression of Negative Emotion and Delay of Gratification Outcome Variables Within Group Variance e ij Between Group Variance u oj ICC ENE - O 0.3464 0.67454 0.66 Overall Touch 0.30138 0.07512 0.20 Note. E NE - O = Overall Expression of Negative Emotion (average of ENE - I and ENE - P); Overall Touch = sum of Touch and Attempted Touch. ICC = Between Group Variance / Sum of Within and Between Group Variance. 244 APPENDIX H Model Specification of Covariates 245 Fo r models in Questions 1 (MRA, MM & ENE - O) and 3 (MRA, MM, &Overall Touch), covariates included lagged maternal verbosity (e.g., Total Comments), toddler gender, linear elapsed time, and quadratic elapsed time. Lagged maternal verbosity was included in mode ls specified with lagged Verbal Comfort, Verbal Orientation to Delay, Positive Commands, Negative Commands, Minimizing, as well as Appropriate Mind - Related Comments, and Non - Attuned Mind - related Comments. Child gender was specified as a level 2 predictor ( of overall intercept) in models predicted from level - 1 Maternal Regulatory Attempts, Appropriate and Non - Attuned Mind - related Comments for both outcomes. Covariates for questions 2 (TRS & ENE - O) and 4 (TRS & Overall Touch) were included based on correlat ional and main effect findings from the Intercept - Only models for both gender (higher for boys) with ENE - O but not with Overall Touch. Additionally, while gender wa s a significant predictor of differences in overall mean of ENE - O (higher for boys), it was not related to Overall Touch (Intercept - Only Model). their ENE - O, it was ne gatively associated with Overall Touch (correlational analyses). Additionally, while expressive vocabulary was not a significant predictor of ENE - O , it was marginally and negatively associated with Overall Touch ( = - 0.002, p = 0.08) suggesting that higher scores in expressive vocabulary (mother - report) were associated with lower Overall Touch, an association that was not found for ENE - O (intercept - only model). Therefore, gender was included in random coefficient models predicting ENE - O from lagged Toddler Regulatory Strategies (Question 2), and expressive language was included in random coefficient models predicting Overall Touch from lagged Toddler Regulatory Strategies (Question 4). Time varying 246 predictors, inc luded Elapsed Time - Linear (ET) and Elapsed Time 2 Quadratic (ET 2 ), and were used as level - 1 explanatory variables in all models to account for linear and quadratic effects of time on average ENE - O and Overall Touch and were specified with fixed slopes. Ta ble H1 Variable names and labels for study covariates Covariate Label Covariate Name Computed Included in Questions Variable Level Lagged Maternal Verbosity Verb Total Comments per 10s 1 & 3 Level 1 Linear Elapsed Time ET 10s 1 - 4 Level 1 Quadratic Elapsed Time ET 2 (10s) 2 1 - 4 Level 1 Toddler Gender Gender 0 = Girls, 1 = Boys 1 - 3 (Not 4) Level 2 Toddler Expressive Language Vocab Vocabulary Production Only 4 Level 2 Note . Level - 1 indicates that variable was observed in 10s intervals. Level - 2 indic ates variable was observed at dyad level. 247 APPENDIX I Models of Non - Significant Predictors 248 Question 1A. Model results for non - significant predictors in Question 1A (Phy - Comf, M - JNT - Dis, M - VO2D, and AMM) are presented in Tables I1 - I4. Lagged Materna l Physical Comfort . A two - level hierarchical model examined the - O) while controlling for effects of linear elapsed time, quadratic elapsed time, and toddler gender . The model was specified with the following equation: ENE - O ij 00 Gender (Gender j ) Phy - Comf (Phy - Comf i - 1j ) + ETL (ET ij ET 2 (ET 2 ij ) + u 0j + e ij Unexpectedly, lagged maternal Physical Comfort (e.g., giving hugs) was not associated with ENE - units higher for boys (scale of 0 - 3) after accounting for effects for all predictors. Additionally, the current model was compared against a model (Model 2) in which slopes be tween lagged maternal Physical Comfort and ENE - O were allowed to vary between dyads. However, results of deviance tests ( = chi squared difference) indicated that allowing the relationship between lagged maternal Physical Comfort and ENE - O to vary among dyads would not significantly help to explain more variance in ENE - O between toddlers ( = 4.20, df = 2, p > 0.05). Additionally, the p - value associated with random slope was not significant suggesting that the association between lagged maternal Physical Comfort and ENE - O does not vary between dyads. Effects of lagged maternal Physical Comfort on ENE - O remain non - significant in Model 2. 249 Table I1 Effects of Lagged Maternal Physical Comfort on Overall Expression of Negative Emotion Fixed Effects Coefficient S.E Intercept 0.63 *** 0.15 Gender 0.31 ** 0.15 Physical Comfort 0.01 0.05 Linear ET 0.00 0.00 Quadr atic ET - 0.00 0.00 Random Effects Variance Components SD u 0j 0.72 *** 0.85 e ij 0.32 0.57 Note. Elapsed Time (ET) is in 10s intervals. * p < .05, ** p < .01, *** p < .001, p < .10 Lagged Maternal Joined Distraction . A two - level hierarchical model examine d the effects of lagged maternal Joined (ENE - O) while controlling for effects of linear elapsed time, quadratic elapsed time, and toddler gender. The model was specified with the following equation: ENE - O ij 00 Gender (Gender j ) M - JNT - Dis ( M - JNT - Dis - 1j ) + ETL (ET ij ET 2 (ET 2 ij ) + u 0j + e ij Unexpectedly, lagged maternal Joined Distraction (e.g. child plays with toes and mother - O ( overall expression of negative emotion was 0.31 units higher for boys (scale of 0 - 3) after accounting for effects for all predictors. Additionally, the current model was compared against a model (Model 2) in which slopes betwe en lagged maternal Joined Distraction and ENE - O were allowed to vary between dyads. However, results of deviance tests ( = chi squared difference) indicated that allowing the relationship between lagged maternal Joined Distraction and ENE - O to vary among dyads would not significantly help to explain more variance in ENE - O between toddlers ( = .59, df = 2, p > 0.05). Additionally, the p - value associated with random slope was not significant suggesting that the association between lagged maternal Joined Distraction and ENE - O does not vary 250 between dyads. Effects of lagged maternal Joined Distraction on ENE - O remain non - significant in Model 2. Table I2 Effects of Lagged Maternal Joined Distraction Comfort on Overall Expression of Negative Emotion Fixed Effects Coefficient S.E Intercept 0.63 *** 0.15 Gender 0.31 ** 0.15 Joined Distraction - 0.08 0.06 Linear E T 0.00 0.00 Quadratic ET - 0.00 0.00 Random Effects Variance Components SD u 0j 0.72 *** 0.85 e ij 0.32 0.57 Note. Elapsed Time (ET) is in 10s intervals. * p < .05, ** p < .01, *** p < .001, p < .10 Lagged Maternal Verbal Orientation to Delay. A two - level hierarchical model examined Emotion (ENE - O) while controlling for effects of linear elapsed time, quadratic elapsed time, lagged maternal verbosity, and toddler gender. The model was specified with the following equation: ENE - O ij 00 Gender (Gender j ) M - VO2D ( M - VO2D i - 1j ) + ETL (ET ij ET 2 (ET 2 ij Verb (Verb i - 1j ) + u 0j + e ij - O but was in the units higher for boys (scale of 0 - 3) after accounting for effe cts for all predictors. Additionally, the current model was compared against a model (Model 2) in which slopes between lagged maternal Verbal Orientation to Delay and ENE - O were allowed to vary between dyads. Results of deviance tests ( = chi squared difference) indicated that allowing the relationship between lagged maternal Verbal Orientation to Delay and ENE - O to vary among dyads would 251 significantly improve model fit = 9.20, df = 2, p < .05. Indeed, when the associ ation between lagged maternal Verbal Orientation to Delay and ENE - O was allowed to vary between dyads, the p - value associated with the random slope coefficient was significant ( p < .01), suggesting that dyads varied in the association between lagged matern al Verbal Orientation to Delay and ENE - O. Effects of lagged maternal Verbal Orientation to Delay on ENE - O remain non - significant in Model 2. Table I3 Effects of Lagged Maternal Verbal Orientation to Delay on Overall Expression of Negative Emotion Model 1 Model 2 Fixed Effects Coefficient S.E Coefficient S.E Intercept 0.51 *** 0.15 0.51 *** 0.15 Gender 0.31 * 0.15 0.31 * 0.15 Verbal Orientation to Delay - 0.02 0.03 - 0.01 0.03 Linear ET 0.00 0.00 0.00 0.00 Quadratic ET - 0.00 0.00 - 0.00 0.00 Verbosity 0 .04 *** 0.01 0.04 *** 0.01 Random Effects Variance Components SD Variance Components SD u 0j 0.73 *** 0.85 0.74 *** 0.86 u M - VO2D , j - - 0.01 ** 0.12 e ij 0.32 0.56 0.31 0.55 Deviance 2902.29 2893.08 Note. Elapsed Time (ET) is in 10s intervals. *p < .05, * Lagged Maternal Appropriate Mind - Related Comments. A two - level hierarchical model examined the effects of lagged maternal Appropriate Mind - Expression of Negative Emotion (ENE - O) while controll ing for effects of linear elapsed time, quadratic elapsed time, lagged maternal verbosity, and toddler gender. The model was specified with the following equation: ENE - O ij 00 Gender (Gender j ) AMM (AMM i - 1j ) + ETL (ET ij ET 2 (ET 2 ij Verb (Verb i - 1j ) + u 0j + e ij 252 Unexpectedly, lagged maternal Appropriate Mind - Related Comments id while toddler is pointing at keys) was not significantly associated with ENE - O, but associations were in the expected higher for boys (scale of 0 - 3) after acco unting for effects for all predictors. Additionally, the current model was compared against a model (Model 2) in which slopes between lagged maternal Appropriate Mind - Related Comments and ENE - O were allowed to vary between dyads. However, results of devian ce tests ( = chi squared difference) indicated that allowing the relationship between lagged maternal Appropriate Mind - Related Comments and ENE - O to vary among dyads would not significantly help to explain more variance in ENE - O between toddlers ( = 0.05, df = 2, p > 0.05). Additionally, the p - value associated with random slope was not significant suggesting that the association between lagged maternal Appropriate Mind - Related Comments and ENE - O does not vary between dyads. Effects of la gged maternal Appropriate Mind - Related Comments on ENE - O remain non - significant in model I4. Table I4 Effects of Lagged Appropriate Mind - Related Comments on Overall Expression of Negative Emotion Fixed Effects Coefficient S.E Intercept 0.51 *** 0.15 G ender 0.31 * 0.15 Appropriate Mind - Related Comments - 0.03 0.05 Linear ET 0.00 0.00 Quadratic ET - 0.00 0.00 Verbosity 0.04 *** 0.01 Random Effects Variance Components SD u 0j 0.73 *** 0.85 e ij 0.32 0.56 Note. Elapsed Time (ET) is in 10s intervals. * p < .05, ** p < .01, *** p < .001, p < .10 Question 1B. Model results for non - significant predictors in Question 1B (Neg - Comd and NMM) are presented in Tables I5 - I6. 