MOMENTARY INTERPERSONAL BEHAVIORS AND PHYSICAL INDICATORS OF MOMENTARY ANXIETY: A TEST OF INTERPERSONAL COMPLEMENTARITY By Nick Schade A THESIS Submitted to Michigan State University in partial fulfillment of the requirement for the degree of Psychology –Master of Arts 2013 ABSTRACT MOMENTARY INTERPERSONAL BEHAVIORS AND PHYSICAL INDICATORS OF MOMENTARY ANXIETY: A TEST OF INTERPERSONAL COMPLEMENTARITY By Nick Schade Behavioral variability in interpersonal interactions can be structured by two dimensions: communion (warmth to coldness) and agency (dominance to submissiveness). These interactions typically display complementarity – interactants are similar in warmth and opposite in dominance – and interactions with more complementarity are associated with more positive outcomes. From an interpersonal perspective, individuals employ behaviors that minimize anxiety. Therefore, anxiety is one possible result of deviation from complementarity. To test this hypothesis at the behavioral level, I evaluated the association between different indicators of affect and interpersonal behaviors across different time intervals in an interaction between a participant and a confederate. Deviation from warmth complementarity within the interaction was associated with increased negative affect. Similar effects were not identified for deviations from dominance complementarity. In addition, participants’ cold behavior was significantly related to participants’ skin conductance within the interaction, and liking the confederate less, and reporting greater negative affect. A number of null results across tests of hypotheses regarding physiological reactions and regarding dominance complementarity provide new direction for future analysis and research. TABLE OF CONTENTS LIST OF TABLES………………………………………………………………... v LIST OF FIGURES………………………………………………………………. vi INTRODUCTION………………………………………………………………… The measurement of interpersonal behavior…………………………….… Interpersonal models of behavioral influence……………………………... Approaches to measuring interpersonal influence………………………… Anxiety and interpersonal dominance……………………………………... Anxiety and interpersonal warmth………………………………………… Present Study………………………………………………………………. Hypotheses……………………………………………………………….... 1 1 2 4 7 9 11 11 METHODS………………………………………………………………………... Participants………………………………………………………………… Procedure…………………………………………………………………... Measures…………………………………………………………………… Interaction Procedure……………………………………………………… Psychophysiological Data Recording, Reduction, and Analysis………….. Joystick Coding……………………………………………………………. Analysis……………………………………………………………………. 14 14 14 15 17 18 19 20 RESULTS…………………………………………………………………………. Questionnaire Internal Consistencies and Descriptive Statistics………….. Observer-coded Interpersonal Behavior Internal Consistencies and Descriptive Statistics………………………………………………………. Hypothesis 1: Self-reported interpersonal style will relate to observercoded interpersonal behavior in a laboratory paradigm…………………… Hypothesis 2: Self-reported traits will relate to self-reported interpersonal sensitivities such that traits will correlate with their opposite sensitivity…. Hypothesis 3: Observer-coded cold behaviors will relate to anxiety……… Across-dyad warmth and liking the confederate…………………... Across-dyad warmth and negative affect………………………….. Across-dyad warmth and skin conductance……………………….. Within-dyad warmth and skin conductance……………………….. Hypothesis 4: Deviation from complementarity between observer-coded participant and confederate warm behaviors will provoke participant anxiety after accounting for main effects of participant and confederate warmth……………………………………………………………………... Within-dyad warmth complementarity and skin conductance…….. Across-dyad warmth complementarity and skin conductance…….. Across-dyad warmth complementarity and liking the confederate... 25 25 iii 26 29 31 32 33 34 35 36 38 38 39 40 Across-dyad warmth complementarity and negative affect……….. Hypothesis 5: Deviations from complementarity between observer-coded participant and confederate dominant behavior will provoke participant anxiety……………………………………………………………………... Within-dyad dominance complementarity and skin conductance…. Across-dyad dominance complementarity and skin conductance... Across-dyad dominance complementarity and liking the confederate………………………………………………………… Across-dyad dominance complementarity and negative affect……. Hypothesis 6: The effect of observer-coded confederate behavior on skin conductance will correspond to self-reported interpersonal sensitivities….. Gender Effects on Interpersonal Processes……………………………… 41 DISCUSSION……………………………………………………………………... Alternative Tests of Interpersonal Complementarity……………………… The Effect of the Task on Tests of Interpersonal Complementarity. The Effect of the Operationalization of Anxiety on Tests of Interpersonal Complementarity……………………………………. The Effect of the Interactants on Tests of Interpersonal Complementarity…………………………………………………... The Effect of the Nature of Association on Tests of Interpersonal Complementarity…………………………………………………... Potential Differences between Communal and Agentic Dimensions……... Moderating effects of Interpersonal Sensitivities………………………….. Conclusion…………………………………………………………………. 49 50 51 APPENDICES……………………………………………………………………... Appendix A: Tables……………………………………………………….. Appendix B: Figures………………………………………………………. Appendix C: Materials ….…………..…………………………………….. 61 62 87 85 REFERENCES…………………………………………………………………….. 91 iv 41 41 42 43 44 45 45 53 54 55 57 58 58 LIST OF TABLES Table 1: Descriptive Statistics and Cronbach’s Alphas…………………………... 63 Table 2: Bivariate Correlations…………………………………………………… 71 Table 3: Summary of Hypothesis Tests…………………………………………... 69 Table 4: Regression Summaries…………………………………………………... 75 v LIST OF FIGURES Figure 1: Interpersonal Circle…………………………………………………….. vi 84 INTRODUCTION From an interpersonal perspective (Sullivan, 1953; Leary, 1957), personality is defined as the “relatively enduring pattern of recurrent interpersonal situations which characterize a human life” (Sullivan, 1953, p110). The functional purpose of one’s approach to interpersonal situations is “to reduce anxiety, ward off disapproval, and maintain self-esteem” (Leary, 1957, p. 119). Behavioral variability in these interpersonal situations can be structured by two dimensions, communion (warmth to coldness) and agency (dominance to submissiveness) (Wiggins, 2003). Communion is defined by the degree to which an individual acts in a manner that is friendly, warm, interested, affiliative, and close with others, versus cold, unfriendly, distant, detached, and private. It is the “condition of being part of a larger social or spiritual entity, and is manifested in strivings for intimacy, union, and solidarity” (Pincus, Lukowitsky, & Wright, 2010; p528). Agency reflects the degree to which an individual exhibits behavior that is controlling, leading, advising, dominant, and ambitious versus submissive, unassertive, docile, and dependent. It is the “condition of being a differentiated individual and it is manifest in striving for mastery and power which enhance and protect that differentiation” (Wiggins, 1991). The measurement of interpersonal behavior Interpersonal assessment instruments are available to measure warmth and dominance across a number of ‘surfaces’ that vary in content and specificity. Surfaces for which interpersonal measures have been developed include efficacy (i.e., confidence in the ability to perform a behavior; Locke & Sadler, 2007), values (i.e., hypothetical preferences; Locke, 2000), perceptions of others’ behavior (Kiesler, Schmidt, & Wagner, 1997), problems (i.e., behaviors an individual finds hard to do or does too much; Alden, Wiggins, & Pincus, 1990; Horowitz, Alden, Wiggins, & Pincus, 2000), and sensitivities (i.e., behaviors which an individual finds aversive; 1 Hopwood et al., 2011). Methods have also been developed to measure both specific behaviors (Strong & Hills, 1986; Sadler, Ethier, Gunn, Duong, & Woody, 2009; Markey, Lowmaster, & Eichler, 2010) and aggregate traits (Wiggins, Trapnell, & Phillips, 1988; Markey & Markey, 2009; Moskowitz, 1994; Fournier, Moskowitz, & Zuroff, 2008). The existence of interpersonal measurement surfaces for each of these types of interpersonal behavior permits direct comparisons of interpersonal functioning across different levels using common metrics. These surfaces take the form of an interpersonal circumplex (IPC), or circular arrangement of interpersonal behaviors involving blends of communion and agency. Unlike simplex models, in a circumplex characteristics around the circumference represent a psychologically meaningful continuum, and the distance between behaviors in circular space is quantitatively interpretable (Gifford & O’Connor, 1987; Gurtman, 1992; Adams & Tracey, 2004). IPC measures are commonly divided into eight octants, four defining the dimensional nodes of communion and agency and four defining the possible combinations of the dimensions (See Fig. 1). Interpersonal models of behavioral influence Interpersonal theorists (Kiesler, 1996; Horowitz et al., 2006; Sadler & Woody, 2003) have followed Lewin (1936) in postulating that behaviors are influenced by a combination of traits and environmental situations. This general hypothesis is broadly supported by evidence that interpersonal traits predict behavior (Buss & Craik, 1983; Back, Schmukle, & Egloff, 2009; Paunonen & Ashton, 2001; Fleeson & Gallagher, 2009), traits can be reliably inferred from interpersonal behavior (Riemann & Angleitner, 1993; Ambady, Hallahan, & Rosenthal, 1995; Borkenau & Liebler, 1992; Borkenau, Mauer, Riemann, Spinath, & Angleitner, 2004; Oltmanns, Friedman, Fiedler, & Turkheimer, 2004; Friedman, Oltmanns, & Turkheimer, 2007) and certain 2 situations can have more or less limiting effects on interpersonal behaviors (Mischel, 1977; Snyder & Ickes, 1985; Moskowitz, Ringo Ho, Turcotte-Tremblay, 2007). Interpersonal theory also specifies the nature of the influences on interpersonal functioning by postulating that an individual’s behavior will arise from a combination of interpersonal traits (i.e., stable individual differences in interpersonal style that can be defined by a particular blend of agency and communion) and predictable reactions to others’ behaviors (i.e. population normative tendencies to respond to blends of others’ agentic or communal behavior in systematic ways) (Carson, 1969; Kiesler, 1996; Sadler, et al., 2009). Specifically, behaviors that are friendly prompt (“invite, pull, elicit, draw, entice, or evoke,” Kiesler, 1983) friendly behaviors in return, promoting security, while unfriendly behaviors prompt unfriendly responses. Conversely, dominant guiding behaviors prompt submissive behaviors, while submissive behaviors are associated with reciprocal dominant behaviors, promoting esteem. This pattern of behavioral interaction in which interactants’ warmth coheres and interactants’ dominance reciprocates is known as complementarity (Carson, 1969). Complementarity provides a framework for predictions about the interpersonal processes that reduce anxiety and promote security and self-esteem (Sullivan, 1953; Leary, 1957; Wiggins, 2003; Pincus, Lukowitsky, & Wright, 2010). Empirical evidence for complementarity and its theoretical implications is reasonably strong. Complementarity manifests across a variety of different relational roles (Tracey, Ryan & Jaschik-Herman, 2001; Markey, Funder, & Ozer, 2003; Strong et al., 1988; Tracey, 1994) and typically increases over the course of a relationship (Markey & Kurtz, 2006). Complementarity is associated with interpersonal closeness, (Yaughn & Nowicki, 1999) cohesion, (Ansell, Kurtz, & Markey, 2008; O’Connor & Dyce, 1997; Shechtman & Horowitz, 2006; Tiedens & Fragale, 3 2003) and satisfaction (Dryer & Horowitz, 1997; Locke & Sadler, 2007). Married couples exhibit a higher degree of complementarity than divorced or separated couples (Tracey, Ryan, & Jaschik-Herman, 2001), and specific patterns of complementarity are associated with effective psychotherapy (Tracey, Sherry, & Albright, 1999; Tracey, 2004; Henry, Schacht, & Strupp, 1986; Tasca & McMullen, 1992). Within interactions, complementarity also relates to psychophysiological indicators of anxiety, as would be predicted by the general hypothesis that one of the primary functions of smooth interpersonal relations is anxiety regulation. In this context, anxiety refers more to global distress or negative arousal rather than the contemporary reference to fear or worry. Dyadic dissimilarity in warm behaviors is associated with increases in blood pressure and heart rate and attempts to restore complementarity nonverbally (Smith & Ruiz, 2007). Similarly, individuals experience greater cardiovascular reactivity when disclosing problems to others with unmitigated warmth (Fritz, Nagurney, & Helgeson, 2003). There is less research on the effects of complementarity on the dominance axis relative to the affiliation axis (Smith, Gallo, & Ruiz, 2003). Ansell and colleagues (2012) employed the joystick technique to code behavior in a conflictual interaction between parents and adolescents and found that relative increases in adolescent cortisol were associated with deviations from dominance complementarity. Similarly, dominant men experience blood pressure reactivity when interacting with dominant partners (Newton, Watters, Philhower, & Weigel, 2005). Approaches to measuring interpersonal influence Despite this general evidence, how to optimally study complementarity remains unclear. Bluhm, Widiger, and Miele (1990) rated participants’ interpersonal behavioral style in interactions with confederates who performed roles based on the four extremes of communion 4 and agency and found that participant warmth was more related to confederate behavior whereas participant dominance was more related to self-reported interpersonal dominance. Strong and colleagues (1988) observed interactions between participants and confederates who performed scripted roles emphasizing one of eight possible interpersonal styles corresponding to the octants of the IPC. When they recorded each interpersonally meaningful behavior and summed them across interactions, they found that participants displayed behaviors predicted by complementarity. Sadler and Woody (2003) investigated the combined effect of interpersonal traits and situations (defined by the interpersonal behavior of a dyadic partner) on behavior in a laboratory interaction using multiple informants. They estimated interpersonal traits with self- and informant-reports and interpersonal behavior using self-, partner- and observer-reports in a structural equation model. They then predicted the behavior of each interactant with their own traits and the other person’s behavior for both dominance and warmth. Trait dominance predicted dominant behavior (path coefficient = .48) and trait warmth predicted warm behavior (path coefficient = .40), showing the relevance of traits for behavioral predictions. Consistent with complementarity and the importance of situations, behavior was also predicted by the other individual for warm (path coefficient = .39) and dominant behavior (path coefficient = -.29). In each of these studies, as with similar examinations of complementarity (e.g., Strong, et al., 1988; Bluhm, Widiger, & Miele, 1990; Markey, Funder, & Ozer, 2003), raters coded global ratings across the interaction rather than the pattern of behaviors within the interaction. However, interpersonal theory postulates that complementarity