253 Lagged Maternal Negative Commands . A two - level hierarchical model examined the effects of la (ENE - O) while controlling for effects of linear elapsed time, quadratic elapsed time, lagged maternal verbosity, and toddler gender. The model was specified with the following equa tion: ENE - O ij 00 Gender (Gender j ) Neg - Comd ( Neg - Comd i - 1j ) ETL (ET ij ET 2 (ET 2 ij Verb (Verb i - 1j ) + u 0j + e ij was not associated with ENE - O but in the exp overall expression of negative emotion was 0.31 units higher for boys (scale of 0 - 3) after accounting for effects for all predictors. Additionally, the current model was compared against a model (Model 2) in w hich slopes between lagged maternal Negative Commands and ENE - O were allowed to vary between dyads. However, results of deviance tests ( = chi squared difference) indicated that allowing the relationship between lagged maternal Negative Commands and ENE - O to vary among dyads would not significantly help to explain more variance in ENE - O between toddlers ( = 3.73, df = 2, p > 0.05). However, when the association between lagged maternal Negative Commands and ENE - O was allowed to vary between dyads, the p - value associated with the random slope coefficient was significant (p <.05), suggesting that dyads varied in the associat ion between lagged maternal Negative Commands and ENE - O. Effects of lagged Negative Commands on ENE - O remain non - significant in Model 2. Results of more parsimonious model are presented in Table I5. 254 Table I5 Effects of Lagged Negative Commands on Over all Expression of Negative Emotion Fixed Effects Coefficient S.E Intercept 0.51 *** 0.15 Gender 0.31 * 0.15 Negative Commands 0 0.02 Linear ET 0.00 0.00 Quadratic ET - 0.00 0.00 Verbosity 0.04 ** 0.01 Random Effects Variance Components SD u 0j 0.72 *** 0.85 e ij 0.32 0.56 Note. Elapsed Time (ET) is in 10s intervals. * p < .05, ** p < .01, *** p < .001, p < .10 Maternal Non - Attuned Mind - Related Comments. A two - level hierarchical model examined the effects of lagged maternal Non - Attuned Mind - Related Comm Expression of Negative Emotion (ENE - O) while controlling for effects of linear elapsed time, quadratic elapsed time, lagged maternal verbosity, and toddler gender. The model was specified with the following equation: ENE - O ij 00 Gend er (Gender j ) NMM (NMM i - 1j ETL (ET ij ) + ET 2 (ET 2 ij Verb (Verb i - 1j ) + u 0j + e ij Unexpectedly, lagged maternal Non - Attuned Mind - said when toddler is reaching for keys, said while toddler shows no overt signs of fatigue ) was not associated with ENE - O, but was in the expected higher for boys (scale of 0 - 3) after accounting for effects for all predictors. Additionally, the current model was compared against a model (Model 2) in which slopes between lagged maternal Non - Attuned Mind - Related Comments and ENE - O were allowed to vary between dyads. However, results of deviance tests ( = chi squared difference) indicated that allowing the relationship between lagged maternal Non - Attuned Mind - Related Comments and ENE - O to 255 vary among dyads would not significantly help to explain more variance in ENE - O between toddlers ( = 3.47, df = 2, p > 0.05). However, when the association between lagged maternal Non - Attuned Mind - Related Comments and ENE - O was allowed to vary between dyads, the p - value associated with the random slope coefficient was significant (p <.05), suggesting t hat dyads varied in the association between lagged maternal Non - Attuned Mind - Related Comments and ENE - O. Effects of lagged maternal Non - Attuned Mind - Related Comments on ENE - O remain non - significant in Model 2. Results of more parsimonious model are present ed in Table I6. Table I6 Effects of Lagged Non - Attuned Mind - Related Comments on Overall Expression of Negative Emotion Fixed Effects Coefficient S.E Intercept 0.51 *** 0.15 Gender 0.31 * 0.15 Non - Attuned Mind - Related Comments 0.05 0.06 Linear ET 0.00 0.00 Quadratic ET - 0.00 0.00 Verbosity 0.04 *** 0.001 Random Effects Variance Components SD u 0j 0.73 *** 0.85 e ij 0.32 0.56 Note. Elapsed Time (ET) is in 10s intervals. * p < .05, ** p < .01, *** p < .001, p < .10 Question 2A. Model results for non - s ignificant predictor in Question 2A (T - INI - VDis) is presented in Table I7. Lagged Toddler Initiated Verbal Distraction. A two - level hierarchical model examined the effects of lagged Toddler Initiated Verbal Distraction Emotion (ENE - O) while controlling for effects of linear elapsed time, quadratic elapsed time, and toddler gender. The model was specified with the following equation: ENE - - INI - Vdis (T - INI - VDisi - 1j) + 256 sings) was not associated with ENE - O but was in the negat ive direction as expected (Table I7). of 0 - 3) after accounting for effects for all predictors. Additionally, the current model was compared against a model (Mode l 2) in which slopes between lagged Toddler Initiated Verbal Distraction and ENE - O were allowed to vary between dyads. However, results of deviance tests ( = chi squared difference) indicated that allowing the relationship between lagged Toddler Initiated Verbal Distraction and ENE - O to vary among dyads would not significantly help to explain more variance in ENE - O between toddlers ( = 1. 36, df = 2, p > 0.05). However, when the association between lagged Toddler Initiated Verbal Distraction and ENE - O was allowed to vary between dyads, the p - value associated with the random slope coefficient was trending toward significance ( p <.01), sugges ting that dyads tended to vary in the association between lagged Toddler Initiated Verbal Distraction and ENE - O. Effects of lagged Toddler Initiated Verbal Distraction on ENE - O remain non - significant in Model 2. Results of more parsimonious model are prese nted in Table I7. Table I7 Effects of Lagged Initiated Verbal Distraction on Overall Expression of Negative Emotion Fixed Effects Coefficient S.E Intercept 0.63 *** 0.15 Gender 0.31 * 15 Initiated Verbal Distraction - 0.08 0.18 Linear ET 0.00 0.00 Qu adratic ET - 0.00 0.00 Random Effects Variance Components SD u 0j 0.72 *** 0.85 eij 0.32 0.57 Question 2B. Model results for non - significant predictors in Question 2 B (T - Vkeys - 2M, T - Vdes - 2M, T - ComfSk) are presented in Appendix I (Table I8 - I10). 257 Lagged Toddler Verbal Keys Bids to Mom . A two - level hierarchical model examined the effects of lagged Toddler Verbal Keys Bids to Mom Em otion (ENE - O) while controlling for effects of linear elapsed time, quadratic elapsed time, and toddler gender. The model was specified with the following equation: ENE - - Vkeys - 2M (T - Vkeys - 2M i - 1j ) + ETL (ET ij ET 2 (ET 2 ij ) + u 0j + e ij Unexpectedly, lagged Toddler Verbal Keys Bids to Mom (e.g., r in response to mom ) was not associated with ENE - overall expression of negative emotion was 0.31 units higher for boys (scale of 0 - 3) after accounting for effects for all predictors. Additionally, the current model was compared against a model (Model 2) in which slopes between lagged Toddler Verbal Keys Bids to Mom and ENE - O were allowed to vary between dyads. Results of deviance tests ( = chi squared difference) indicated that allowing the relationship between lagged Toddler Verbal Keys Bids to Mom and ENE - O to vary among dyads would significantly improve model fit = 12.56, df = 2, p < 0.01). Indeed, when the assoc iation between lagged Toddler Verbal Keys Bids to Mom and ENE - O was allowed to vary between dyads, the p - value associated with the random slope coefficient was significant ( p <.001), suggesting that dyads varied in the association between lagged Toddler Verbal Keys Bids to Mom and ENE - O. Effects of lagged Toddler Verbal Keys Bids to Mom on ENE - O remain non - significant in Model 2 but in the same negative direction as expected. 