influences individuals at the momentary level – behaviors bid for behaviors. Moreover, Tracey (2004) demonstrated that the effect of complementarity is stronger to the degree that assessments focus on specific behaviors in an 5 interaction as opposed to aggregated interactions or stable traits. This finding sheds light on inconsistency in the examination of trait complementarity (Orford, 1986), which is more likely to be observed when ratings of interpersonal style are global and distal. Sadler, Ethier, Gunn, Duong, and Woody (2009) developed a method of coding momentary interpersonal behavior in dyadic interactions in real time to more effectively model complementarity as it has been theorized in the interpersonal literature (e.g., Kiesler, 1996) and demonstrated in the empirical literature (e.g., Tracey, 2004). In this method trained coders watch a dyadic interaction and concurrently manipulate a computer joystick along dimensions which correspond to the dimensions of the IPC in order to sample warm and dominant behaviors, one interactant at a time, every 0.5 seconds. They applied this method to the videotaped interactions from Sadler and Woody (2003) and found that aggregated behavioral warmth was associated with observational ratings of warmth (r = .65) and aggregated behavioral dominance was associated with observational ratings of dominance (r = .81). They tested complementarity within each interaction by correlating each individual’s momentary warmth and dominance time series with the corresponding time series of the other individual. When averaging across all dyads, results were consistent with complementarity for both warmth (mean r = .51) and dominance (mean r = -.43). Markey, Lowmaster, and Eichler (2010) applied a similar method to interactions between unacquainted individuals who performed cooperative tasks. These dyads displayed complementarity with regards to warmth (mean warmth r = .36) and dominance (mean dominance r = -.32). Furthermore, dyads that deviated more from warmth complementarity also tended to like each other less and perform worse on the cooperative tasks. Thus interpersonal complementarity manifests across traits, relations, interactions, and momentary behaviors, 6 prompting the question: what mechanism elicits and maintains complementarity? If, as Sullivan (1953) suggests, interpersonal processes are meant to reduce anxiety by satisfying needs for security and self-esteem, then deviations from complementarity would be expected to increase anxiety. Anxiety and interpersonal dominance According to interpersonal theory, deviations from dominance complementarity – when both individuals are dominant or both are submissive – should be associated with anxiety. Given that interpersonal traits predict behavior and that certain types of behavioral interactions are associated with anxiety, it stands to reason that interpersonal behaviors that an individual finds irritating or anxiety provoking might have a trait-like form. Hopwood and colleagues (2011) investigated this hypothesis by examining individual differences in self-reported sensitivities to different interpersonal behaviors. Drawing on the principle of complementarity, Hopwood et al. hypothesized that individuals higher in self-reported warmth would report greater sensitivity to cold behaviors and individuals higher in self-reported dominance would report greater sensitivity to dominant behavior. Although findings supported the hypothesis regarding the warmth dimension, trait dominance correlated with sensitivities to submissiveness, suggesting overall that people are most irritated by their opposites, not their a-complement, as would be predicted by interpersonal theory (Kiesler, 1996). How can the finding that people tend to report being irritated by those who are complementary to themselves on agency be consolidated with evidence reviewed above indicating that deviations from complementarity are associated with anxiety? One issue is that complementarity theoretically occurs at a more contextually specific level than broad aggregate attitudes about others (e.g., Tracey, 2004), as discussed above. This may explain, for instance, 7 why assortative mating effects on dominance have tended to be weak and mixed, with some studies showing positive spousal correlations on dominance (Humbad, Donnellan, Iacono, McGue, Burt, 2010; Barleds, 2005; Vandenberg, 1972; although see also McCrae et al., 2008). It is possible that people, in the aggregate, would like others to be like themselves, whereas in specific interactions with or responses to them they would prefer others to be complementarity. For example, a dominant individual could prefer others to be dominant in general, but to display complementary behaviors in specific interactions. If this were the case, then aggregate measures of traits and sensitivities may show different patterns of associations than momentary measures of behaviors and anxiety. Momentary assessments would indicate an association between acomplementarity on dominance and anxiety at the level of specific behavioral interchanges, even for people who report that they are most irritated, in general, by their interpersonal opposite. Another potential explanation involves how anxiety is measured. Participant self-reports of interpersonal sensitivities may be affected by biases, such as potential efforts to maintain or justify one’s identity. For instance, one may be less likely to endorse behaviors that reflect one’s own personality as aversive. In this case, whereas individuals would report being most irritated by their opposite, fewer verbal indicators would indicate that a-complementarity on dominance is associated with anxiety. This finding would be consistent with previous research by Assor, Aronoff, and Messe (1986), who recorded skin conductance in subjects watching interactions with someone whom they expected to meet. They found that subjects high in trait dominance experienced greater reactivity when the individual they expected to meet displayed dominant behavior, whereas more submissive subjects experienced greater reactivity when the individual displayed submissive behavior. Moreover, more dominant participants evaluated the person they observed more highly to the extent that they were dominant in the interaction, similar to the 8 result reported by Hopwood et al. Alternatively, the phrasing in the ISC asks how much a given behavior “bothers” the respondent, rather than specifically referring to anxiety or other negative affective experiences. Finally, it is possible that individuals prefer interacting with others who are similarly dominant or submissive, as suggested by the Hopwood et al. findings. This would be consistent with the similarity-attraction hypothesis (Byrne, 1971), but would contrast with an interpersonal perspective and research reviewed above on the relation between reciprocal dominance and anxiety. In this case, both self-reported and other methods would indicate that interpersonal complementarity is associated with anxiety. Therefore, one purpose of this study is to distinguish and relate different anxiety responses and sensitivities to dominance a-complementarity. Anxiety and interpersonal warmth Another purpose of this study is to investigate the nature of the association between anxiety, individuals’ warm behavior, and the complementarity of that warm behavior. Interpersonal theory postulates that similarity in warmth or coldness during an interaction is associated with decreased anxiety. However, research also indicates that interpersonal coldness in general increases cardiovascular reactivity (Vella & Friedman, 2009; Smith & Allred, 1989; Stroud, Tanofsky-Kraff, Wilfey, & Salovey, 2000; Hardy & Smith, 1988; Newton & Bane, 2000), cortisol (Pope & Smith, 1991), and skin conductance (Gormly, 1974; Gallo, Smith, & Kircher, 2000). Schade et al. (2012) analyzed data with Sadler, et al.’s joystick coding method in a pilot study of 9 dyads to investigate the relation between momentary behavior and skin conductance. Consistent with previous research, complementarity between interactants arose with regards to warmth (mean r = .50; mean weighted coherence = .47) and dominance (mean r = -.45; mean weighted coherence = .44). In this pilot study, elevated skin conductance was 9 associated with others’ momentary cold behaviors (mean r = -.13, mean weighted coherence = .40), indicating that individuals are generally more bothered by cold behaviors than by warm behaviors. Furthermore, affiliative behavior is typically rated as more desirable than unaffiliative behavior (Hill & Safran, 1994) and satisfaction in an interaction is associated with warm values (Locke & Sadler, 2007). If warm behavior is more desirable, more satisfying, and less anxiety provoking, why do some individuals nonetheless display cold behaviors? One possible explanation is that individuals who adopt cold, unaffiliative interpersonal strategies do not necessarily do so because they desire their dyadic partners to be interpersonally distant, but because they expect them to be so (Locke, 2005; Kivlighan, Marsh-Angelone, & Angelone, 1994; Kivlighan & Angelone, 1992; Dodge & Somberg, 1987; Dodge & Crick, 1990; Wood, Harms, & Vazire, 2010; Krueger & Clement, 1994; Graziano, Bruce, Sheese, & Tobin, 2007). In interpersonal terms, they expect their bids for affiliation to be rejected, violating security needs, so instead choose not to make those bids as a way of minimizing anxiety. If cold behavior can be attributed to false expectations about others’ cold behavior, then complementary cold behavior would not necessarily serve to optimally alleviate anxiety. Thus, while interpersonal complementarity theory suggests that a person behaving in a cold manner would be put at ease by another person behaving coldly towards him, other evidence suggests that cold behavior may be generally aversive, meaning a person behaving coldly would still prefer warm behavior. Given that individuals generally prefer warm behaviors, it may have been previously impossible to dissociate the two effects. That is, research investigating the main effect of behavioral warmth (e.g., Gallo, Smith, & Kircher, 2000) has neglected individual differences, while research investigating individual differences in response to behavioral warmth (e.g., Smith & Ruiz, 2007) 10 has neglected the main effect of behavioral warmth. Therefore, a goal of this study is to determine whether cold behaviors, dissimilarity in warm behaviors, or both provoke anxiety. Present Study The overall purpose of the present study is to examine how deviations from interpersonal complementarity are associated with anxiety. Taking into account that complementarity is conceived in interpersonal theory and manifests most strongly at the behavioral level (Tracey, 2004), I will employ Sadler et al.’s joystick method (2009) to measure momentary warmth and dominance and relate these behaviors to skin conductance measured at the same time scale. In accordance with findings from the pilot study (described below), observed interactions will be between one participant and one confederate, who will more readily vary interpersonal behavior when meeting an unacquainted participant. Consistent with interpersonal theory, I expect objective indicators of anxiety to relate to deviations from complementarity in the interaction. Moreover, as cold behaviors are considered a barrier to communion, I expect cold behaviors to relate to anxious physiological reactions. Finally, I expect physiological responses to interpersonal behavior to be moderated by interpersonal sensitivities. Hypotheses Hypothesis 1: Self-reported interpersonal style will relate to observer-coded interpersonal behavior in a laboratory paradigm, replicating Sadler et al., 2009: A: Observer-coded participant warmth during the interaction will correlate positively with self-reported trait warmth and self-reported warm behavior during the interaction across dyads. 11 B: Observer-coded participant dominant behavior during the interaction will correlate positively with self-reported trait dominance and self-reported dominant behavior during the interaction. Hypothesis 2: Self-reported traits will relate to self-reported interpersonal sensitivities such that traits will correlate with their opposite sensitivity, replicating Hopwood, et al., 2011: A: Participant self-reported trait warmth will negatively correlate with self-reported sensitivity to warmth. B: Participant self-reported trait dominance will negatively correlate with self-reported sensitivity to dominance. Hypothesis 3: Observer-coded cold behaviors will relate to anxiety: A: Mean observer-coded participant behavioral warmth in an interaction will correlate negatively with self-reported anxiety and positively with measures of liking across dyads. B: Mean observer-coded confederate behavioral warmth in an interaction will correlate negatively with self-reported anxiety and positively with measures of liking across dyads. C: Observer-coded participant warm behaviors will relate to participant peripheral psychophysiological indicators of anxiety within interactions. Hypothesis 4: Deviation from complementarity between observer-coded participant and confederate warm behaviors will provoke participant anxiety after accounting for main effects of participant and confederate warmth: A: Deviations from complementarity between observer-coded participant and confederate warmth within interactions will relate to participant peripheral physiological indicators of anxiety within dyads. 12 B: Mean deviations from complementarity between observer-coded participant and confederate warmth during an interaction will correlate positively with self-reported anxiety and negatively with measures of liking across dyads. Hypothesis 5: Deviations from complementarity between observer-coded participant and confederate dominant behavior will provoke participant anxiety: A: Deviations from complementarity between observer-coded participant and confederate dominance will relate to peripheral physiological indicators of anxiety within dyads. B: Mean deviations from complementarity in observer-coded dominant behaviors during an interaction will correlate negatively with self-reported anxiety and positively with measures of liking across dyads. C: Mean dominance complementarity will correlate negatively with skin conductance and measures of liking across dyads. Hypothesis 6: The effect of observer-coded confederate behavior on skin conductance will correspond to self-reported interpersonal sensitivities: A: Self-reported sensitivity to coldness will predict a stronger positive relationship between observer-coded confederate cold behavior and peripheral physiological indicators of anxiety. B: Self-report sensitivity to dominance will predict a stronger positive relationship between observer-coded confederate dominant behavior and peripheral physiological indicators of anxiety. 13 METHODS Participants Seventy two (41 female) undergraduates participated in the present study for class credit or volunteered. They primarily described themselves as Caucasian or white (N = 60) and ranged in age from 18 to 23 (M = 19.77, SD = 1.52). Due to equipment failure, observer-coded interpersonal behaviors were available for only 63 participants (38 female). Procedure Participants were recruited through the MSU HPR System and were compensated with course credit. After arriving in the lab, the participants were briefed that they would be participating in a study on personality, behavior, and physiology which would include both questionnaires and an interaction task with peripheral psychophysiology recording. Informed consent was be obtained, wherein the participant was be led to believe the interaction would be with another participant and that a random decision determined which participant would be measured with physiological recordings. Participants next completed the pre-interaction questionnaires measuring interpersonal traits, behaviors, sensitivities, problems, and affects. Next, electrodes were affixed to the distal phalanges of the index and middle fingers. A first skin conductance baseline of 2 minutes was acquired prior to meeting the confederate. A second baseline of 1 minute was acquired after the confederate entered the room while the instructions for the interaction task were being given. These baselines will serve to control for individual differences in net skin conductance, permitting analysis of changes in skin conductance. After the interaction, the participant completed additional questionnaires assessing perception of the interaction, affect, and measures of liking. Finally the participants were debriefed, informed of the other interactant’s confederate status, and thanked for his or her participation. 14 Measures Means, standard deviations, and alphas for each of the scales in this study are presented in Table 1. The Social Behavior Inventory (SBI; Moskowitz, 1994) is a 46-item measure of interpersonal behavior with dimensions corresponding to warmth and dominance. Respondents are asked to report on interpersonal behavior on a 6-point scale (from “never” to “almost always”). As in previous studies (e.g. Sadler & Woody, 2003; Sadler, et al., 2009), the SBI will be used to measure interpersonal traits, the participant’s rating of the confederate’s behavior during the interaction, and the participant’s rating of his or her own behavior during the interaction. The SBI has four subscales: dominance, submissiveness, warmth, and coldness. A warmth score was calculated by subtracting coldness from warmth and dominance score was calculated by subtracting submissiveness from dominance (Moskowitz, 1994). The International Personality Item Pool - Interpersonal Circumplex (IPIP-IPC; Markey & Markey, 2009) is a 32-item self-report measure of interpersonal traits organized around the interpersonal circumplex. Participants rate the degree to which each statement is accurate of them on a 5-point scale from “Very Inaccurate” to “Very Accurate.” The IPIP-IPC provides scores on each of the interpersonal circumplex’s eight octants, from which warmth and dominance scores can be computed via geometry (Wiggins et al., 1989; Gurtman, 1996). I included this as an additional measure of interpersonal traits because the IPIP-IPC, as opposed to the SBI, produces octant scores, resulting in more reliable estimates of warmth and dominance vectors. The Inventory of Interpersonal Problems - Short Circumplex Form (IIP-SC; Soldz et al., 1995; Hopwood, Pincus, DeMoor, & Koonce, 2008) is a 32-item self-report measure of 15 interpersonal problems organized around the interpersonal circumplex model of social behavior and personality (Alden, Wiggins, & Pincus, 1990; Horowitz, Alden, Wiggins, & Pincus, 2000). It assesses interpersonal behaviors that participants report as “hard to do” or “do too much.” This measure produces scores for each of the interpersonal circumplex’s eight octants. Octant scale scores were converted into vectors representing warm interpersonal problems (M = -0.21, SD = .16), and dominant interpersonal problems (M = .06, SD = .15). A global index of interpersonal adjustment, elevation, was calculated by taking the mean of standardized octant scores (M = 1.01, SD = .14). The Interpersonal Sensitivities Circumplex (ISC; Hopwood, Ansell, Pincus, Wright, Lukowitsky, & Roche, 2011) is a 64-item self-report measure used to assess the degree to which the participant is bothered by the behaviors of others on an 8-point scale from “Not at all, never bothers me” to “Extremely, always bothers me.” The ISC provides measures of the respondent’s sensitivity to each of the eight octants of the interpersonal circumplex. Cronbach’s alpha for the ISC octant scales ranged from .73 to .86. These octant scale scores were converted into a vector of sensitivity to warmth (M = .09, SD = .54) and a vector of sensitivity to dominance (M = .03, SD = .62). Standardized scores were averaged to reflect the general degree to which the behaviors of others bother the participant (M = -.40, SD = .68). The Positive and Negative Affect Schedule (PANAS; Watson et al., 1988) consists of 10item Positive Affect and Negative Affect scales. Participants are asked to rate the degree to which they have experienced a particular emotion over the past week on a 5-point scale (from “very slightly or not at all” to “very much”). I chose to employ the PANAS, as opposed to a measure of state anxiety, in order to index the more general negative affect described by 16 interpersonal theory. The PANAS was given both before and after the interaction task in order to measure the effect of the interaction on affective experiences during the task. Finally, participants’ ratings of liking the confederate, finding the confederate attractive, and enjoying the task were assessed after the interaction. Participants were instructed to report how much they liked the confederate, how much they would want to see the confederate again, how much they would want to be friends with the confederate, how attractive they found the confederate, and how enjoyable they found the interaction task on a 9-point likert scale. The mean of the first three items was retained as an index of how much the participant liked the confederate (alpha = .92). Interaction Procedure Participants took part in a dyadic interaction with a confederate. This interaction lasted 15 minutes and was videotaped with the participant’s consent. In a pilot study, Schade et al. (2012) manipulated the structure of the interaction (i.e., whether the interaction was organized around discussion prompts or not) and nature of interaction dyad (i.e., whether the interaction included two participants or a confederate). Pilot data suggested that unacquainted strangers typically did not display a wide diversity of behaviors along the affiliative dimension in open-ended tasks, seemingly preferring a more cordial interaction, whereas a confederate was able to manifest more variability when instructed to do so. Therefore, I employed a confederate for the interaction task. The confederate was a male university undergraduate with a background in the performing arts. Previous research (Sadler, et al, 2009; Sadler & Woody, 2003) indicates limited gender effects for complementarity, so the confederate was not selected on gender. However, I sampled both genders to test gender effects. The confederate was trained in joystick coding (see below) and was required to meet the same inter-rater agreement standard of .40 with other well-trained 17 judges in order to appreciate the full range and meaning of interpersonal dimensions. He was instructed to vary his affiliative and dominant behavior throughout the interaction. To aid in this task, a list of interpersonal behaviors varying across these dimensions was developed (see Appendix). The confederate was instructed to perform each of these behaviors at least twice over the course of the interaction. Furthermore, his observer-rated behaviors were checked to see whether warmth and dominance were appropriately orthogonal and when they varied appropriately. When appropriate, I delivered feedback about adherence to the instructions. Members of the dyad were prompted to learn as much as possible about one another and were provided with discussion prompts (see instructions in the Appendix). These prompts were taken from Aron’s sharing game (Aron, Melinat, Aron, Vallone, & Bator, 1997), a series of questions designed in order to induce closeness and familiarity in an experimental setting. Example prompt questions include “What would constitute a ‘perfect’ day for you?” and “If you could wake up tomorrow having gained any one quality or ability, what would it be?” (see all prompts in the Appendix). Psychophysiological Data Recording, Reduction, and Analysis Continuous skin conductance (SC) activity was recorded by 2 Ag-AgCl electrodes (Biosemi, Amsterdam, The Netherlands) affixed to the distal phalanges of the index and middle fingers. During data acquisition, the Common Mode Sense active electrode and Driven Right Leg passive electrode form the ground, as per Biosemi’s design specifications. All bioelectric signals were digitized at 512 Hz using ActiView software (BioSemi). Offline Analyses were performed using LedaLab Software, a Matlab-based program for the analysis of skin conductance (MathWorks, Massachusetts, United States). SC data were down-sampled to 50 Hz, 18 utilizing spline interpolation (third-degree polynomial), for storage and analysis. Finally, data collected during the interaction task was segmented into contiguous epochs lasting 500 ms each. Joystick Coding Seven trained judges utilized a method developed by Sadler (see Sadler, et al., 2009) to provide a real time assessment of the participant’s and confederate’s affiliative and dominant behaviors during the interactions. Judges watched the videotaped interactions and oriented the joystick in accordance with their perception of the participant’s verbal and nonverbal interpersonal behavior. The orientation of the joystick maps on to Cartesian coordinates with axes reflecting the interpersonal dimensions such that movements in each direction reflect the type of interpersonal behavior (e.g., right for affiliation, up for dominance) and the distance from the center reflects the extremity of that behavior. As in Sadler et al., data was recorded every 0.5 seconds, yielding a total of 1800 observations per person per interaction. Each observation produced both an affiliation score and a dominance score ranging from -1000 to 1000. The zeroes of each score range represent neither warm nor cold and neither dominant nor submissive, respectively. Before coding, judges were trained to be familiar with positions of the interpersonal behaviors using the training scripts developed by Sadler et al. In addition, judges were provided a list of adjectives and required to locate the corresponding joystick position. Finally, judges practiced using the joystick on a series of dyadic interaction videos. To ensure that a judge’s ratings of one individual would not be biased by ratings of the other individual, judges were prohibited from rating the same interaction twice in a row. I then truncated the first 5 seconds from each data series to ensure judges have adequate time to move their joysticks away from the neutral position (Sadler, et al., 2009). Inter-rater reliability for warmth was examined at the 19 momentary level by correlating different judges’ ratings of warm behavior for these practice dyadic interactions. Inter-rater reliability for dominance was examined in a similar manner. Following the instructions of Sadler et al., I used an inter-rater correlation of .40 as a benchmark for acceptable judge training. Seven judges met this benchmark and coded the interaction videos. Analyses The first set of study hypotheses involved the association between self-reported interpersonal style and joystick coded interpersonal behavior. In order to test these hypotheses, I calculated observer-coded warmth and dominance scores for each participant by averaging across each coder. I correlated, across interactions (A) average observer-coded participant warmth and self-reported trait warmth; and (B) average observer-coded participant dominance and self-reported trait dominance. The second set of hypotheses involved the relation between self-reported interpersonal traits and self-reported sensitivities. I tested this hypothesis by correlating, across interactions, (A) self-reported trait warmth to self-reported sensitivity to warmth; and (B) self-reported trait dominance to self-reported sensitivity to dominance. The third set of hypotheses involved the association between confederate and participant interpersonal coldness and participant anxiety. In order to test this hypothesis, I calculated a mean observer-coded warmth score for each participant by averaging across each coder. I then correlated the average observer-coded participant warmth with self-reported anxiety across interactions; as well as average observer-coded participant warmth with measures of liking. Selfreported anxiety was indexed as the residual score from regressing pre-interaction negative negative affect on post-interaction negative affect. The PANAS NA score was used instead of the single anxiety item to enhance reliability and to account for the possibility that the subjective 20 experience of negative affect associated with interpersonal patterns might differ slightly across participants. I also calculated a mean observer-coded warmth score for the confederate in each interaction by averaging across each coder. I then correlated the average observer-coded confederate warmth with self-reported anxiety; as well as average observer-coded confederate warmth with measures of liking across interactions. I also tested the relation between momentary interpersonal coldness and anxiety within interactions. I employed cross-correlation because it can inform on delays in reaction to other’s interpersonal behavior. In a cross-correlation, 2n correlations are performed by correlating time series X with time series Y at t+n and t-n intervals, identifying which time lag n best adheres to the assumptions of complementarity. Although cross-correlations provide a useful initial indicator of any relation between time series, such an approach can lead to spurious correlations due to potentially independent trends (Cappella, 1996; Raffalovich, 1994; Warner, 1998). Correlating time series with independent trends may produce spurious correlations (such as the association between Venetian sea levels and British bread prices, which each trend upward over time; Sober, 2001). Therefore, I repeated the cross-correlation using the residuals that are left after detrending. Spurious correlations can also result from autocorrelation, which occurs when time series data follows independent sinusoidal trends. In other words, just as two phenomena with linear trends (e.g., Venetian sea levels and British bread prices) manifest a spurious correlation, two phenomena with cyclical trends (e.g., daylight hours and gift giving) can manifest spurious associations as well. Previous research suggests that individual interpersonal behavioral data (Wagner, Kiesler, & Schmidt, 1995) as well as skin conductance data (Lackner et al., 2010) take sinusoidal shapes and that different dyads’ behavioral waves can differ in wavelength (Sadler, et al., 2009). For example, different dyads might exhibit differing rates of 21 dominance turn taking, which might result in different dominance wavelengths. This sinusoidal trend would raise the concern that behaviors and skin conductance reactivity fall on sine waves and that the degree of similarity in their wavelengths is merely coincidental. To test against this possibility, I compared participant behavior in an interaction to the behavior of the confederate in different interactions. Behavior ratings for a participant that correlate more strongly with confederate behavior in other interactions (a