258 Table I8 Effects of Lagged Toddler Verbal Keys Bids to Mom on Overall E xpression of Negative Emotion Model 1 Model 2 Fixed Effects Coefficient S.E Coefficient S.E Intercept 0.63 *** 0.15 0.64 *** 0.15 Gender 0.31 * 0.15 0.30 * 0.15 Verbal Keys Bids to Mom - 0.01 0.06 - 0.06 0.06 Linear ET 0.00 0.00 0.00 0.00 Quadratic ET - 0.00 0.00 - 0.00 0.00 Random Effects Variance Components SD Variance Components SD u 0j 0.72 *** 0.85 0.74 *** 0.86 u T - Vkeys - 2M , j - - 0.09 *** 0.31 e ij 0.32 0.56 0.32 0.56 Deviance 2917.97 2905.41 Note. Elapsed Time (ET) is in 10s intervals. * p < .05 , ** p < .01, *** p < .001, p < .10 Lagged Toddler Verbal Desire Bids to Mom . A two - level hierarchical model examined the effects of lagged Toddler Verbal Desire Bids to Mom Emotion (ENE - O) while controlling for ef fects of linear elapsed time, quadratic elapsed time, and toddler gender. The model was specified with the following equation: ENE - O ij 00 Gender (Gender j ) T - Vdes - 2M (T - Vdes - 2M i - 1j ) + ETL (ET ij ET 2 (ET 2 ij ) + u 0j + e ij Unexpectedly, lagged Toddler Verbal Desire Bids to Mom (e.g., ) wa s not associated with ENE - O (Table I9 boys (scale of 0 - 3) after accounting for effects for all predictors. Additionally, the current model was compared against a model (Mo del 2) in which slopes between lagged Toddler Verbal Desire Bids to Mom and ENE - O were allowed to vary between dyads. However, results of deviance tests ( = chi squared difference) indicated that allowing the relationship between lagged Toddler Verbal Desire Bids to Mom and ENE - O to vary among dyads would not significantly help to explain more variance in ENE - O between toddlers ( = 1.35 , df = 2, p > 0.05). 259 However, when the association between lagged Toddler Verbal Desire Bids to Mom and ENE - O was allowed to vary between dyads, the p - value associated with the random slope coefficient was marginally significant ( p < .10), suggesting tha t dyads varied in the association between lagged Toddler Verbal Desire Bids to Mom and ENE - O. Effects of lagged Toddler Verbal Desire Bids to Mom on ENE - O remain non - significant in Model 2; however, the association becomes negative as expected. Results of more parsimonious model are presented in Table I9. Table I9 Effects of Lagged Toddler Verbal Desire Bids to Mom on Overall Expression of Negative Emotion Fixed Effects Coefficient S.E Intercept 0.63 *** 0.15 Gender 0.31 * 0.15 Verbal Desire B ids to Mom 0.02 0.10 Linear ET 0.00 0.00 Quadratic ET - 0.00 0.00 Random Effects Variance Components SD u 0j 0.72 *** 0.85 e ij 0.32 0.57 Note. Elapsed Time (ET) is in 10s intervals. * p < .05, ** p < .01, *** p < .001, p < .10 Lagged Toddler Physical Comfort Seeking . A two - level hierarchical model examined Emotion (ENE - O) while controlling for effects of linear elapsed time, quadratic elapsed time, and toddler ge nder. The model was specified with the following equation: ENE - O ij 00 Gender (Gender j ) T - ComfSk ( T - ComfSk i - 1j ) + ETL (ET ij ET 2 (ET 2 ij ) + u 0j + e ij Unexpectedly, lagged Toddler Physical Comfort Seeking (e.g., reaching arms up to ) was not associated with EN E - O (Table I10). of 0 - 3) after accounting for effects for all predictors. Additionally, the current model was compared against a model (Model 2) in which slopes between lagged Toddler Physical Comfort 260 Seeking and ENE - O were allowed to vary between dyads. Results of deviance tests ( = chi squared difference) indicated that allowing the relationship between lagged Toddler Physical Comfort Seeking and ENE - O to vary among dyads would significantly improve model fit = 30.96, df = 2, p < 0.001). Indeed, when the assoc iation between lagged Toddler Physical Comfort Seeking and ENE - O was allowed to vary between dyads, the p - value associated with the random slope coefficient was significant ( p < .001), suggesting that dyads varied in the association between lagged Toddler Comfort Seeking and ENE - O. Effects of lagged Toddler Physical Comfort Seeking on ENE - O remain non - significant in Model 2. Table I10 Effects of Lagged Toddler Physical Comfort Seeking on Overall Expression of Negative Emotion Model 1 Model 2 Fixed Effe cts Coefficient S.E Coefficient S.E Intercept 0.63 *** 0.15 0.57 *** 0.15 Gender 0.31 * 0.15 0.36 * 0.15 Physical Comfort Seeking 0.01 0.05 0.04 0.06 Linear ET 0.00 0.00 0.00 0.00 Quadratic ET - 0.00 0.00 - 0.00 0.00 Random Effects Variance Components SD Variance Components SD u 0j 0.72 *** 0.85 0.77 *** 0.88 u T - ComfSk , j - - 0.17 *** 0.41 e ij 0.32 0.57 0.30 0.55 Deviance 2918.22 2887.26 Note. Elapsed Time (ET) is in 10s intervals. * p < .05, ** p < .01, *** p < .001, p < .10 Question 3A. Model results for non - significant predictors in Question 3A (V - Comf, M - JNT - Dis, M - VO2D, and AMM) are presented in Appendix I (Table I11 - I14). Lagged Maternal Verbal Comfort . A two - level hierar chical model examined the effects of lagged maternal Verbal Comfort on toddle 261 linear elapsed time, quadratic elapsed time, lagged maternal verbosity, and toddler gender. 39 The model was specified with the following equation: Overall Touch ij 00 Gender (Gender j ) V - Comf (V - Comf i - 1j ) + ETL (ET ij ET 2 (ET 2 ij Verb (Verb i - 1j ) + u 0j + e ij the expected direction (Table current model was compared against a model (Model 2) in which slopes between lagged maternal Verbal Comfort and Overall Touch were a llowed to vary between dyads. However, results of deviance tests ( = chi squared difference) indicated that allowing the relationship between lagged maternal Verbal Comfort and ENE - O to vary among dyads would not significantly help to explain more variance in ENE - O between toddlers ( = 0.21, df = 2, p > 0.05). Additionally, the p - value associated with random slope was not significant suggesting that the association between lagged maternal Verbal Comfort and Overall Touch does not vary between dyads. Effects of lagged maternal Verbal Comfort on Overall T ouch remain non - significant in Model 2. 39 L agged maternal verbosity and t oddler gender did not have significant main effects on Overall Touch in an y of the models for Questions 3A or 3B. Toddler gender was not correlated with Touch, Attempted Touch or Overall Touch but was retained in all models given previous findings in literature. Models for Overall Touch were compared with and without lagged mate rnal verbosity and results did not differ, but verbosity was retained in models for consistency in models for Questions 1 and 3. 262 Table I11 Effects of Lagged Maternal Verbal Comfort on Overall Touch Fixed Effects Coefficient S.E Intercept 0.57 *** 0.08 Gender - 0.02 0.06 Verbal Comfort - 0.004 0.03 Linear ET - 0.01 *** 0.00 Quadratic ET 0. 00 *** 0.00 Verbosity 0.01 0.01 Random Effects Variance Components SD u 0j 0.08 *** 0.28 e ij 0.28 0.53 Note. Elapsed Time (ET) is in 10s intervals. * p < .05, ** p < .01, *** p < .001, p < .10 Lagged Maternal Joined Distraction . A two - level hierarchical model examined the effects of lagged maternal Joined effects of linear elapsed time, quadratic elapsed time, and toddler gender. The model was specified with the following equation: Overall Tou ch ij 00 Gender (Gender j ) M - JNT - Dis ( M - JNT - Dis i - 1j ) + ETL (ET ij ET 2 (ET 2 ij ) + u 0j + e ij Unexpectedly, lagged maternal Joined Distraction (e.g., child plays with toes and mother Touch, but was in expected Additionally, the current model was compared against a model (Model 2) in which slopes between lagged maternal Joined Distraction and Ov erall Touch were allowed to vary between dyads. Results of deviance tests ( = chi squared difference) indicated that allowing the relationship between lagged maternal Joined Distraction and Overall Touch to vary among dyads would significantly improve model fit = 6.84, df = 2, p < 0.05. However, the p - value as sociated with random slope was not significant suggesting that the association between lagged 263 maternal Joined Distraction and Overall Touch does not vary between dyads. Effects of lagged maternal Joined Distraction on Overall Touch remain non - significant i n Model 2. Table I12 Effects of Lagged Maternal Joined Distraction on Overall Touch Model 1 Model 2 Fixed Effects Coefficient S.E Coefficient S.E Intercept 0.60 *** 0.11 0.60 *** 0.11 Gender - 0.02 0.06 - 0.02 0.06 Joined Distraction - 0.04 0.05 - 0.05 0 .04 Linear ET - 0.01 *** 0.00 - 0.01 *** 0.00 Quadratic ET 0.00 ** 0.00 0.00 ** 0.00 Random Effects Variance Components SD Variance Components SD u 0j 0.08 *** 0.28 0.08 *** 0.29 u M - JNT - Dis , j - - 0.03 0.17 e ij 0.28 0.53 0.28 0.53 Deviance 2466.92 2460.08 Note. Elapsed Time (ET) is in 10s intervals. * p < .05, ** p < .01, *** p < .001, p < .10 Lagged Maternal Verbal Orientation to Delay . A two - level hierarchical model examined Touch while controlling for effects of linear elapsed time, quadratic elapsed time, lagged maternal verbosity, and toddler gender. The model was specified with the following equation: Overall Touch ij 00 Gender (Gender j ) M - VO2D ( M - VO2D i - 1j ) + E TL (ET ij ET 2 (ET 2 ij Verb (Verb i - 1j ) + u 0j + e ij gender were not signifi cantly associated. Additionally, the current model was compared against a model (Model 2) in which slopes between lagged maternal Verbal Orientation to Delay and Overall Touch were allowed to vary between dyads. However, results of deviance tests ( = chi squared difference) indicated that allowing the relationship between lagged maternal Verbal Orientation to Delay and ENE - O to vary among dyads would not significantly help to explain more variance in ENE - O between 264 toddlers ( = 5.32, x, df = 2, p > 0.10). Additionally, the p - value associated with random slope was not significant suggesting that the association between lagged maternal Verbal Orientation to Delay and Overall Touch does not vary between dyads. Effects of lagged maternal Verb al Orientation to Delay on Overall Touch remain non - significant in Model 2. Table I13 Effects of Lagged Maternal Verbal Orientation to Delay on Overall Touch Fixed Effects Coefficient S.E Intercept 0.57 *** 0.11 Gender - 0.02 0.06 Verbal Orientation to Delay 0.01 0.01 Linear ET - 0.01 *** 0.00 Quadratic ET 0.00 ** 0.00 Verbosity 0.01 0.01 Random Effects Variance Components SD u 0j 0.08 *** 0.28 e ij 0.28 0.53 Note. Elapsed Time (ET) is in 10s intervals. * p < .05, ** p < .01, *** p < .001, p < .10 Lag ged Maternal Appropriate Mind - Related Comments . A two - level hierarchical model examined the effects of lagged maternal Appropriate Mind - Overall Touch while controlling for effects of linear elapsed time, quadratic elapsed time , lagged maternal verbosity, and toddler gender. The model was specified with the following equation: Overall Touch ij 00 Gender (Gender j ) AMM (AMM i - 1j ) + ETL (ET ij ET 2 (ET 2 ij Verb (Verb i - 1j ) + u 0j + e ij Unexpectedly, lagged maternal Ap propriate Mind - keys) was not associated with Overall Touch but were in expected direction (Table I14). ouch and gender were not significantly associated. Additionally, the current model was compared against a model (Model 2) in which slopes between lagged maternal Appropriate Mind - Related Comments and Overall Touch were allowed 265 to vary between dyads. Howeve r, results of deviance tests ( = chi squared difference) indicated that allowing the relationship between lagged maternal Appropriate Mind - Related Comments and Overall Touch to vary among dyads would not significantly help to explain more varianc e in Overall Touch between toddlers ( = 0.10, df = 2, p > 0.05). However, when the association between lagged maternal Appropriate Mind - Related Comments and Overall Touch was allowed to vary between dyads, the p - value associated with the random s lope coefficient was significant ( p <.05), suggesting that dyads varied in the association between lagged maternal Appropriate Mind - Related Comments and Overall Touch . Effects of lagged maternal Appropriate Mind - Related Comments on Overall Touch remain non - significant in Model 2. Results of more parsimonious model are presented in Table I14. Table I14 Effects of Lagged Appropriate Mind - Related Comments on Overall Touch Fixed Effects Coefficient S.E Intercept 0.57 *** 0.11 Gender - 0.02 0.06 Appropriate Mind - Related Comments - 0.007 0.04 Linear ET - 0.01 *** 0.00 Quadratic ET 0.00 ** 0.00 Verbosity 0.02 0.01 Random Effects Variance Components SD u 0j 0.08 *** 0.28 e ij 0.28 0.53 Note. Elapsed Time (ET) is in 10s intervals. * p < .05, ** p < .01, *** p < .00 1, p < .10 Question 3B. Model results for non - significant predictors (Pun, Min, NMM) are presented in Appendix I (Table I15 - I17). Lagged Maternal Punitive Reactions . A two - level hierarchical model examined the effects of lagged maternal Punitive Reactio effects of linear elapsed time, quadratic elapsed time, and toddler gender. The model was specified with the following equation: 266 Overall Touch ij 00 Gender (Gender j ) M - Pun ( M - Pun - 1j ) ETL ( ET ij ET 2 (ET 2 ij ) + u 0j + e ij nificantly associated. Additionally, the current model was compared against a model (Model 2) in which slopes between lagged maternal Punitive Reactions and Overall Touch were allowed to vary between dyads. However, results of deviance tests ( = chi squared difference) indicated that allowing the relationship between lagged maternal Punitive Reactions and Overall Touch to vary among dyads would not significantly help to explain more variance in Overall Touch between toddlers ( = 0.023, d f = 2, p > 0.05). Additionally, the p - value associated with random slope was not significant suggesting that the association between lagged maternal Punitive Reactions and Overall Touch does not vary between dyads. Effects of lagged maternal Punitive React ions on Overall Touch remain non - significant in Model 2 but directionality became negative as expected. Table I15 Effects of Lagged Punitive Reactions on Overall Touch Fixed Effects Coefficient S.E Intercept 0.60 *** 0.11 Gender - 0.02 0.06 Punitive Re actions - 0.003 0.06 Linear ET - 0.01 *** 0.00 Quadratic ET 0.00 ** 0.00 Random Effects Variance Components SD u 0j 0.08 *** 0.28 e ij 0.28 0.53 Note. Elapsed Time (ET) is in 10s intervals. * p < .05, ** p < .01, *** p < .001, p < .10 Lagged Maternal Minim izing . A two - level hierarchical model examined the effects of 267 elapsed time, quadratic elapsed time, lagged maternal verbosity, and toddler gender. The model was s pecified with the following equation: Overall Touch ij 00 Gender (Gender j ) Min (Min i - 1j ) + ETL (ET ij ET 2 (ET 2 ij Verb (Verb i - 1j ) + u 0j + e ij not associa significantly associated. Additionally, the current model was compared against a model (Model 2) in which slopes between lagged maternal Minimizing and Overall Touch were allow ed to vary between dyads. However, results of deviance tests ( = chi squared difference) indicated that allowing the relationship between lagged maternal Minimizing and Overall Touch to vary among dyads would not significantly help to explain more variance in Overall Touch between toddlers ( = 0.66 , df = 2, p > 0.05). However, when the association between lagged maternal Minimizing and Overall Touch was allowed to vary between dyads, the p - value associated with the random slope coefficient was significant ( p <.05), suggesting that dyads varied in th e association between lagged maternal Minimizing and Overall Touch . Effects of lagged maternal Minimizing on Overall Touch remain non - significant in Model 2. Results of more parsimonious model are presented in Table I16. Table I16 Effects of Lagged Minim izing on Overall Touch Fixed Effects Coefficient S.E Intercept 0.57 *** 0.11 Gender - 0.02 0.06 Minimizing - 0.03 0.05 Linear ET - 0.01 *** 0.00 Quadratic ET 0.00 ** 0.00 Verbosity 0.01 0.01 Random Effects Variance Components SD u 0j 0.08 *** 0.28 e ij 0. 28 0.53 Note. Elapsed Time (ET) is in 10s intervals. * p < .05, ** p < .01, *** p < .001, p < .10 268 Lagged Maternal Non - Attuned Mind - Related Comments . A two - level hierarchical model examined the effects of lagged maternal Non - Attuned Mind - Related Comments on time, lagged maternal verbosity, and toddler gender. The model was specified with the following equation: Overall Touch ij 00 Gender (Gender j ) NMM (N MM i - 1j ) + ETL (ET ij ET 2 (ET 2 ij Verb (Verb i - 1j ) + u 0j + e ij Unexpectedly, lagged maternal Non - Attuned Mind - and gend er were not significantly associated. Additionally, the current model was compared against a model (Model 2) in which slopes between lagged maternal Non - Attuned Mind - Related Comments and Overall Touch were allowed to vary between dyads. However, results of deviance tests ( = chi squared difference) indicated that allowing the relationship between lagged maternal Non - Attuned Mind - Related Comments and Overall Touch to vary among dyads would not significantly help to explain more variance in ENE - O between toddlers ( = 5.03, df = 2, p > 0.10). Additionally, the p - value associated with random slope was not significant suggesting that the association between lagged maternal Non - Attuned Mind - Related Comments and Overall Touch does not vary between dyads. Effects of lagged maternal Non - Attuned Mind - Related Comments on Overall Touch remained non - significant in Model 2. 269 Table I17 Effects of Lagged Maternal Non - Attuned Mind - Related Comments on Overall Touch Fixed Effects Coefficient S.E Intercept 0.57 *** 0.11 Ge nder - 0.02 0.06 Non - Attuned Mind - Related Comments - 0.06 0.04 Linear ET - 0.01 ** 0.00 Quadratic ET 0.00 ** 0.00 Verbosity 0.01 0.01 Random Effects Variance Components SD u 0j 0.08 *** 0.28 e ij 0.28 0.53 Note. Elapsed Time (ET) is in 10s intervals. * p < .05, ** p < .01, *** p < .001, p < .10 Question 4A. Model results for non - significant predictors in Question 4A ( T - INI - Vdes ) are presented in Appendix I (Table I18). Lagged Toddler Initiated Verbal Desire . A two - level hierarchical model examined the effe cts of lagged Toddler Initiated Verbal Desire model was specified with the following equation: Overall Touch ij = 00 Vocab (Vocab j T - INI - Vdes (T - INI - Vdes i - 1j ) + ETL (ET ij ET 2 (ET 2 ij ) + u 0j + e ij Unexpectedly, lagged Toddler Initiated Verbal Desire not associated with Overall Touch (Table I18). On average, less Over all Touch was marginally associated with higher scores on expressive language such that for every unit increase in vocabulary production, Overall Touch decreased by 0.002 in an average interval (mean per interval = .27, range per interval 0 - 6). Additiona lly, the current model was compared against a model (Model 2) in which slopes between lagged Toddler Initiated Verbal Desire and Overall Touch were allowed to vary between dyads. However, results of deviance tests ( = chi squared difference) indicated that allowing the relationship between lagged Toddler Initiated Verbal Desire and Overall Touch to vary among dyads would not significantly help to explain 270 more variance in Overall Touch between toddlers ( = 2.17, df = 2, p > 0.05). However, when the association between lagged Toddler Initiated Verbal Desire and Overall Touch was allowed to vary between dyads, the p - value associated with the random slope coefficient was significant (p <.05), suggesting that dyads varied in the association between lagged Toddler Initiated Verbal Desire and Overall Touch . Effects of lagged Toddler Initiated Verbal Desire on Overall Touch remain non - significant in Model 2. Results of more parsimonious model are presented i n Table I18 . Table I18 Effects of Lagged Toddler Initiated Verbal Desire on Overall Touch Fixed Effects Coefficient S.E Intercept 0.69 *** 0.13 Productive Vocab - 0.00 0.001 Initiated Verbal Desire 0.09 0.07 Linear ET - 0.01 ** 0.00 Quadratic ET 0.00 ** 0.00 Random Effects Variance Components SD u 0j 0.08 *** 0.28 e ij 0.28 0.53 Note. Elapsed Time (ET) is in 10s intervals. * p < .05, ** p < .01, *** p < .001, p < .10 Question 4B. Model results for non - significant predictors (T - JNT - VDis, T - JNT - nonVDi s, T - Vkeys - 2M, T - Vctrl - 2M, and T - Vdes - 2M are presented in Appendix I (Table I19 - I23). Lagged Toddler Joined Verbal Distraction. A two - level hierarchical model examined the effects of lagged Toddler Joined Verbal Distraction language. The model was specified with the following equation: Overall Touch