pseudo-dyad) would cast doubt on the correlation between behaviors within that participant’s interaction. Notably however, previous tests of spurious correlation in joystick data have been negative across several samples (Sadler et al., 2009). I also performed spectral analysis to find the best-fit wave forms for momentary behavior and skin conductance after detrending. The first step in spectral analysis involved determining which frequencies or wavelengths best capture the form of the data. This was done by using ordinary least squares regression on the functional form X = b1 + b2*cos(wt) + b3*sin(wt) + e where b1 is the mean after detrending and w is equal to 2*pi/period. This regression will be performed k/2 times, for k/2 different periods, where k is the length of the full time series. These analyses produce a periodogram which indicates which frequencies account for the greatest amount of variance in the data. Once the wave which best describes a time series is identified, linear regression can be used to determine the extent to which that wave accounts for variance in another time series wave. This results in an index of variance known as coherence (Gottman, 1979). The degree to which the cycles are in phase (i.e., the degree to which the peaks in one cycle match the peaks of another) will also be calculated. Cross-correlational and spectral analysis will be conducted with the data from each dyad, resulting in 60 sets of results. For each 22 type of test (correlation before detrending, correlation after detrending, and phase), I then performed a binomial probability test to determine whether significantly more interactions exhibit the hypothesized relationship than would be expected by chance. This will produce a p value and significance test for each type of test. I performed a one-sample t-test to determine whether the mean correlations differed significantly from zero. Fisher-transformation was performed when comparing within-dyad coefficients across dyads. I used this method to test the association between confederate warmth and participant skin conductance, participant warmth and participant skin conductance, and participant warmth and confederate warmth. I also used this method to test the effect of deviations from complementarity on participant skin conductance. In order to do so, I calculated indices of momentary deviations from complementarity. The index of momentary deviation from warmth complementarity is equal to the absolute value of the difference between participant momentary warmth and confederate momentary warmth. The index of momentary deviation from dominance complementarity is equal to the absolute value of the sum of participant momentary dominance and confederate momentary dominance. I expected each of these indexes to be positively associated with skin conductance and I expected these associations to maintain after controlling for main effects of participant and confederate interpersonal behavior in hierarchical linear regression models. Furthermore, I expected the mean deviation from complementarity in an interaction to predict negative changes in affect and liking the interaction less. Finally, I also expected that the effect of observer-coded confederate behavior on skin conductance would be moderated by self-reported interpersonal sensitivities. For example, I expected cold behavior to increase skin conductance and I expected this effect to be greater for individuals who report being more sensitive to cold behavior. Each of the methods of assessing 23 the relation between observer-coded confederate cold behavior and peripheral physiological indicators of anxiety produced a coefficient for each interaction (e.g., cross-correlation before and after detrending, spectral phase, coherence). To test these final hypotheses, I correlated each of these coefficients with self-reported sensitivity to coldness and repeated this process with regards to dominance. 24 RESULTS Questionnaire Internal Consistencies and Descriptive Statistics I first calculated the internal consistencies of the paper measures used in this study (see Table 1). Alphas for the SBI, ISC, and PANAS scales were comparable to those in previous studies. However, Cronbach’s alphas were low for IPIP-IPC Cold-Hearted (DE) (.36), Unassured-Submissive (HI) (.23), and Unassuming-Ingenuous (JK) scales (.19). The relatively lower internal consistencies for these scales compared to other scales in the study may be due to the fact that each scale has relatively fewer items and those items are heterogeneous in content. IPIP-IPC items are somewhat heterogeneous in content whereas SBI items, for example, are all about behaviors. However, this would not explain the difference between the alphas in this study and those of the original study (Markey & Markey, 2009). These low alphas are also consistent with previous reports of low IPIP-IPC octant alphas in studies of college undergraduates (Yalch, personal communication, February 23rd, 2013). Because of the low alpha coefficients of some of the octant scales, the IPIP-IPC measures of interpersonal traits were not considered in the analyses. The IIP-SC Domineering (PA) scale had a low alpha of .36. Having examined the internal consistencies of these measures, I calculated scale and vector means and standard deviations (see Table 1). I had administered several items regarding the participants’ ratings of liking the confederate, finding the confederate attractive, and enjoying the interaction. Internal consistency between the three items regarding liking the confederate was .92. The mean (M = 3.77, SD = 1.22) of these three items was retained as an index of how much the participant liked the confederate. Participant ratings of liking the confederate less than previous studies of unacquainted dyads (M = 7.06) and these ratings varies more (SD = 0.91, Markey, Lowmaster, & 25 Eichler, 2010). These differences may be due to the confederate’s tendency to behave more coldly than interactants in previous studies. Observer-coded Interpersonal Behavior Internal Consistencies and Descriptive Statistics Next, I examined the association between different judges’ ratings of interpersonal behavior within dyads. The first question was whether judges were synchronized in their assessment of interpersonal behaviors over time. Starting the joystick program at the inappropriate time or being inattentive to changes in interpersonal behavior as they occur may result in different kinds of problems for internal consistency. To check for these judge errors, I first examined the time series for synchronization with cross-correlations with lags of 10. If judges had delayed reactions or improperly synched the joystick coding program with the video, I would expect a local maximum correlation at a lag that differed from zero. When the local maximum correlation did differ from zero, I adjusted the relative time position by moving that time series up or down to match the others. This was true in of one case of confederate warmth and dominance and was attributed to the judge beginning the joystick program before the interaction began. If judges were inattentive or failed to notice changes in interpersonal behavior as they occurred, I would expect the correlation to increase with lag and for there to be no local maximum correlation. This was true of 8 cases. These judges’ ratings for that interactant in that interaction were removed from the analysis. Next, I calculated corrected judge-total correlations for each judge across all dimensions and interactants. Of the 9 problematic observer ratings, 7 were attributable to the three judges with the lowest judge-total correlation (M = .53, SD = .04), who were removed from the analyses. The four judges with the highest judge-total correlation (M = .60, SD = .02) were retained for these analyses. I examined the time series coded by these judges for internal consistency. The 26 mean alpha for momentary warmth was .60 (SD = .14) and the mean alpha for dominance was .76 (SD = .17). These alphas are comparable to those of previous research using the same method demonstrating momentary alphas ranging from .60 to .77 (Lizdek et al., 2012). The mean alpha for confederate momentary warmth (M = .65, SD = .15) was higher than that for participant momentary warmth (M = .55, SD = .12). This difference was statistically significant (t = 3.77, p < .01). This difference may be due to the increased variance in confederate warmth (mean SD = 83.56) compared to participant warmth (mean SD = 78.94; Cohen’s d =.20). Mean alphas for momentary dominance did not differ between participant and confederate (t = -0.45, p = .66). To obtain aggregate momentary scores within an interaction, I averaged ratings for the interpersonal dimensions across judges. Time series were calculated by averaging across the time series of the four joystick judges who were the most consistent, on average, across the entire dataset. Mean warmth and dominance scores across the interaction were retained as interaction-level state measures of those behaviors. When the means for each of the joystick judges were treated as items, Cronbach’s alpha for participant warmth was .79, whereas for dominance it was .83, suggesting that judges generally agreed which participants were more dominant and which were warmer across dyads. Alphas for the confederate’s behaviors were similar with an alpha for warmth of .73 and .82 for dominance. Having examined internal consistency, I computed the means. Participant mean observercoded warmth ranged from -8.54 to 330.59 (M = 160.75, SD = 56.92), suggesting that participants tended to be warm towards the confederate. Participant mean observer-coded dominance ranged from -337.94 to 222.28 (M = -.36.81, SD = 132.24). Confederate means for 27 warmth ranged from -107.32 to 197.68 (M = 70.37, SD = 63.43) whereas dominance means ranged from -232.72 to 318.09 (M = 100.94, SD = 109.58). Overall, participants were significantly more warm (t = 7.71, p < .001) and less dominant (t = -6.75, p < .001) than the confederate. This difference may be attributable to the strong social press to behave agreeably as observed by the base-rate of agreeable behaviors in previous research (mean participant warmth across interaction = 194.95, Sadler et al., 2009). This result also provides evidence that the confederate effectively varied his interpersonal behaviors around the circumplex. Consistent with this interpretation, the mean standard deviation of confederate behaviors within dyads was 83.56 for warmth and 201.48 for dominance, suggesting that the confederate varied his interpersonal behaviors within the interactions. As expected, mean warmth and dominance within-person were not significantly correlated for either the participant (r = .17, p = .18; see Table 2 for bivariate correlations) or the confederate (r = .16, p = .20) across dyads. Within dyads, the average correlation between confederate warmth and dominance (M = .05, SD = 0.2) did not differ significantly from zero (t = 1.78, p = .08). However, the average within-dyad correlation between participant warmth and dominance (M = -.14, SD = .19) was significantly less than zero (t = -5.47, p < .001), further suggesting that participants typically behaved more warmly when submissive. Demonstrating robust complementarity, participant and confederate dominance across dyads were strongly negatively correlated (r = -.62, p < .001). Mean warmth scores across dyads were positively, but not significantly, correlated (r = .19, p = .14). Within dyads, the mean correlation between participant and confederate warmth (M = .36, SD = .20) was significantly greater than zero (t = 13.10, p < .001) and the mean correlation between participant and confederate dominance (M = -.68, SD = .12) was significantly less than zero (t = -29.48, p < 28 .001). After removing linear trends, these mean correlations changed to .35 for warmth and -.66 for dominance and remained significantly different from zero. To test against the possibility that these associations were due to coincidental linear and sinusoidal trends in interpersonal behavior, I tested the complementarity within-pseudodyads formed by comparing a participant’s interpersonal behavior within one dyad to the confederate’s behavior in another. Mean correlations between participant and confederate warmth did not differ significantly from zero before detrending (mean r = .00, t = 0.19, p = .85) or after detrending (mean r = .01, t = 0.52, p = .61). Mean within-dyad complementarity coefficients did not significantly differ from zero either before (mean r = .03, t = 1.84, p = .07) or after detrending (mean r = .02, t = 1.75, p = .09). Thus, within-dyad complementarity manifest between the behaviors of individuals who were interacting with one another, but not between the behaviors of individuals who were not. Overall, these results replicate previous findings suggesting that interpersonal behaviors are complementary across half-second intervals (Sadler et al., 2009). Hypothesis 1: Self-reported interpersonal style will relate to observer-coded interpersonal behavior in a laboratory paradigm To test the first hypothesis, I calculated the correlation between mean observer-coded participant warmth and dominance and self-reported warm and dominant behaviors across dyads. Mean observer-coded participant warmth was significantly associated with self-reported warm behaviors (r = .38, p < .01), whereas it was not significantly associated with self-reported dominant behaviors (r = .14, p = .29). Similarly, mean observer-coded participant dominance was significantly associated with self-reported dominant behaviors (r = .35, p < .01), and not 29 associated with self-reported warm behaviors (r = .03, p = .81). These findings were consistent with the first hypothesis (see Table 3 for summary of hypothesis tests). Self-reported interpersonal behaviors on the SBI were also predicted by self-reported interpersonal traits on the SBI. Warm traits predicted self-reported warm behaviors (r = .30, p < .01) and dominant traits predicted self-reported dominant behaviors (r = .41, p < .01). However, the association between self-reported interpersonal traits and mean observer-coded participant behaviors was not as strong. Trait dominance was not significantly associated with observerrated dominant behaviors (r = .13, p = .30) and the association between trait warmth and observer-rated warm behaviors trended towards significance (r = .23, p < .10), consistent with the expectation that individuals’ report of their own behaviors over brief intervals more accurately predict observer ratings of those behaviors than does that individuals’ report of their own interpersonal traits. This may be due to the influence of individual’s self-perception on their report of their own behaviors. Individuals may be more likely to interpret their ambiguous behaviors as consistent with their self-reported interpersonal traits. Participant reports of their own behaviors may be influenced by the evaluative quality of those behaviors, particularly with regards to warmth. I next compared observer-coded confederate behaviors to participant-reported confederate interpersonal behaviors. Observer-coded mean confederate warmth was significantly associated with participant-reported confederate warmth during the interaction (r = .35, p < .01), but not with participant-reported confederate dominance during the interaction (r = .21, p < .10). The association between mean observer-coded confederate dominance and participant-reported