ij 00 Vocab (Vocab j T - JNT - VDis (T - JNT - VDis i - 1j) + ETL (ET ij ET 2 ( ET 2 ij ) + u 0j + e ij Unexpectedly, lagged Toddler Joined Verbal Distraction (e.g., mother starts counting and child joins in ) was not associated with Overall Touch but was in the negative direction as 271 expected (Table I19). On average, less Overall Touch was marginally associated with higher scores on expressive language such that for every unit increase in vocabulary production, Overall Touch decreased by 0.002 in an average interval (mean per interval = .27, range per interval 0 - 6).Additionally, the curre nt model was compared against a model (Model 2) in which slopes between lagged Toddler Joined Verbal Distraction and Overall Touch were allowed to vary between dyads. However, results of deviance tests ( = chi squared difference) indicated that a llowing the relationship between lagged Toddler Joined Verbal Distraction and Overall Touch to vary among dyads would not significantly help to explain more variance in Overall Touch between toddlers ( = 0.11, df = 2, p > 0.05). Additionally, the p - value associated with random slope was not significant suggesting that the association between lagged Toddler Joined Verbal Distraction and Overall Touch does not vary between dyads. Effects of lagged Toddler Joined Verbal Distraction on Overall Touch r emain non - significant in Model 2. Table I19 Effects of Lagged Toddler Joined Verbal Distraction on Overall Touch Fixed Effects Coefficient S.E Intercept 0.69 *** 0.13 Productive Vocab - 0.00 0.001 Joined Verbal Distraction - 0.02 0.04 Linear ET - 0.01 * * 0.00 Quadratic ET 0.00 ** 0.00 Random Effects Variance Components SD u 0j 0.08 *** 0.28 e ij 0.28 0.53 Note. Elapsed Time (ET) is in 10s intervals. * p < .05, ** p < .01, *** p < .001, p < .10 Lagged Toddler Joined Non - Verbal Distraction . A two - level hierarchical model examined the effects of lagged Toddler Joined Non - expressive language. The model was specified with the following equation: Overall Touch ij 00 Vocab (Vocab j T - JNT - nonVDis (T - JNT - nonVDis i - 1j) + 272 ETL (ET ij ET 2 (ET 2 ij ) + u 0j + e ij Unexpectedly, lagged Toddler Joined Non - Verbal Distraction (e.g., mother initiates play and child joins in, mother points to a direction away from keys and child looks where she is pointing ) was not associated with Overall Touch but was in expected negative direction (Table I20). On average, less Overall Touch was marginally associated with higher scores on expr essive language such that for every unit increase in vocabulary production, Overall Touch decreased by 0.002 in an average interval (mean per interval = .27, range per interval 0 - 6).Additionally, the current model was compared against a model (Model 2) i n which slopes between lagged Toddler Joined Non - Verbal Distraction and Overall Touch were allowed to vary between dyads. However, results of deviance tests ( = chi squared difference) indicated that allowing the relationship between lagged Toddl er Joined Non - Verbal Distraction and Overall Touch to vary among dyads would not significantly help to explain more variance in Overall Touch between toddlers ( = 3.73, df = 2, p > 0.05). Additionally, the p - value associated with random slope was not significant suggesting that the association between lagged Toddler Joined Non - Verbal Distraction and Overall Touch does not vary between dyads. Effects of lagged Toddler Joined Non - Verbal Distraction on Overall Touch remain non - significant in Model 2. Table I20 Effects of Lagged Toddler Joined Non - Verbal Distraction on Overall Touch Fixed Effects Coefficient S.E Intercept 0.69 *** 0.13 Productive Vocab - 0.00 0.001 Joined Non - Verbal Distraction Linear ET - 0.01 ** 0.00 Quadratic ET 0.00 ** 0.00 Random Effects Variance Components SD u 0j 0.08 *** 0.28 e ij 0.28 0.53 Note. Elapsed Time (ET) is in 10s intervals. * p < .05, ** p < .01, *** p < .001, p < .10 273 Lagged Toddler Verbal Keys Bids to Mom . A two - level hierarchical model examined the effect s of lagged Toddler Verbal Keys language. The model was specified with the following equation: Overall Touch ij = 00 Vocab (Vocab j T - Vkeys - 2M (T - Vkeys - 2M i - 1j ) + ETL (ET ij ET 2 (ET 2 ij ) + u 0j + e ij Unexpectedly, lagged Toddler Verbal Keys Bids to Mom (e.g., e.g., response to mom ) was not associated with Overall Touch (Table I21). On average, less Overall Touch was marginally associated with higher scores on expressive language such that for every unit increase in vocabulary production, Overall Touch decreased by 0. 002 in an average interval (mean per interval = .27, range per interval 0 - 6). Additionally, the current model was compared against a model (Model 2) in which slopes between lagged Toddler Verbal Keys Bids to Mom and Overall Touch were allowed to vary b etween dyads. However, results of deviance tests ( = chi squared difference) indicated that allowing the relationship between lagged Toddler Verbal Keys Bids to Mom and Overall Touch to vary among dyads would not significantly help to explain m ore variance in Overall Touch between toddlers ( = 5.47, df = 2, p > 0.05). Additionally, the p - value associated with random slope was not significant suggesting that the association between lagged Toddler Verbal Keys Bids to Mom and Overall Tou ch does not vary between dyads. Effects of lagged Toddler Verbal Keys Bids to Mom on Overall Touch remain non - significant in Model 2. 274 Table I21 Effects of Lagged Toddler Verbal Keys Bids to Mom on Overall Touch Fixed Effects Coefficient S.E Inter cept 0.69 *** 0.13 Productive Vocab - 0.00 0.001 Verbal Keys Bids to Mom 0.01 0.03 Linear ET - 0.01 ** 0.00 Quadratic ET 0.00 ** 0.00 Random Effects Variance Components SD u 0j 0.08 *** 0.28 e ij 0.28 0.53 Note. Elapsed Time (ET) is in 10s intervals. * p < .05, ** p < .01, *** p < .001, p < .10 Lagged Verbal Self - Control - Bids to Mom . A two - level hierarchical model examined the effects of lagged Toddler Verbal Self - Control Bids to Mom controlling for effects of linear language. The model was specified with the following equation: Overall Touch ij 00 Vocab (Vocab j T - Vctrl - 2M (T - Vctrl - 2M i - 1j ) + ETL (ET ij ET 2 (ET 2 ij ) + u 0j + e ij Unexpectedly, lagged Toddler Self - Control - Bids to Mom (e.g., e.g., said while looking at or in response to mom ) was not associated wit h Overall Touch but was in expected negative direction (Table I22). On average, less Overall Touch was marginally associated with higher scores on expressive language such that for every unit increase in vocabulary production, Overall Touch decreased by 0. 002 in an average interval (mean per interval = .27, range per interval 0 - 6). Additionally, the current model was compared against a model (Model 2) in which slopes between lagged Toddler Verbal Self - Control Bids to Mom and Overall Touch were allowed t o vary between dyads. However, results of deviance tests ( = chi squared difference) indicated that allowing the relationship between lagged Toddler Verbal Self - Control Bids to Mom and Overall Touch to vary among dyads would not significantly h elp to explain more variance in Overall Touch between toddlers ( = 2.14, df = 2, p > 0.05). 275 Additionally, the p - value associated with random slope was not significant suggesting that the association between lagged Toddler Verbal Self - Control Bi ds to Mom and Overall Touch does not vary between dyads. Effects of lagged Toddler Verbal Self - Control Bids to Mom on Overall Touch remain non - significant in Model 2. Table I22 Effects of Lagged Toddler Verbal Self - Control Bids to Mom on Overall Touc h Fixed Effects Coefficient S.E Intercept 0.70 *** 0.13 Productive Vocab - 0.00 0.001 Verbal Self - Control Bids to Mom - 0.12 0.14 Linear ET - 0.01 ** 0.00 Quadratic ET 0.00 ** 0.00 Random Effects Variance Components SD u 0j 0.08 *** 0.28 e ij 0.28 0.5 3 Note. Elapsed Time (ET) is in 10s intervals. * p < .05, ** p < .01, *** p < .001, p < .10 Lagged Toddler Verbal Desire Bids to Mom . A two - level hierarchical model examined the effects of lagged Toddler Verbal Desire l Touch expressive language. The model was specified with the following equation: Overall Touch ij 00 Vocab (Vocab j T - Vdes - 2M (T - Vdes - 2M i - 1j ) + ETL (ET ij ET 2 (ET 2 ij ) + u 0j + e ij Unexpectedly, lagged Toddler Verbal Desire Bids to Mom (e.g., e.g., m ) was not associated with Overall Touch (Table I23). On average, less Overall Touch was marginally associated with higher scores on expressive language such that for every unit increase in vocabulary production, Overall Touch decreased by 0.002 in an aver age interval (mean per interval = .27, range per interval 0 - 6). Additionally, the current model was compared against a model (Model 2) in which slopes between lagged Toddler Verbal Desire Bids to Mom and Overall Touch were 276 allowed to vary between dyad s. Results of deviance tests ( = chi squared difference) indicated that allowing the relationship between lagged Toddler Verbal Desire Bids to Mom and Overall Touch to vary among dyads would significantly improve model fit = 10.92, df = 2, p < 0.01). Indeed, when the association between lagged Toddler Verbal Desire Bids to Mom and Overall Touch was allowed to vary between dyads, the p - value associated with the random slope coefficient was significant ( p <.05), suggesting that dyads varied in the association between lagged Toddler Verbal Desire Bids to Mom and Overall Touch . Effects of lagged Toddler Verbal Desire Bids to Mom on Overall Touch remain non - significant in Model 2. Table I23 Effects of Lagged Toddler Verbal Desire Bids to Mom on Overall Touch Model 1 Model 2 Fixed Effects Coefficient S.E Coefficient S.E Intercept 0.69 *** 0.13 0.70 *** 0.13 Productive Vocab - 0.00 0.001 - 0.00 0.001 Verbal Desire Bids to Mom 0.07 0.14 0.12 0.10 Linear ET - 0.01 ** 0.00 - 0.01 * ** 0.00 Quadratic ET 0.00 ** 0.00 0.00 ** 0.00 Random Effects Variance Components SD Variance Components SD u 0j 0.08 *** 0.28 0.08 *** 0.28 u T - Vdes - 2M , j - - 0.13 * 0.35 e ij 0.28 0.53 0.28 0.53 Deviance 2428.82 2417.90 Note. Elapsed Time (ET) is in 10 s intervals. * p < .05, ** p < .01, *** p < .001, p < .10 277 APPENDIX J Acknowledge ment of Grants 278 Acknowledge ment of Grants Data from this study were provided from the Early Head Start Research and Evaluation Project, which was supported by grant # 90YF0010, Pathways Project: Research into Directions for Family Health and Service Use, from the Administration on Children, Youth, and Families, Department of Health and Human Services, Rachel F. Schiffman, Ph.D., R.N., Principal Investigator, Michigan State University. The cross - site national evaluation data were collected under subcontract to Mathematica Policy Research, Inc., Princeton, NJ, which was responsible for the national Early Head Start program evaluation under contract 105 - 95 - 1936 with the A dministration for Children and Families, U.S. Department of Health and Human Services. Key MPR staff were John M. Love (Project Director), Ellen Kisker (Principal Investigator), and Jeanne Brooks - Gunn (Principal Investigator), under the supervision of the project manager for ACYF, Helen H. Raikes. Coding work was partially supported by the USDA National Institute of Food and Agriculture Hatch project MICL02393 . 279 REFERENCES 280 REFERENCES Ainsworth, M. D. S., Blehar, M. C., Waters, E., & Wall, S. (1978). Patterns of attachment: A psychological study of the strange situation. Hillsdale, NJ: Lawrence Erlbaum Associates. Altman DG (1991). Practical statistics for medical research. London: Chapman and Hall. Arnott, B., & Meins, E. (2007). L inks among antenatal attachment representations, postnatal mind - mindedness, and infant attachment security: A preliminary study of mothers and fathers. Bulletin of the Menninger Clinic, 71(2), 132 - 149. doi:10.1521/bumc.2007.71.2.132 Becker Razuri, E., Hil es Howard, A. R., Purvis, K. B., & Cross, D. R. (2017). Mental state language development: The longitudinal roles of attachment and maternal language. Infant Mental Health Journal , 38 (3), 329 - 342. Bell, M. A., & Deater - Deckard, K. D. (2007). Biological sy stems and the development of self regulation: Integrating behavior, genetics, and psychophysiology. Journal of Developmental and Behavioral Pediatrics , 28, 409 420. doi:10.1097/DBP.0b013e3181131fc7 Bernier, A., Carlson, S. M. and Whipple, N. (2010), From External Regulation to Self Child Development, 81: 326 339. doi: 10.1111/j.1467 - 8624.2009.01397.x Bernier, A., McMahon, C. A., & Perrier, R. (2017). Maternal mind - mindedness school readiness: A longitudinal study of developmental processes. Developmental Psychology, 53 (2), 210 - 221. Blair, C. (2002). School readiness - integrating cognition and emotion in a neurobiological conceptualization of children's funct ioning at school entry. American Psychologist, 57(2), 111 - 127. doi:10.1037//0003 - 066X.57.2.111 Blair, C., & Raver, C. C. (2012). Child development in the context of adversity: Experiential canalization of brain and behavior. The American Psychologist, 67 ( 4), 309 - 318. Bocknek, E. L., Brophy Herb, H. E., & Banerjee, M. (2009). Effects of parental supportiveness on toddlers' emotion regulation over the first three years of life in a low income African American sample. Infant Mental Health Journal , 30 (5), 45 2 - 476. Bowlby, J. (1969). Attachment and loss. Vol 1. Attachment. New York: Basic Books. Bridgett, D., Burt, N., Edwards, E., & Deater - Deckard, K. (2015). Intergenerational transmission 281 of self - regulation: A multidisciplinary review and integrative conce ptual framework. Psychological Bulletin, 141(3), 602 - 654. doi:10.1037/a0038662 Brophy - Herb, H. E., Bocknek, E. L., Vallotton, C. D., Stansbury, K. E., Senehi, N., Dalimonte Merckling, D., & Lee, Y. (2015). Toddlers with early behavioral problems at higher family demographic risk benefit the most from maternal emotion talk. Journal of Developmental and Behavioral Pediatrics : JDBP, 36(7), 512 - 520. Brophy - Herb, H., Stansbury, K., Bocknek , E., & Horodynski, M. (2012). Modeling maternal emotion - related social ization behaviors in a low - income sample: Relations with toddlers' self - regulation. Early Childhood Research Quarterly, 27(3), 352 - 364. doi:10.1016/j.ecresq.2011.11.005 Buckner, J. C., Mezzacappa, E., & Beardslee, W. R. (2003). Characteristics of resilien t youths living in poverty: The role of self - regulatory processes. Development and Psychopathology, 15(1), 139 - 162. doi:10.1017/S0954579403000087 Buckner, J. C., Mezzacappa, E., & Beardslee, W. R. (2009). Self - regulation and its relations to adaptive func tioning in low income youths. American Journal of Orthopsychiatry, 79(1), 19 - 30. doi:10.1037/a0014796 Burman, J. T., Green, C. D. and Shanker, S. (2015), On the Meanings of Self - Regulation: Digital Humanities in Service of Conceptual Clarity. Child Develo pment, 86: 1507 1521. doi: 10.1111/cdev.12395 Byrk, A., & Raudenbush, S. (1992). Hierarchical linear models. Thousand Oaks, Ca: Sage. Calkins, S. D., & Dedmon, S. E. (2000). Physiological and behavioral regulation in two - year - old children with aggressiv e/destructive behavior problems. Journal of abnormal child psychology , 28 (2), 103 - 118. Calkins, S. D., Graziano, P. A., Berdan, L. E., Keane, S. P., & Degnan, K. A. (2008). Predicting cardiac vagal regulation in early childhood from maternal child relatio nship quality during toddlerhood. Developmental psychobiology , 50 (8), 751 - 766. Calkins, S. D., & Johnson, M. C. (1998). Toddler regulation of distress to frustrating events: Temperamental and maternal correlates. Infant Behavior and Development, 21(3), 37 9 395. doi:10.1016/S0163 - 6383(98)90015 - 7 Carlson, S. M. (2003), Executive Function in Context: Development, Measurement, Theory, and Experience. Monographs of the Society for Research in Child Development, 68: 138 151. doi: 10.1111/j.1540 - 5834.2003.068030 12.x Cassidy, J.. (1994). Emotion Regulation: Influences of Attachment Relationships. Monographs of the Society for Research in Child Development, 59(2/3), 228 249. http://doi.org/10.2307/1166148 282 Centifanti, L. C.M., Meins, E. and Fernyhough, C. (2016), C allous - unemotional traits and impulsivity: distinct longitudinal relations with mind - mindedness and understanding of others. Journal of Child Psychology and Psychiatry, 57: 84 92. doi: 10.1111/jcpp.12445 Cicchetti, D., Ackerman, B. P., & Izard, C. E. (199 5). Emotions and emotion regulation in developmental psychopathology. Development and psychopathology , 7 (1), 1 - 10. Cline, R. J., Harper, F. W., Penner, L. A., Peterson, A. M., Taub, J. W., & Albrecht, T. L. (2006). Parent communication and child pain and d istress during painful pediatric cancer treatments. Social science & medicine , 63 (4), 883 - 898. Cohen, J. D. (1960). A coefficient of agreement for nominal scales. Educational and Psychological Measurement, 20(1), 37 46. Cole, P., Tan, P., Hall, S., Zhang, Y., Crnic, K., Blair, C., & Li, R. (2011). Developmental changes in anger expression and attention focus: Learning to wait. Developmental Psychology, 47(4), 1078 - 1089. doi:10.1037/a0023813 Cole, P. M., Martin, S. E., & Dennis, T. A. (2004). Emotion regul ation as a scientific construct: Methodological challenges and directions for child development research. Child Development, 75(2), 317 - 333. doi:10.1111/j.1467 - 8624.2004.00673.x Cole, P. M. (2014). Moving ahead in the study of the development of emotion r egulation. International Journal of Behavioral Development, 38(2), 203 - 207. doi:10.1177/0165025414522170 Cole, P. M., Armstrong, L. M., & Pemberton, C. K. (2010). The role of language development of emotion regulation. In S. D. Calkins (Ed.) & M. A. Bell (Ed.), Child development at the intersection of emotion and cognition.(pp. 59 - 77). Washington, DC: American Psychological Association. Cole, P. M., Dennis, T. A., Smith - Simon, K. E. and Cohen, L. H. (2009), Preschoolers' Emotion Regulation Strategy Under standing: Relations with Emotion Socialization and Child Self - regulation. Social Development, 18: 324 352. doi: 10.1111/j.1467 - 9507.2008.00503.x Davidov, M. and Grusec, J. E. (2006), Untangling the Links of Parental Responsiveness to Distress and Warmth t o Child Outcomes. Child Development, 77: 44 58. doi: 10.1111/j.1467 - 8624.2006.00855.x Demers, I., Bernier, A., Tarabulsy, G. M., & Provost, M. A. (2010). Mind - mindedness in adult and adolescent mothers: Relations to maternal sensitivity and infant attachm ent. International Journal of Behavioral Development, 34 (6), 529 - 537. doi:10.1177/0165025410365802 Demers, I., Bernier, A., Tarabulsy, G., & Provost, M. (2010). Maternal and child characteristics as antecedents of maternal mind - mindedness. Infant Mental H ealth Journal, 31 (1), 94 - 112. doi:10.1002/imhj.20244 283 Diener, M. L., & Mangelsdorf, S. C. (1999). Behavioral strategies for emotion regulation in toddlers: Associations with maternal involvement and emotional expressions. Infant Behavior and Development, 22 (4), 569 - 583. doi:10.1016/S0163 - 6383(00)00012 - 6 Eiden, R., Edwards, E., & Leonard, K. (2007). A conceptual model for the development of externalizing behavior problems among kindergarten children of alcoholic families: Role of parenting and children's sel f - regulation. Developmental Psychology, 43(5), 1187 - 1201. doi:10.1037/0012 - 1649.43.5.1197 Eisenberg, N., Champion, C., & Ma, Y. (2004). Emotion - related regulation: An emerging construct. Merrill - Palmer Quarterly (1982 - ), 50(3), 236 - 259. doi:10.1353/mpq.20 04.0016 