confederate dominance during the interaction trended towards significance (r = .23, p = .07) while its association with participant-reported confederate warmth during the interaction also 30 failed to reach significance (r = .16, p = .20). The failure of these associations to reach statistical significance may be due to a lack of power, as these associations are small to moderate (Cohen, 1988). Furthermore, associations on the off-diagonal suggest that participants may have perceived the confederates’ dominant behaviors as somewhat warm. This violation of the assumption of orthogonality is consistent with previous research on the halo effect suggesting that liking an individual may cause that individual to be seen as both warm and dominant at the level of global appraisal of personality traits (Anusic et al., 2009). When controlling for participants’ rating of liking the confederate, the association between participant-reported confederate warmth and observer-coded confederate dominance decreased to .10 (p = .42), suggesting that this failure to find cross-method orthogonality may be due to the influence of the halo effect on participant perceptions of the confederate. Overall, these findings only provide partial support for the first hypothesis because the association between participant-rated and observer-rated confederate dominance did not reach statistical significance whereas the correlation for warmth was significant. Hypothesis 2: Self-reported traits will relate to self-reported interpersonal sensitivities such that traits will correlate with their opposite sensitivity I calculated the correlation between interpersonal sensitivity vectors and interpersonal trait vectors. The ISC sensitivity to warmth vector was significantly negatively correlated with the SBI trait warmth vector (r = -.31, p = .01), comparable to previous research showing this negative association between warmth sensitivity and warm interpersonal traits (rs range from .27 to -.34, Hopwood et al., 2011). Sensitivity to warmth was not significantly related to the SBI dominance vector (r = -.12, p = .30), consistent with previous research (rs range from -.22 to .05). The ISC sensitivity to dominance vector was not as strongly associated with the SBI 31 dominance vector (r = -.14, p = .26) as it was to trait measures of dominance in previous research (rs range from -.25 to -.26). The SBI warmth vector was not strongly associated with sensitivity to dominance (r = .13, p = .28), consistent with the previous findings (rs range from .20 to .27). Overall, these findings were consistent with the hypothesis with regards to warmth, but not as consistent with regards to dominance. This finding may be a function of the content of SBI items. These items focus on behaviors rather than personality traits. To the extent that personality traits more directly relate to interpersonal sensitivities than do habitual behaviors, the SBI may not bear as clear an association to interpersonal sensitivities. Next, I tested the association between interpersonal sensitivities and interpersonal problems. Warm interpersonal problems were significantly negatively correlated with sensitivity to warmth (r = -.34, p < .01), while it was not significantly correlated with sensitivity to dominance (r = .13, p = .30). Dominant interpersonal problems were significantly negatively correlated with sensitivity to dominance (r = -.43, p < .01), while it was not significantly correlated with sensitivity to warmth (r = -.02, p = .87). Overall, these findings were consistent with the second hypothesis and previous research demonstrating a negative association between interpersonal problems and interpersonal sensitivities (e.g., -.27 between dominant problems and dominant sensitivities, Hopwood et al., 2011). Hypothesis 3: Observer-coded cold behaviors will relate to anxiety The third hypothesis – that observer-coded interpersonal warm behaviors and participant anxiety would be associated – was tested both across dyads and within dyads. I tested this hypothesis with three indicators of anxiety: self-reported negative affect, self-reported liking the confederate, and skin conductance reactivity. Negative affect reflects the most general distress described by interpersonal theory to be associated with deviations from complementarity. 32 However, self-reported negative affect was not measured within dyads. Skin conductance reactivity was measured to inform on the participants’ affective experience within dyads. For across-dyad tests, I calculated the effect of self-reported warm behaviors and mean observercoded warm behaviors on self-reported liking the confederate, self-reported negative affect, and mean skin conductance reaction. For within-dyad tests, I calculated the effect of interpersonal behaviors on skin conductance across half-second intervals for each dyad. I then calculated the whether the proportion of dyads demonstrating the hypothesized association differed significantly from 50% and whether the mean effect for the dyads differed significantly from zero. Across-dyad warmth and liking the confederate I first tested the hypothesis that participant and confederate cold interpersonal behaviors should be associated with increased anxiety at the interaction level by computing the relation between means of observer-coded interpersonal warmth and participant ratings of liking the confederate based on the assumption that participants who experienced less anxiety during the interaction would rate liking the confederate more. This composite liking score was not significantly associated with either mean confederate warmth (r = .10, p = .46) or mean participant warmth (r = .19, p = .14). I next tested this association with participant perceptions of their own warm behavior and the confederate warm behaviors. Participants’ rating of how much they liked the confederate was strongly associated with their perceptions of the confederate’s warm behaviors during the interaction (r = .61, p < .01) and with their perceptions of their own warm behaviors during the interaction (r = .48, p < .01). These findings provide partial support of the hypothesis in that participants who liked the confederate behaved more warmly toward him. While correlation coefficients between observer-coded warm behaviors and participants’ 33 ratings of liking the confederate were positive but small, they did not reach statistical significance. Power analysis suggests that a total sample size of 100 would have a power of .8 in detecting a significant association with an effect of this size. Across-dyad warmth and negative affect Self-reported anxiety was measured with pre- and post-interaction PANAS negative affect scales. Overall, participants experienced a significant decline (t = 3.54, p < .001) in negative affect from before the interaction (M = 1.70, SD = 0.59) to after the interaction (M = 1.37, SD = 0.51). To test whether warm behaviors moderated change in negative affect during the interaction, I calculated correlations between warm behaviors, the average between pre- and post-interaction negative affect, and the residual of post- on pre-interaction negative affect. I first examined this association with observer-coded behaviors. Observer-coded participant warmth did not significantly correlate with post-interaction negative affect when controlling for preinteraction negative affect (r = -.00, p = .98). Participant warmth was more strongly, but not significantly, related to mean negative affect (r = -.10, p = .43). Observer-coded confederate warmth was not significantly related to the residual score (r = .02, p = .82) and trended towards associating significantly with the average negative affect (r = .24, p < .10). I next examined the associations between negative affect and interpersonal warmth as rated by the participant on the SBI. Self-reported warmth during the interaction was associated with the average negative affect score (r = -.27, p < .05) and negative affect residual score (r = -.34, p < .01). Participantreported confederate warmth was associated with the average negative affect (r = -.25, p < .05) and the post-interaction negative affect after controlling for pre-interaction negative affect (r = .25, p < .05). As with ratings of liking the confederate, examination of the change in selfreported negative affect provided partial support for the hypothesis. Participants’ reports of warm 34 behaviors significantly predicted negative affect, but observer-coded warm behaviors did not. These findings may suggest that participants’ perceptions of others’ behaviors are more important than other observers’ perceptions of others’ behavior to participants’ affective experience. Additionally, the failure to find significant effects with observer-coded warmth may be a result of testing these effects across different methods and, conversely, significant associations between of self-reported behaviors and self-reported negative affect may be artifacts of method overlap. Across-dyad warmth and skin conductance Mean observer-coded cold behaviors were compared to mean skin conductance measures to evaluate the relationship between interpersonal behaviors and skin conductance reactions across dyads. Overlapping video and skin conductance data were available for 55 dyads. I evaluated this association by calculating residual scores from regressing mean skin conductance during the interaction on mean baseline skin conductance. Mean confederate warmth did not increment mean baseline skin conductance in predicting mean skin conductance during the interaction (r = -.10, p = .48). When mean skin conductance was averaged between the baseline and the interaction, that average was not significantly associated with mean confederate warmth (r = .18, p = 19). Mean participant warmth trended towards significance in predicting skin conductance above baseline skin conductance (r = -.26, p < .10) and was less strongly associated with mean skin conductance (r = .08, p = .54). While the coefficients were negative as predicted, the associations did not achieve statistical significance, failing to support the hypothesis. I also tested this hypothesis using participants’ reports of their own interpersonal warmth and of the confederate’s interpersonal warmth. Participant-rated confederate warmth was not significantly correlated with the skin conductance residual (r = -.07, p = .61) and trended towards 35 significantly predicting mean skin conductance (r = .25, p < .10). Self-reported behavioral warmth trended towards significantly predicting mean skin conductance (r = -.24, p < .10) but was less strongly associated with the negative affect residual score (r = -.14, p = .31). These findings were also inconsistent with the hypothesis. Within-dyad warmth and skin conductance The hypothesis that warm interpersonal behaviors would relate to skin conductance within dyads was tested by calculating indices of within-dyad association for each dyad and employing one-sample t-tests to determine whether the average coefficient significantly differed from zero and binomial probability tests to determine whether significantly more interactions exhibit the hypothesized relationship than would be expected by chance. I calculated three indices of association between momentary confederate warmth and momentary participant skin conductance. The first was the correlation between the raw time series data, the second was the correlation between the detrended time series data, and the third was the correlation between best fit wave approximations of confederate warmth and participant skin conductance. The first index measured the simple correlation between the two time series. The second index removed linear trends from the time series to reduce spurious correlations due to linear trends. The third index smoothed the data into the single wave that best accounted for variance in the data and measured the correlation between those smoothed time series. The same three indices were computed for the association between participant warmth and skin conductance. I first tested to whether the mean correlations between momentary skin conductance and momentary participant and confederate warm behaviors significantly differed from zero with one-sample t-tests and then tested whether there were a significant difference in the number of cases with positive versus negative Pearson’s r coefficients. The mean association between 36 momentary confederate warmth and momentary participant skin conductance (M = 0.00, SD = 0.25) was not significant (t = 0.09, p = .93) and 28 cases (50%) exhibited a negative association. The mean association between participant warmth and skin conductance (M = -0.06, SD = 0.24) trended towards being significantly less than zero, (t = -1.76, p < .10). Moreover, there were significantly more cases with a negative r coefficient (N = 36, 64%) than with a positive r coefficient (p < .05). I repeated the one-sample t-test after removing linear trends from the time series data. After detrending, the association between confederate warm behaviors and participant skin conductance (M = -0.03, SD = 0.22) remained nonsignificant (t = -1.07, p = .29). The number of cases with a negative association between confederate warmth and skin conductance increased to 32, but this proportion remained nonsignificant (p = .35). However, the mean correlation between momentary participant warmth and skin conductance (M = -.06, SD = .18) significantly differed from zero (t = -2.41, p < .05) and the same number of cases displayed a negative r coefficient. Partial support was found for the within-dyad association between behavioral warmth in an interaction and the participant’s skin conductance. Skin conductance was higher in moments that participants were behaving coldly, but not necessarily those in which the confederate was behaving more coldly. Best-fit waves for interpersonal behavior and skin conductance time series were conducted to determine whether they were associated at the level of underlying waveforms. On 2 average, skin conductance data was better described by single waves (mean R = .34) than were 2 interpersonal behaviors (mean R = .12). Best-fit wavelengths for confederate warmth ranged from 46 to 1091 seconds (M = 505.78, SD = 319.37) while best-fit wavelengths for participant warmth ranged from 10 to 454 seconds (M = 410.46, SD = 252.24). When participants’ warmth and skin conductance time series data were modeled as the waves that best accounted for variance across those time series, the mean correlation between the participant’s skin 37 conductance and the participant’s behavioral warmth (M = -0.07, SD = 0.37) did not differ significantly from zero (t = -0.98, p = .33) nor did the correlation of the skin conductance best-fit wave with the best-fit wave for confederate warmth (M = -.04, SD = 0.37, t = -0.85, p = .40). This finding may relate to the finding that the best-fit wavelengths for skin conductance did not significantly correlate with those for participant warmth (r = .19, p = .16) or confederate warmth (r = .07, p = .60), suggesting that the associations between skin conductance and warmth are not better accounted for by single best-fit waves. These findings might differ if the data were better described by the combination of two or more waves. In future analyses, I plan to examine the coherence between time series by decomposing that association into the mean weighted association between time series across multiple waves of differing frequencies using crossspectral analysis. Hypothesis 