Eisenberg, N., Cumberland, A., & Spinrad, T. L. (1998). Parental socialization of emotion. Psychological Inquiry, 9(4), 241 - 273. doi:10.1207/s15327965pli0904_1 Eisenberg, N., Cumberland, A., Spinrad, T. L., Fabes, R. A., Shepard, S. A., Reiser, M ., ... & Guthrie, I. K. (2001). The relations of regulation and emotionality to children's externalizing and internalizing problem behavior. Child development , 72 (4), 1112 - 1134. Eisenberg, N., & Morris, A. S. (2002). Children's emotion - related regulation. In H.Reese, & R.Kail (Eds.), Advances in child development and behavior (Vol. 30, pp. 189 229). San Diego, CA: Academic Press. Eisenberg, N., Morris, A. S. and Spinrad, T. L. (2005) Emotion - Related Regulation: The Construct and its Measurement, in Handb ook of Research Methods in Developmental Science (ed D. M. Teti), Blackwell Publishing Ltd, Oxford, UK. doi: 10.1002/9780470756676.ch21 Eisenberg, N., Spinrad, T. L. (2004). Emotion - related regulation: Sharpening the definition. Child Development, 75(2), 334 - 339. doi:10.1111/j.1467 - 8624.2004.00674.x Eisenberg, N., Spinrad, T., & Eggum, N. (2010a). Emotion - related self - regulation and its relation to children's maladjustment. (pp. 495 - 525). PALO ALTO: ANNUAL REVIEWS. doi:10.1146/annurev.clinpsy.121208.13120 8 (2010b). Relations among maternal socialization, effortful control, and maladjustment in early childhood. Development and Psychopathology, 22(3), 507 525. h ttp://doi.org/10.1017/S0954579410000246 Eisenberg, N., Sadovsky, A., & Spinrad, T. L. (2005). Associations of emotion - related regulation with language skills, emotion knowledge, and academic outcomes. New Directions for Child and Adolescent Development, 2 005(109), 109 - 118. doi:10.1002/cd.143 Eisenberg, N., Valiente, C., & Sulik, M. (2009). How the study of regulation can inform the study of coping. (pp. 75 - 86). San Francisco: Wiley Periodicals. doi:10.1002/cd.244 284 E kas, N. V., Braungart - Rieker, J. M., Lick enbrock, D. M., Zentall, S. R. and Maxwell, S. M. (2011), Toddler Emotion Regulation With Mothers and Fathers: Temporal Associations Between Negative Affect and Behavioral Strategies. Infancy, 16: 266 294. doi: 10.1111/j.1532 - 7078.2010.00042.x Ensor, R., Devine, R. T., Marks, A., & Hughes, C. (2014). Mothers' cognitive references to 2 Year Olds predict theory of mind at ages 6 and 10. Child Development, 85(3), 1222 - 1235. doi:10.1111/cdev.12186 Feldman, R., Dollberg, D., & Nadam, R. (2011). The expression and regulation of anger in toddlers: Relations to maternal behavior and mental re presentations. Infant Behavior and Development, 34(2), 310 - 320. doi:10.1016/j.infbeh.2011.02.001 Fenson, L., Dale, P. S., Reznick, J. S, Thal, D., Bates, E., Hartung, J. P., Pethick, S., & Reilly, J. S. (1993). The MacArthur Communicative Development Inven technical manual. San Diego, CA: Singular Publishing Group. Fivush, R., & Baker - Ward, L. (2005). The search for meaning: Developmental perspectives on internal state language in autobiographical memory. Journal of Cognition and De velopment, 6(4), 455 - 462. doi:10.1207/s15327647jcd0604_1 Fonagy, P., & Target, M. (1997). Attachment and reflective function: Their role in self - organization. Development and Psychopathology, 9(4), 679 - 700. Retrieved from http://ezproxy.msu.edu/login?url= http://search.proquest.com/docview/619173200?accoun tid=12598 Fonagy, P., & Target, M. (2002). Early intervention and the development of self - regulation. Psychoanalytic Inquiry, 22(3), 307 - 335. doi:10.1080/07351692209348990 Fonagy, P. (2004), Early - Life Trauma and the Psychogenesis and Prevention of Violence. Annals of the New York Academy of Sciences, 1036: 181 200. doi: 10.1196/annals.1330.012 Fonagy, P. (2006) The Mentalization - Focused Approach to Social Development, in Handbook of Mentalization - Base d Treatment (eds J. G. Allen and P. Fonagy), John Wiley & Sons, Ltd, Chichester, UK. doi: 10.1002/9780470712986.ch3 Fonagy, P., Gergely, G., Jurist E. L., & Target M. (2002). Affect Regulation, Mentalization, and the Development of the Self. Karnac; Londo n: 2002. Fonagy, P., Steele, M., Steele, H., Moran, G., & Higgitt, A. (1991). The capacity for understanding mental states - the reflective self in parent and child and its significance for security of attachment. Infant Mental Health Journal, 12(3), 201 - 218. doi:10.1002/1097 - 0355(199123)12:3<201::AID - IMHJ2280120307>3.0.CO;2 - 7 285 Fox, N. A., & Calkins, S. D. (2003). The development of self - control of emotion: Intrinsic and extrinsic influences. Motivation and Emotion, 27 (1), 7 - 26. doi:10.1023/A:1023622324898 Freelon, D. (2013). ReCal OIR: Ordinal, interval, and ratio inter coder reliability as a web service. International Journal of Internet Science, 8(1), 10 - 16. Friedlmeier, W., Corapci, F., & Benga, O. (2015a). Coding system for maternal regulatory attemp ts in a delay situation with toddlers. Unpublished instrument. Grand Valley State University, MI. Friedlmeier, W., Corapci, F., & Benga, O. (2015b). Coding System for Delay of Gratification Task Child Behavior and Regulatory Attempts. Unpublished instru ment. Grand Valley State University, MI. Friedlmeier, W., Corapci, F., & Benga, O. (2015c). Coding System for Delay of Gratification Task - Intensity of Emotion Expression. Unpublished instrument. Grand Valley State University, MI. Gagné C, Bernier A, McMah on CA. (2018). The role of paternal mind mindedness in preschoolers' self regulated conduct. Inf Child Dev . (e2081) 1 - 12 . https://doi - org.proxy2.cl.msu.edu/10.1002/icd.2081 Gilliom, M., Shaw, D., Beck, J., Schonberg, M., & Lukon, J. (2002). Anger regulation in disadvantaged preschool boys: Stra tegies, antecedents, and the development of self - control. Developmental Psychology , 38(2), 222 - 235. doi:10.1037//0012 - 1649.38.2.222 Gottman, J. M., Katz, L. F., & Hooven, C. (1996). Parental meta - emotion philosophy and the emotional life of families: Theo retical models and preliminary data. Journal of Family Psychology, 10 (3), 243 - 268. doi:10.1037/0893 - 3200.10.3.243 Graziano, P. A., Calkins, S. D., & Keane, S. P. (2011). Sustained attention development during the toddlerhood to preschool period: Associati ons with toddlers' emotion regulation strategies and maternal behavior. Infant and Child Development, 20(6), 389 - 408. doi:10.1002/icd.731 Grolnick, W. S., Kurowski, C. O., McMenamy, J. M., Rivkin, I., & Bridges, L. J. (1998). Mothers' strategies for regul ating their toddlers' distress. Infant Behavior and Development , 21(3), 437 - 450. doi:10.1016/S0163 - 6383(98)90018 - 2 Grolnick, W. S., Bridges, L. J. and Connell, J. P. (1996) . Emotion Regulation in Two - Year - Olds: Strategies and Emotional Expression in Four Contexts. Child Development , 67: 928 941. doi: 10.1111/j.1467 - 8624.1996.tb01774.x Howe, N., Rinaldi, C. M., & Recchia, H. E. (2010). Patterns in mother - child internal state discourse across four contexts. Merrill - Palmer Quarterly , 56(1), 1 - 20. doi:10.1353 /mpq.0.0042 286 Kopp, C. B. (1989). Regulation of distress and negative emotions: A developmental view. Developmental Psychology , 25(3), 343 - 354. doi:10.1037/0012 - 1649.25.3.343 Kopp, C. B. (1982). Antecedents of self - regulation: A developmental perspective. D evelopmental Psychology , 18(2), 199 - 214. doi:10.1037/0012 - 1649.18.2.199 Koren - Karie, N., Oppenheim, D., Dolev, S., Sher, E., & Etzion - Carasso, A. (2002). Mothers' insightfulness regarding their infants' internal experience: Relations with maternal sensiti vity and infant attachment. Developmental Psychology, 38 (4), 534 - 542. doi:http://dx.doi.org/10.1037/0012 - 1649.38.4.534 Krippendorff, K. (1970). Bivariate agreement coefficients for reliability of data. Sociological Methodology , 2 , 139 - 150. Krippendorff, K . (2004). Content analysis: An introduction to its methodology (2 nd ed.). Thousand Oaks, CA: Sage. Laranjo, J., Bernier, A., Meins, E., & Carlson, S. (2010). Early manifestations of children's theory of mind: The roles of maternal mind - mindedness and infa nt security of attachment. Infancy, 15 (3), 300 - 323. doi:10.1111/j.1532 - 7078.2009.00014.x Laranjo, J., Bernier, A., Meins, E., & Carlson, S. (2014). The roles of maternal mind - mindedness and infant security of attachment in predicting preschoolers' understanding of visual perspective taking and false belief. Journal of Experimental Child Psycho logy, 125 , 48 - 62. doi:10.1016/j.jecp.2014.02.005 Lengua, L. J., & Long, A. C. (2002). The role of emotionality and self - regulation in the appraisal coping process: Tests of direct and moderating effects. Journal of Applied Developmental Psychology, 23 (4), 471 - 493. doi:10.1016/S0193 - 3973(02)00129 - 6 Lengua, L., & Sandler, I. (1996). Self - regulation as a moderator of the relation between coping and symptomatology in children of divorce. Journal of Abnormal Child Psychology, 24(6), 681 - 701. doi:10.1007/BF0166 4734 Lengua, L. J. (2002). The contribution of emotionality and self - regulation to the understanding - 161. doi:10.1111/1467 - 8624.00397 Love, J. M., Kisker, E. E., Ross, C., Raikes, H., Constantine, J., Boller, K.. . Vogel, C. (2005). The effectiveness of early head start for 3 - year - old children and their parents: Lessons for policy and programs. Developmental Psychology, 41(6), 885 - 901. doi:10.1037/0012 - 1649.41.6.885 Love, J. M., Kiske r, E. E., Ross, C. M., Schochet, P. Z., Brooks - - Smith, C. (2002). Making a difference in the lives of infants and toddlers and their families: The impacts of Early Head Start. Princeton, NJ: Mathematica Policy Research. 