4: Deviation from complementarity between observer-coded participant and confederate warm behaviors will provoke participant anxiety after accounting for main effects of participant and confederate warmth Within-dyad warmth complementarity and skin conductance To test the hypothesis within dyads, I first calculated time series of momentary deviation from warmth complementarity by taking the absolute value of the difference between participant and confederate momentary warmth time series. Next, I calculated a partial correlation between skin conductance and deviation from warmth complementarity controlling for participant and confederate warmth for each interaction. The mean partial correlation was .03 (SD = 0.11). The mean partial correlation was significantly greater than zero in a one-sample t-test (t = 2.01, p < .05). I repeated this process after removing linear trends from the data. The resulting mean partial correlation was .01 (SD = 0.10), which was not significantly different than zero (t = 0.69, p = .50). This difference in findings between raw and detrended findings suggests that any 38 association between momentary anxiety and momentary warmth complementarity maybe be best accounted for by general trends in the interaction. After modeling each of these time series as the wave that best described them, the mean partial correlation (M = .02, SD = 0.27) did not significantly differ from zero (t = 0.65, p = .52). This finding suggests that any within-dyad association between momentary skin conductance and momentary deviation from warmth complementarity is not best explained by a single best-fit wave underlying each time series. This may be because warmth and skin conductance time series data are better explained by multiple waves, in which case simultaneous consideration of a range of wavelengths would be more appropriate. Across-dyad warmth complementarity and skin conductance To test this association across dyads, I entered mean participant warmth, mean confederate warmth, and mean participant skin conductance during the task instruction baseline in the first block of a hierarchical linear regression and mean deviation from warmth complementarity in a second block predicting mean skin conductance during the interaction. Mean deviation from warmth complementarity did not significantly increment baseline skin conductance and each individual’s warmth in predicting skin conductance during the interaction (Beta = -.14, t = -1.38, p = .17, see Table 4 for all regression model summaries). However, after detrending, the second step of the regression was significant (Beta = -.13, t = -2.72, p < .01). The second step of the regression was not significant when deviation from the complementarity of best-fit waves was considered (Beta = .09, t = 1.81, p = .08). These results were inconsistent with the hypothesis, which predicted an increase rather than a decrease in skin conductance in interactions with more deviation from warmth complementarity. This difference in sign may be because individuals who feel more insecure may take additional efforts to complement the other 39 person, but react to the anxiety with these efforts too slowly to affect within-dyad associations between anxiety and deviation from warmth complementarity. I also tested this association across-dyads with participant’s self-report and confederaterated warmth during the interaction. An SBI score for deviation from warmth complementarity was calculated by taking the absolute value of the difference between participant ratings of their own warmth and ratings of confederate warmth. This deviation from complementarity score did not increment baseline skin conductance and participant ratings of each person’s warmth in predicting mean skin conductance during the interaction (Beta = -.18, t = -0.74, p = .46). This finding is similar to the results for the main effect of warmth suggesting participant SBI ratings of warmth are not strongly associated with skin conductance reactivity. Across-dyad warmth complementarity and liking the confederate To test the effect of deviation from warmth complementarity on how much the participant rated liking the confederate, I used a hierarchical linear regression, as before. Before detrending, mean deviation from warmth complementarity did not significantly increment each individual’s warmth in predicting how much the participant would rate liking the confederate (Beta = -.37, t = -1.33, p = .19). This association did not reach significance despite a strong beta coefficient. The beta was weaker after detrending (Beta = -.01, t = -0.03, p = .97) and when smoothing into best-fit waves (Beta = .01, t = .07, p = .95). The difference in beta coefficients between trended and detrended time series suggest that liking a person is more related to complementarity over longer time intervals (i.e., 15 minutes) rather than shorter time intervals (i.e., 0.5 seconds). I performed analogous hierarchical regressions with the participant’s SBI ratings of warm behaviors during the interaction. The deviation from warmth complementarity did not increment 40 the participant’s perception of each interactants’ warm behaviors during the interaction in predicting how much the participant would rate liking the confederate (Beta = .29, t = 0.91, p = .37). These findings did not support the hypothesis that more warmth complementarity should be associated with better liking the other person. Across-dyad warmth complementarity and negative affect I next tested the association between across-dyad warmth complementarity and negative affect. Before detrending, mean deviation from warmth complementarity did not significantly increment pre-interaction negative affect and each person’s warm behaviors in predicting postinteraction negative affect (Beta = .20, t = 0.94, p = .35). However, after detrending, this prediction became significant (Beta = .21, t = 2.01, p < .05). This effect was not replicated with best-fit waves (Beta = -.08, t = -0.73, p = .47). This finding suggests that the degree to which interactants are similar or different in warmth across brief time intervals can have an effect on the affective experience of those interactants. I also tested this association across dyads with participants’ self-report warmth and participant perception of confederate warmth during the interaction. Deviation from complementarity between participant ratings of self and others’ interpersonal behaviors did not significantly increment baseline negative affect and each individual’s behaviors in predicting post-interaction negative affect (Beta = -.33, t = -1.20, p = .24). Thus, the findings supported the hypothesis with regards to observer-rated warm behaviors, but not self-reported warm behaviors. Hypothesis 5: Deviations from complementarity between observer-coded participant and confederate dominant behavior will provoke participant anxiety Within-dyad dominance complementarity and skin conductance 41 The fifth hypothesis – that deviation from dominance complementarity should increase anxiety – was tested both within dyads and between dyads. To test this hypothesis within-dyads, I calculated mean partial correlations between momentary deviation from dominance complementarity and skin conductance controlling for each person’s dominance. The mean partial correlation between deviation from dominance complementarity and skin conductance within interactions was -0.01 (SD = .11), which did not significantly differ from zero (t = -0.86, p = .40). After detrending, the mean changed to -0.002 (SD = .10), which also did not significantly differ from zero (t = -0.17, p = .87). With each of these analyses, 31 interactions (55%) demonstrated a positive partial correlation, which did not significantly differ from 50% (p = .504). Use of best-fit waves in this analysis also yielded a negative result (M = -.02, t = -1.17, p = .25). With this analysis, 23 interactions (42%) demonstrated the predicted associations. These findings were inconsistent with the hypothesis that momentary skin conductance would be positively associated with momentary deviations from dominance complementarity. These findings may differ from those of warmth for a number of reasons. Theoretically, security and self-esteem, the domains of interpersonal anxiety relating to warmth and dominance, respectively, may manifest in different physiological modalities. Alternatively, the physiological consequence of deviations from dominance complementarity may be different depending on whether interactants are both dominant or are both submissive. Across-dyad dominance complementarity and skin conductance To determine the relation across dyads, I employed a hierarchical regression to determine whether mean deviation from dominance complementarity incremented each interactant’s dominance and the participant’s mean baseline skin conductance in predicting the participant’s mean skin conductance during the interaction. This second block of the regression did not 42 significantly increment the first in predicting mean skin conductance (Beta = -.02, t = -0.24, p = .81). The results after detrended were the same (Beta = -.05, t = -1.06, p = .92), as were the results when using best-fit waves (Beta = -.11, t = -0.88, p = .38). These findings were inconsistent with the hypothesis that mean skin conductance would be positively associated with mean deviations from dominance complementarity. These results extend the previous results in suggesting that observer-coded dominance complementarity is not associated with skin conductance across interactions. I also tested this association with participant’s self-reported dominance and perceived confederate dominance during the interaction. Deviation from complementarity between participant ratings of self and others’ interpersonal behaviors did not significantly increment baseline skin conductance and each individual’s behaviors in predicting mean skin conductance during the interaction (Beta = .26, t = 1.86, p < .10). This finding suggests that individuals may experience greater arousal when they see others as being similarly dominant or submissive. Across-dyad dominance complementarity and liking the confederate I tested the association between deviations from dominance complementarity and liking the confederate at the interaction level. I performed a hierarchical linear regression to determine whether mean observer-coded deviations from dominance complementarity would increment each interactant’s observer-coded dominance in predicting how much the participant would rate liking the confederate. Before detrending, the second block of this regression did not significantly increment the first (Beta = -.07, t = -0.36, p = .72). This finding was also the case after detrending (Beta = -.09, t = -0.71, p = .48) and when using best-fit waves (Beta = .16, t = 1.18, p = .24). 43 I also examined this hypothesis with participant’s ratings of perceived interpersonal behaviors on the SBI. In the analogous hierarchical regression, deviation from dominance complementarity on the SBI did not increment SBI ratings of participant and confederate dominance in predicting how much the participant would report liking the confederate (Beta = .11, t = 0.31, p = .76). Overall, in contrast with the hypothesis, these results suggest that deviation from dominance complementarity does not affect the degree to which participants liked the confederate. Across-dyad dominance complementarity and negative affect To test the hypothesis that mean deviation from dominance complementarity would increment each person’s mean dominance and the participant’s baseline negative affect in predicting post-interaction negative affect, I performed a hierarchical linear regression. Before detrending, mean observer-rated deviation from dominance complementarity did not significantly increment either person’s dominance and pre-interaction negative affect in predicting post-interaction negative affect (Beta = .04, t = 0.28, p = .78). This was also true of the detrended mean observer-rated deviation from dominance complementarity (Beta = .14, t = 1.40, p = .17) and deviation from the complementarity of best-fit dominance waves (Beta = -.04, t = -0.41, p = .69). I also tested this hypothesis with participants’ perceptions of their own dominant behavior and others’ dominant behavior with the SBI. Participants’ perception of their own and the confederate’s dominant behavior were entered in the first block of a hierarchical regression with pre-interaction negative affect. Deviation from participant-perceived complementarity was entered in the second block to predict post-interaction negative affect. This second block did not significantly increment the first (Beta = -.04, t = -0.13, p = .90). Overall, in contrast with the 44 hypothesis, these results suggest that deviation from dominance complementarity does not affect the degree to which participants experienced an increase in negative affect. Hypothesis 6: The effect of observer-coded confederate behavior on skin conductance will correspond to self-reported interpersonal sensitivities Finally, I tested the sixth hypothesis by correlating across-dyad measures of interpersonal sensitivities with within-dyad correlations between confederate behaviors and participant skin conductance. To test this with warmth, I correlated within-dyad associations between confederate warmth and participant skin conductance with participants’ self-reported interpersonal sensitivity to warmth vector. Interpersonal sensitivity to warmth was not significantly correlated with within-dyad associations between confederate warmth and participant skin conductance before detrending (r = -.08, p = .54), after removing linear trends (r = .00, p = .99), or when modeling the time series data as the single waves that best accounted for variance in those time series (r = .09, p = .53). I repeated with dominance and found that interpersonal sensitivity to dominance was not significantly associated with experiencing greater skin conductance when the other person was dominant. This was true before detrending (r = -.13, p = .36), after detrending (r = -.02, p = .87), and with best-fit waves (r = -.14, p = .31). Overall, these results suggest that individual differences in skin conductance reactivity to others’ interpersonal behaviors are not best explained by self-reported interpersonal sensitivities. Gender Effects on Interpersonal Processes The present study sampled both male and female participants to determine whether the hypothesized effects differed as a function of gender. I performed independent sample t-tests to determine whether study variables differed as a function of gender. I first examined the effects of gender on interpersonal traits and behaviors. Across dyads, females rated themselves as 45 significantly more warm (t = 2.31, p < .05) and more prone to warm interpersonal problems (t = 3.53, p < .01). Male and female participants did not significantly differ on other self-reported interpersonal measures or on observer ratings of interpersonal behavior. I next examined the moderating effect of gender on the association between self-reported interpersonal behaviors and observer ratings of interpersonal behaviors. In a hierarchical regression testing the moderating effect of gender on the association between self- and otherreported warm behaviors, the interaction did not significantly increment the main effects (Beta = .005, t = .04, p = .97). The interaction between gender and dominance trended towards significant (Beta = .23, t = 1.89, p < .10), suggesting that the association between self-reported dominant behaviors and observer-rated dominant behaviors may be stronger for males. Regression summaries for findings presented here are detailed in Table 4. Summaries for the nonsignificant regressions are available on request. I next examined the moderating effect of gender on the association between self-reported interpersonal traits, problems, and sensitivities. In this study, gender did not have a significant main effect on sensitivity to warmth. The moderating effect of gender on the association between trait warmth and sensitivity to warmth trended towards significance (Beta = .20, t = 1.67, p < .10). Results for the moderating effect of gender on the association between warm interpersonal problems and sensitivity to warmth also trended towards significance (Beta = .20, t = 1.74, p < .10). These findings suggest that warm females may be more sensitive to interpersonal coldness than warm males. Gender moderation of the effects of trait dominance and dominant interpersonal problems on sensitivity to dominance did not approach significance. I next examined the effect of gender on the association between interpersonal coldness and distress. None of the within-dyad effects of interpersonal processes on skin conductance 46 differed depending on gender, so I tested the across-dyad effects of confederate warmth on across-dyad measures of distress. The only moderating effect approaching significance was the effect on the association between confederate warmth and liking (Beta = .25, t = 1.90, p < .10). Specifically, confederate warmth was more strongly associated with liking the confederate among male participants. Given the limitations of the study design, it is unclear as to whether this potential effect is driven by participant gender or by same-sex dyads. I examined the moderating effect of gender on participant warmth in the same manner. None of these moderating regressions approached significance. I also investigated the effect of gender on the association between interpersonal complementarity and distress. Within-dyad effects of interpersonal complementarity on skin conductance did not significantly differ between genders in independent samples t-tests. I tested across-dyad effects of gender on the association between complementarity and anxiety with hierarchical regression. The moderating effects of gender on liking the confederate were significant for detrended warmth complementarity (Beta = -.36, t = -3.00, p < .01) and trended towards significance for detrended dominance complementarity (Beta = -.27, t = -1.90, p < .10), suggesting that the association between complementarity and liking the confederate was stronger for female participants. Gender did not moderate the effects for increase in skin conductance or negative affect. Finally, I employed hierarchical regression to determine whether the interaction between gender and interpersonal sensitivities incremented each in predicting detrended within-dyad associations between confederate behaviors and participant skin conductance. The moderating effect of gender on the association between across-dyad sensitivity to warmth and within-dyad skin conductance response to warmth was significant (Beta = .36, t = 2.63, p < .05), suggesting 47 that male participants who reported being sensitive to coldness had greater skin conductance reactions while the confederate was behaving coldly compared to female participants who reported being sensitive to coldness. Overall, these findings suggest that the effects of interpersonal processes on anxiety may depend on the gender of the interactants. However, given the limitations of the interaction task it is unclear whether these differences are due to the different-sex versus same-sex dyadic interaction or purely due to gender. Alternatively, given the number of moderating regressions, these effects may be due to type I error. 48 DISCUSSION According to interpersonal theory, complementarity arises as an interchange of interpersonal bids and reactions to those bids. These bids are expressed, received, and reacted to at the level of behaviors (Tracey, 2004), suggesting the relation between a relationships’ degree of trait complementarity and relationship satisfaction may be due to a propensity towards complementary behaviors across interactions and within interactions. The present study tested the effects of interpersonal complementarity on anxiety across time intervals of differing duration. Interpersonal behaviors in this study were complementary across dyads and within dyads and were orthogonal within person, consistent with previous research and the assumptions of interpersonal theory. Observer ratings of warmth complementarity across dyads failed to reach significance, which may be a consequence of using a single confederate who varied his warmth within dyads moreso than between dyads. The relatively stronger effect of complementarity at the level of specific behaviors is consistent with the interpersonal theory of complementarity at different time intervals. One purpose of this study was to determine the effect of interpersonal warmth on anxiety across dyads and within dyads. Interpersonal warmth in the interaction was related to participant’s liking the confederate, reporting less negative affect, and less skin conductance reactivity. These findings are partially consistent with previous research on the desirability of warm behaviors in interpersonal situations (Hill & Safran, 1994; Locke & Sadler, 2007) and the effect of interpersonal coldness on physiological reactions (e.g., Gallo, Smith, & Kircher, 2000). These findings also suggest that the association between an individual’s interpersonal warmth and that individual’s physiological reactions may manifest within interactions, across halfsecond intervals. 49 A second purpose of the present study was to examine the effect of warmth complementarity on anxiety across dyads and within dyads. Interpersonal theory suggests that warm behaviors express communal motives and constitute bids for affiliation (Horowitz et al., 2006). When behavioral responses deviate from complementarity, the motives are frustrated, provoking anxiety. The findings of the present study provide partial support for the hypothesis that warmth complementarity relates to participant’s anxiety across dyads. Specifically, deviation from warmth complementarity within dyads was found to relate to self-reported negative affect across dyads. These findings contribute to a growing body of evidence suggesting that frustration of interpersonal motives is associated with distress and contribute to initial evidence that interpersonal motives may vary within dyads and may be expressed dynamically across the course of an interaction. This study did not find an association between interpersonal complementarity of dominant behaviors and distress within dyads or across dyads. These negative findings were consistent with previous research demonstrating an effect of warmth complementarity, but not dominance complementarity, on satisfaction with an interaction (Markey, Lowmaster, & Eichler, 2010). The failure of these studies to identify an association between dominance complementarity and interpersonal distress may be due to the operationalization of these tests of complementarity, differences between communal and agentic dimensions of motives and behavior, or both. Alternative Tests of Interpersonal Complementarity This study provided an opportunity to investigate interpersonal complementarity in a novel situation. Hypotheses regarding the effects of warmth complementarity were partially supported. Within-dyad associations between warmth complementarity and skin conductance 50 were negative across three indices. However, the association between complementarity during the interaction and increase in negative affect across dyads was significant for detrended time series. In contrast, observer-rated deviation from dominance complementarity did not relate to participants ratings of negative affect or liking the confederate, nor did it relate to skin conductance reactions across dyads or within dyads. The findings of this study, including failure to provide support for several hypotheses regarding complementarity, must be considered in light of how the tests of complementarity effects were operationalized. Complementarity effects may have differed depending on the task, the interactants, the criteria for interpersonal distress, and the nature of the association. The Effect of the Task on Tests of Interpersonal Complementarity Previous examinations of interpersonal complementarity have differed in the interpersonal context in which these behaviors were observed. Tests of complementarity across dyads have differed in the task that dyad was instructed to perform. For example, several previous studies showing complementarity did so in the context of performance-based tasks in which participants were meant to cooperate (Strong et al., 1988; Bluhm, Widiger, & Miele, 1990; Sadler et al., 2009; Markey, Lowmaster, & Eichler, 2010) whereas relatively little research has examined complementarity in dyadic interactions where the task was less structured (Markey, Funder, & Ozer, 2003) or less performance-oriented. In contrast with much of the extant research on complementarity across interactions, the task employed in this study was one in which interactants were instructed to get to know one another. The task was semi-structured in the sense that interactants were given discussion prompts. However, the nature of the task did not demand that they adhere to those discussion prompts. In past tests of complementarity, the tasks demanded that the dyads perform a task efficiently. In such an interpersonal situation, the 51 allocation of control – who is leading and who is following – may be more important to anxiety. In this study, neither interactant had interpersonal motives for the dyad to perform the task efficiently. Consequently, when participants’ interpersonal bids to lead the confederate or be led by the confederate were frustrated with the confederate’s dominant or submissive behavioral responses, respectively, that deviation from complementarity may not have frustrated the participant’s motives as severely as if the participant had more distinct agentic goals. This would make sense of the lack of association between deviation from dominance complementarity and participants’ anxiety. If the interaction task had involved cooperation in achieving an agentic goal, within- and across-dyad frustration of interpersonal motives for allocation of control may have been more closely associated with interpersonal distress. A number of features of the task may have affected interpersonal motives for affiliation during the task. The interaction task was a discussion between the dyads which emphasized intimacy and self-disclosure (Aron et al., 1997), which may have resulted in more affiliative task demands than previous studies of complementarity. Conversely, the confederate was trained to be colder on average (M = 70.37) than participants in this study (M = 160.75) and in previous studies of unacquainted dyads (M = 194.95, Sadler et al., 2009). Moreover, the confederate was instructed to intentionally vary his interpersonal behaviors within the interaction, but not across interactions, which may relate to stronger findings of warmth complementarity within interactions than across interactions. Finally, participants’ behaviors and reactions to the confederates’ behaviors may have been influenced by their reactions to the equipment including the video camera and psychophysiological measures, in the presence of which they may have been more uncomfortable than in typical dyadic interactions. The features of the interaction 52 prompts and confederate instructions may influence the effect of deviation from warmth and dominance complementarity on interpersonal distress. The Effect of the Operationalization of Anxiety on Tests of Interpersonal Complementarity Previous tests of interpersonal complementarity have examined anxiety at the broadest time interval in terms of relationship satisfaction (e.g., Yaughn & Nowicki, 1999) and at intermediate time intervals in terms of satisfaction with an interaction (e.g., Dryer & Horowitz, 1997). The present study tested the effects of interpersonal complementarity on anxiety at intermediate time intervals (15 minute interactions) and brief time intervals (half-seconds). The criterion measures employed in the present study differed in how directly they measured participant anxiety. Overall, these results may suggest that deviation from warmth complementarity – but not from dominance complementarity – results in negative affect and, in some analyses, skin conductance. However, in the present study, skin conductance was not significantly correlated with self-reported negative affect, suggesting that it may not measure anxiety broadly. The present findings would suggest the participant was more physiologically relaxed when behaving warmly and their warm behaviors were more ‘in tune’ with those of the confederate. In interpersonal terms, participants were more at ease when their bids for interpersonal closeness or distance were reciprocated. The increase in skin conductance within dyads when the interactants behaviors deviated from warmth complementarity may suggest that participants became more physiologically aroused when the confederate’s behaviors suggested he wanted to get too close or be too distant. That is, deviations from warmth complementarity may alert interactants to the danger of interpersonal rejections or interpersonal intrusions, which 53 would be consistent with the postulations of interpersonal theory (Horowitz et al., 2006). After acting on those cues to deviations from warmth complementarity, individuals’ negative affective responses may depend on whether their protections against interpersonal rejection or intrusion were successful. Future examination of the effects of complementarity on anxiety within dyads should include investigate other psychophysiological measures affective reactions, such as cardiovascular reactivity (Vella & Friedman, 2009) and cortisol response (Pope & Smith, 1991). The present study was limited by the operationalization of affect in terms of the PANAS and a previously uninvestigated measure of liking the confederate. While the measure was internally consistent, its validity in measuring liking has not previously been studied. Whether or not the measure accurately reflected how much the participant liked the confederate, the degree to which a person is found to be likable is not a direct test of the degree to which interacting with that person provoked anxiety. The use of the PANAS results in a similar limitation. Positive and negative affect factors may not best capture the vector of affect most closely associated with deviations from interpersonal complementarity, nor do they appear to capture the vector most closely associate with physiological arousal measured by skin conductance. The Effect of the Interactants on Tests of Interpersonal Complementarity In the present study, the confederate was the same person across all dyads – a Caucasian male college student. The present study sampled primarily Caucasian college students. Consequently, care must be taken in generalizing the results. The effects in this study may be due to the interactants’ demographic factors including gender and occupation. The effects of complementarity