287 Lyons - Ruth, K., Bronfman, E., & Parsons, E. (1999). Maternal frightened, frightening, or atypical behavior and disorganized infant attachment patterns. Monographs of the Society for Research in Child Development , 67 - 96. Madigan, S., Moran, G., Schuengel, C., P ederson, D. R., & Otten, R. (2007). Unresolved maternal attachment representations, disrupted maternal behavior and disorganized attachment in infancy: Links to toddler behavior problems. Journal of Child Psychology and Psychiatry , 48 (10), 1042 - 1050. McCl elland, M. M., & Cameron, C. E. (2012). Self Regulation in early childhood: Improving conceptual clarity and developing ecologically valid measures. Child Development Perspectives, 6(2), 136 - 142. doi:10.1111/j.1750 - 8606.2011.00191.x MedCalc Statistical Sof tware version 16.2.0 (MedCalc Software bvba, Ostend, Belgium; https://www.medcalc.org; 2016) Meins, E., & Fernyhough, C. (1999). Linguistic acquisitional style and mentalising development: The role of maternal mind - mindedness. Cognitive Development, 14(3) , 363 - 380. doi:10.1016/S0885 - 2014(99)00010 - 6 Meins, E., & Fernyhough, C. (2015). Mind - mindedness coding manual, Version 2.2. Unpublished manuscript. Durham University, Durham, UK. Meins, E. (1998). The effects of security of attachment and material attri bution of meaning on children's linguistic acquisitional style. Infant Behavior and Development, 21(2), 237 - 252. doi:10.1016/S0163 - 6383(98)90004 - 2 Meins, E., Centifanti, L., Fernyhough, C., & Fishburn, S. (2013). Maternal mind - mindedness and children's be havioral difficulties: Mitigating the impact of low socioeconomic status. Journal of Abnormal Child Psychology, 41 (4), 543 - 553. doi:10.1007/s10802 - 012 - 9699 - 3 Meins, E., Fernyhough, C., & Harris - Waller, J. (2014). Is mind - mindedness trait - like or a quality of close relationships? evidence from descriptions of significant others, famous people, and works of art. Cognition, 130(3), 417 - 427. doi:10.1016/j.cognition.2013.11.009 Meins, E., Fernyhough, C., Arnott, B., Turner, M., & Leekam, S. R. (2011). Mother versus Infant Centered correlates of maternal Mind Mindedness in the first year of life. Infancy, 16(2), 137 - 165. doi:10.1111/j.1532 - 7078.2010.00039.x Meins, E., Fernyhough, C., Arnott, B., Turner, M. and Leekam, S. R. (2011), Mother - Versus Infant - Center ed Correlates of Maternal Mind - Mindedness in the First Year of Life. Infancy, 16: 137 165. doi: 10.1111/j.1532 - 7078.2010.00039.x Meins, E., Fernyhough, C., de Rosnay, M., Arnott, B., Leekam, S. R., & Turner, M. (2012). Mind Mindedness as a multidimensiona l construct: Appropriate and nonattuned Mind Related comments independently predict Infant Mother attachment in a socially diverse sample. Infancy, 17(4), 393 - 415. doi:10.1111/j.1532 - 7078.2011.00087.x 288 Meins, E., Fernyhough, C., Fradley, E., & Tuckey, M. (2 001). Rethinking maternal sensitivity: Mothers' comments on infants' mental processes predict security of attachment at 12 months. The Journal of Child Psychology and Psychiatry and Allied Disciplines, 42(5), 637 - 648. doi:10.1017/S0021963001007302 Meins, E., Fernyhough, C., Russell, J. and Clark - Carter, D. (1998), Security of Attachment as a Predictor of Symbolic and Mentalising Abilities: A Longitudinal Study. Social Development, 7: 1 24. doi: 10.1111/1467 - 9507.00047 Meins, E., Fernyhough, C., Wainwright , R., Das Gupta, M., Fradley, E., & Tuckey, M. (2002). Maternal mind mindedness and attachment security as predictors of theory of mind understanding. Child development, 73(6), 1715 - 1726. doi:10.1111/1467 - 8624.00501 Morris, A. S., Silk, J. S., Steinberg, L., Myers, S. S., & Robinson, L. R. (2007). The role of the family context in the development of emotion regulation. Social Development, 16(2), 361 - 388. doi:10.1111/j.1467 - 9507.2007.00389.x Morris, A. S., Silk, J. S., Morris, M. D. S., Steinberg, L., Auco in, K. J., & Keyes, A. W. (2011). The influence of mother - child emotion regulation strategies on children's expression of anger and sadness.Developmental Psychology, 47(1), 213 - 225. doi:10.1037/a0021021 Oppenheim, D. and Koren - Karie, N. (2002), Mothers' i nsightfulness regarding their children's internal worlds: The capacity underlying secure child mother relationships. Infant Ment. Health J., 23: 593 605. doi: 10.1002/imhj.10035 Posner, M. I., & Rothbart, M. K. (2000). Developing mechanisms of self - regula tion. Development and Psychopathology, 12(3), 427 - 441. doi:10.1017/S0954579400003096 Premo, J. E., & Kiel, E. J. (2014). The effect of toddler emotion regulation on maternal emotion socialization: Moderation by toddler gender. Emotion, 14(4), 782 - 793. doi :http://dx.doi.org.proxy2.cl.msu.edu/10.1037/a0036684 Putnam, S. P., Spritz, B. L., & Stifter, C. A. (2002). Mother Child coregulation during delay of gratification at 30 months. Infancy, 3(2), 209 - 225. doi:10.1207/S15327078IN0302_6 Raikes, H., Robinson, J., Bradley, R., Raikes, H., & Ayoub, C. (2007). Developmental trends in self - regulation among low - income toddlers. Social Development, 16(1), 128 - 149. doi:10.1111/j.1467 - 9507.2007.00375.x Raudenbush, S., Byrk, A., Cheong, Y.F., Congdon, R., & du Toit, M. (2011). HLM7: Hierarchical Linear and Nonlinear Modeling. Lincolnwood, IL: Scientific Software International, Inc. Raudenbush, S.W., Bryk, A.S, & Congdon, R. (2013). HLM 7.01 for Windows [Computer software]. Skokie, IL: Scientific Software International, Inc. 289 Roben, C. K. P., Cole, P. M., & Armstrong, L. M. (2013). Longitudinal relations among language skills, anger expression, and regulatory strategies in early childhood. Child Development, 84(3), 891 - 905. doi:10.1111/cdev.12027 Rothbart, M. K., & Bates, J. E. (2006). Temperament. In W. Damon & R. Lerner (Series Eds.) & W. Damon & R. Lerner (Vol. Eds.), Handbook of child psychology: Vol. 3. Social, emotional, and personality development (6th ed., p p. 99 166). Hoboken, NJ: Wiley. Schel, M., Scheres, A. P. J., & Crone, E. A. M. (2014). New perspectives on self - control development: Highlighting the role of intentional inhibition. Neuropsychologia, 65 , 236 - 246. doi:10.1016/j.neuropsychologia.2014.08.022 Schore, J. R., & Schore, A. N. (2008). Modern attachment theory: The central role of affect regulation in development and treatment. Clinical Social Work Journal, 36(1), 9 - 20. doi:http://dx.doi.org/10.1007/s10615 - 007 - 0111 - 7 Senehi, N., Brophy - Herb, H.E., & Vallotton, C. D. (2018). Effects of maternal mentalization - - regulation. Early Childhood Research Quarterly.44 , 1 - 14. https://doi.org /10.1016/j.ecresq.2018.02.001 Sharp, C., & Fonagy, P. (2008). The parent's capacity to treat the child as a psychological agent: Constructs, measures and implications for developmental psychopathology. Social Development, 17(3), 737 - 754. doi:10.1111/j.146 7 - 9507.2007.00457.x Slade, A., Grienenberger, J., Bernbach, E., Levy, D., & Locker, A. (2005). Maternal reflective functioning, attachment, and the transmission gap: A preliminary study. Attachment & Human Development, 7(3), 283 - 298. doi:10.1080/146167305 00245880 Smith, C. L., Calkins, S. D., & Keane, S. P. (2006). The relation of maternal behavior and attachment security to toddlers' emotions and emotion regulation. Research in Human Development, 3(1), 21 - 31. doi:10.1207/s15427617rhd0301_3 Snijders, T. A. B. & Bosker, R. J. (2012). Multilevel Analysis: An introduction to basic and advanced multilevel modeling (2nd ed.). London: SAGE Publications Inc. Spinrad, T. L., Eisenberg, N. and Gaertner, B. M. (2007a), Measures of effortful regulation for young c hildren. Infant Ment. Health J., 28: 606 626. doi: 10.1002/imhj.20156 Spinrad, T. L., Eisenberg, N., Gaertner, B., Popp, T., Smith, C. L., Kupfer, A., . . . Hofer, C. (2007b). Relations of maternal socialization and toddlers' effortful control to children 's adjustment and social competence. Developmental Psychology, 43(5), 1170 - 1186. doi:10.1037/0012 - 1649.43.5.1170 Spinrad, T. L., Eisenberg, N., Gaertner, B., Popp, T., Smith, C. L., Kupfer, A.. . Hofer, C. (2007). Relations of maternal socialization and t oddlers' effortful control to children's adjustment and social competence. Developmental Psychology, 43(5), 1170 - 1186. doi:10.1037/0012 - 1649.43.5.1170 290 Spinrad, T. L., Stifter, C. A., Donelan - Strategies - Regulation. Social Development, 13 : 40 55. doi: 10.1111/j.1467 - 9507.2004.00256.x Swingler, M. M., Perry, N. B., & Calkins, S. D. (2015). Neural plasticity and the development of attention: Intri nsic and extrinsic influences. Development and Psychopathology, 27(2), 443 - 457. doi:10.1017/S0954579415000085 Symons, D. K., Fossum, K. - L. M. and Collins, T.B. K. (2006), A Longitudinal Study of Belief and Desire State Discourse During Mother Child Play a nd Later False Belief Understanding. Social Development, 15 : 676 692. doi: 10.1111/j.1467 - 9507.2006.00364.x Tabachnick, B. G., & Fidell, L. S. (2007). Using multivariate statistics, 5 th ed . Boston: Allyn & Bacon/Pearson Education. Taumoepeau, M. and Ruff man, T. (2006), Mother and Infant Talk About Mental States Relates to Desire Language and Emotion Understanding. Child Development, 77: 465 481. doi: 10.1111/j.1467 - 8624.2006.00882.x Maternal Talk Relates to Child Mental State Language and Emotion Understanding at 15, 24, and 33 Months. Child Development, 79: 284 302. doi: 10.1111/j.1467 - 8624.2007.01126.x Thompson, R. A. (1994). Emotion regulation: A theme in search of definition. Mo nographs of the Society for Research in Child Development, 59(2/3), 25 - 52. doi:10.1111/j.1540 - 5834.1994.tb01276.x Tronick, E. Z. (1989). Emotions and emotional communication in infants. American Psychologist, 44(2), 112 - 119. Retrieved from http://ezproxy. msu.edu/login?url=http://search.proquest.com/docview/63154454?account id=12598 Vallotton, C., & Ayoub, C. (2011 ) Use your words: The role of language in the development of - regulation. Early Childhood Research Quarterly, 26(2), 169 - 181. doi:1 0.1016/j.ecresq.2010.09.002