may be different in populations of different ethnicities or in different professional settings (Moskowitz et al., 2006). The nature of the relationship between the interactants may also have had an effect on the tests of complementarity. Interactants were 54 unacquainted individuals who were ostensible peers as undergraduate students. The effects of complementarity may differ depending on whether the interactants are peers versus whether they are in professional, romantic, or other relationships (Dryenforth et al., 2010). The use of one confederate instead of many was a limitation of this study. For example, initial findings suggest that confederate warmth may have been more important to how much male participants liked the confederate whereas warmth complementarity may have been more important to how much female participants reported liking the confederate. However, given the nature of the study design, same-sex versus different-sex dyadic status was perfectly confounded with participant gender. Consequently, it is unclear which factor is driving these associations. It may be the case that warmth is more important to liking a member of the same sex whereas complementing warmth is more important to liking a member of the opposite sex. Alternatively, male ratings of liking others may be more strongly driven by main effects of warmth whereas female ratings of liking others may be more strongly driven by warmth complementarity. Both may be the case. However, the present study was not able to test these alternative hypotheses. Future tests of interpersonal complementarity should sample dyads of different sexes, male same-sex dyads, and female same-sex dyads to investigate which of these factors may be driving gender effects. Future tests which might wish to manipulate interpersonal warmth may also employ multiple confederates to disentangle the effect of the individual’s behavior from other aspects of that individual. The Effect of the Nature of Association on Tests of Interpersonal Complementarity The present study examined within-dyad associations with three indices: correlations between raw time series, correlations between detrended time series, and correlations between detrended time series as modeled by the waves that best accounted for their variance. Each index 55 captures a different aspect of association and differences in findings between these indices may be attributable to how they measure association. Correlation between raw time series data is the simplest test that may bias the test the effect with general trends across the interaction. These linear trends may obscure momentary effects or lead to spurious correlations. Thus, the trending effect of deviation from warmth complementarity on skin conductance within the interaction found in the present study may be a spurious consequence of linear trends. The possibility that within-dyad associations are spuriously driven by linear trends can be precluded by removing linear trends from the data. Doing so allowed for more direct tests of the association behaviors and affects at the momentary level. However, previous research suggests that the association between two individuals’ behaviors can be expressed in terms of underlying waves (Sadler, et al., 2009). The correlation between detrended time series data does not account for these underlying waves in testing the association between time series. The third index of association was the correlation between best-fit waves. However, interpersonal behaviors over time may not be best described by a single wave. Indeed, in the present study, best-fit waves explained an average of 12% of the variance in the detrended behavioral time series data. The failure of interpersonal behavior to vary sinusoidally over time may explain negative results in the tests of the association of best-fit waves. The entrainment between two interactants’ interpersonal behaviors may be better investigated across a number of wavelengths, such as in cross-spectral analysis (Sadler et al., 2009). In such an analysis, the wavelengths across which interactants’ behaviors are compared are not determined by the degree to which waves of that length account for variance in those behaviors. Rather, the interactants’ behaviors, as approximated by waves, are compared across the same wavelengths. Those comparisons are then weighted depending on the degree to which waves of those lengths account 56 for variance in each individuals’ behaviors. This allows for a more accurate measure of interpersonal entrainment in behavioral time series which are better explained by two or more waves. Overall, tests of the effects of interpersonal complementarity over time may differ depending on the nature of the association in question and depending on the nature of the method in which that question is investigated. Potential Differences between Communal and Agentic Dimensions Differences in results between warmth and dominance complementarity may be attributable to differences between communal and agentic dimensions. Interpersonal anxiety in interpersonal situations can differ depending on the nature of the individual’s motivation. Fulfilling communal motives is thought to preserve security whereas fulfilling agentic motives is thought to preserve self-esteem. These different aspects of anxiety may function differently in relation to interpersonal complementarity. For example, security and self-esteem might have different physiological correlates. Consequently, deviation from dominance complementarity may be more closely associated with a different measure of psychophysiological reactivity. Alternatively, the affective consequences of deviation from dominance complementarity may manifest at different time intervals, explaining why previous research has found an association between dominance complementarity and relational satisfaction over long time intervals (Tracey, Ryan & Jaschik-Herman, 2001), but not brief time intervals (Markey, Lowmaster, & Eichler, 2010). Dominance complementarity further differs from warmth complementarity in that it concerns dissimilarity between interactants whereas warmth complementarity concerns similarity between interactants. Affective consequences of deviation from dominance complementarity might differ depending on whether both individuals are dominant or both are submissive. Indeed, 57 the former case might result when two individuals are expressing their self-esteem. This raises the additional question as to whether the effects of deviation from warmth complementarity differ depending on whether the other person is too warm or too cold. Moderating effects of Interpersonal Sensitivities The present study examined the across-dyad effect of interpersonal sensitivities on the within-dyad association between confederate behaviors and participant skin conductance. The main effect of this association failed to reach significance. However, the findings suggest that the effect of the confederate’s warmth on the participant’s physiological reactivity depend on the interaction between self-reported interpersonal sensitivity to warmth and participant gender (i.e., same-sex versus different-sex dyad status). A similar effect was not found for sensitivity to dominance. This finding may suggest interpersonal sensitivities predict the effect of others’ behaviors on an individual’s skin conductance reactions in same-sex dyads but not different-sex dyads. Additionally, males may more readily characterize their sensitivities in terms of those behaviors which result in physiological affective reactions. Alternatively, given the number of moderating regressions, these effects may be due to type I error. Conclusion Interpersonal complementarity manifests across temporal intervals of differing durations including broad interpersonal tendencies (e.g., Tracey, Ryan & Jaschik-Herman, 2001), interpersonal behaviors across interactions (e.g., Strong et al., 1988), and interpersonal behaviors within interactions (e.g., Sadler et al., 2009). Extant research suggests that deviation from complementarity is associated with distress across relationships (e.g., Tracey, Ryan & JaschikHerman, 2001) and across interactions (Markey, Lowmaster, & Eichler, 2010). The present study provides preliminary evidence regarding the association between interpersonal processes, 58 anxiety, and physiological arousal across and within interactions. Specifically, participant’s perceptions of their own warmth and the confederates’ warmth were associated with how much the participant reported liked the confederate and how much negative affect the participant reported experiencing. Observer and participant perception of the participant’s warm behaviors trended towards significant association with overall changes in skin conductance and observer report of the participant’s warm behaviors associated was significantly associated with skin conductance response within the interaction. Total deviation from warmth complementarity within interactions increased negative affect across interactions. Similar effects were not identified for deviation from dominance complementarity. Overall, this exploratory study provided partial support for the hypothesis that interpersonal anxiety should be associated with cold behavior and deviations from warmth complementarity. Important lessons from this exploratory study point to future direction for further analyses of the study data. First, in this study, I compared the associations of best-fit waves. As 2 evidenced by the low mean R between the detrended time series and the best-fit wave time series, single waves may not best capture the variance in interpersonal behaviors over time. Cross-spectral analysis allows for the comparison of time series data as modeled across many waves of differing wavelengths. While this approach adds some level of complexity in decisionmaking (such as how many wavelengths to consider and of what magnitude), such an approach would allow for a more direct comparison of time series data that is better described by two or more waves. In future analyses I plan to compare the association between time series data across multiple wavelengths in this type of spectral analysis. Second, in this study, I analyzed skin conductance data. While skin conductance provides one physiological measure of affect, I also collected data on EMG and heart-rate during the interactions. EMG and heart-rate index different aspects of the participants’ psychophysiological affective reactions. In future analyses, I plan to 59 compare the association between interpersonal behaviors and these measures of physiological response within dyads. This exploratory study also provided new directions for future studies on the association between interpersonal behaviors and affects across brief time intervals. First, the present study was limited by number of participants, resulting in several inconclusive findings where the effects were sizable, but not significant. Future studies should collect a greater number of participants to enable more power in testing these hypotheses. Second, the present study was also limited by the number of confederates, resulting in demographic confounds and limited generalizability. Future studies should include multiple confederates of both genders to discriminate the effects of gender from same-sex versus different-sex dyad, as well as increase the degree to which study results can be generalized. Third, more dramatic shifts in confederate behavior may increase the power in testing within-dyad hypotheses. In future studies, the confederate should be trained to vary his behavior more dramatically. Careful and more thorough training may enable confederates to more dramatically vary their behavior without being identified as actors. Finally, the present study was limited by the operationalization of anxiety. Future studies should examine the effects of other manifestations of anxiety including EMG, heart rate, cortisol, and/or novel measurements of participant affect within the interaction. 60 APPENDICES 61 Appendix A Tables 62 Table 1. Descriptive Statistics and Cronbach’s Alphas Variable Mean Standard Cronbach’s Deviation Alpha Observer-coded Behaviors Overall Mean Warmth 113.46 76.79 .82 Participant Mean Warmth 160.75 56.92 .79 Confederate Mean Warmth 70.37 63.43 .73 Overall Mean Dominance 27.49 113.17 .86 Participant Mean Dominance -36.81 132.24 .83 Confederate Mean Dominance 100.94 109.58 .82 SBI Dominance 3.07 0.71 .79 SBI Submissiveness 2.32 0.57 .70 SBI Warmth 3.29 0.70 .77 SBI Coldness 1.71 0.57 .73 SBI Warmth Vector 1.57 0.76 SBI Coldness Vector 0.75 0.97 SBI Dominance 3.00 0.54 .70 SBI Submissiveness 2.52 0.56 .71 SBI Warmth 2.81 0.70 .78 SBI Coldness 2.14 0.63 .69 SBI Warmth Vector 0.67 0.99 SBI Dominance Vector 0.47 0.79 SBI Dominance 3.59 0.63 .81 SBI Submissiveness 3.21 0.57 .75 SBI Warmth 4.30 0.59 .80 SBI Coldness 2.70 0.52 .68 Self-Reported Interpersonal Behaviors Participant-Reported Confederate Interpersonal Behaviors Self-Reported Interpersonal Traits (SBI) 63 Table 1 (cont’d) SBI Warmth Vector 1.70 0.83 SBI Coldness Vector 0.38 0.89 (PA) Assured-Dominant 2.42 0.77 .76 (BC) Arrogant-Calculating 1.96 0.54 .52 (DE) Cold-Hearted 2.29 0.58 .36 (FG) Aloof-Introverted 2.84 0.87 .75 (HI) Unassured-Submissive 3.26 0.56 .23 (JK) Unassuming-Ingenuous 368 0.51 .19 (LM) Warm-Agreeable 4.28 0.53 .70 (NO) Gregarious-Extraverted 3.40 0.88 .73 IPIP-IPC Warmth Vector 0.18 0.72 IPIP-IPC Dominance Vector 0.10 0.80 (PA) Domineering 2.68 2.00 .32 (BC) Vindictive 2.51 2.20 .61 (DE) Cold 3.75 3.05 .76 (FG) Socially Avoidant 4.46 3.78 .85 (HI) Nonassertive 6.35 3.75 .80 (JK) Exploitable 5.92 3.36 .76 (LM) Overly-Nurturant 6.31 3.57 .73 (NO) Intrusive 3.31 3.10 .77 IIP Warm Problems Vector -0.21 0.16 IIP Dominant Problems Vector 0.06 0.15 (PA) Sensitive to Control 5.75 1.01 .79 (BC) Sensitive to Antagonism 6.05 1.08 .83 (DE) Sensitive to Remoteness 4.50 1.19 .83 (FG) Sensitive to Timidity 3.72 0.94 .75 Self-Reported Interpersonal Traits (IPIP-IPC) Self-Reported Interpersonal Problems Self-Reported Interpersonal Sensitivities 64 Table 1 (cont’d) (HI) Sensitive to Passivity 4.03 1.26 .86 (JK) Sensitive to Dependence 3.51 1.26 .75 (LM) Sensitive to Affection 3.11 1.07 .75 (NO) Sensitive to Attention-Seeking 4.72 1.00 .73 ISC Sensitive to Warmth Vector 0.09 0.54 ISC Sensitive to Dominance Vector 0.03 0.62 Measure of Liking the Confederate 3.77 1.22 Finding the Confederate Attractive 3.03 1.30 Enjoying the Task 4.23 1.38 Pre-Interaction Positive Affect 3.08 0.83 .90 Pre-Interaction Negative Affect 1.69 0.59 .86 Post-Interaction Positive Affect 2.96 0.89 .92 Post-Interaction Negative Affect 1.37 0.51 .90 Outcome Measures 65 .92 Table 2. Bivariate Correlations 1 2 3 4 5 6 1. SBI Trait WRM 2. SBI Trait DOM .10 3. SBI Participant State WRM .30* .19 4. SBI Participant State DOM .15 .41** .37** 5. SBI Confederate State WRM -.01 -.02 .64** .12 6. SBI Confederate State DOM .09 -.20t .02 -.22t .06 7. Observer-coded Participant WRM .23t .01 .38** .14 .24t .13 8. Observer-coded Participant DOM .06 .13 .03 .35** -.10 -.10 9. Observer-coded Confed. WRM -.15 -.04 .19 -.14 .35** .21t 10. Observer-coded Confed. DOM -.01 -.19 -.09 -.37** .11 .23t 11. ISC WRM Vector -.31** -.12 -.19 -.28t .11 -.17 12. ISC DOM Vector. .13 -.14 .05 -.10 -.05 .01 13. ISC Elevation .07 .08 .19 .09 .13 -.07 14. IIP WRM Vector .50** .10 .08 .10 -.22t -.01 15. IIP DOM Vector -.24t .60** -.06 .30* -.17 -.15 16. IIP Elevation -.02 -.41** -.18 -.22t -.10 .18 17. Average Positive Affect .05 .37** .29* .48* .07 -.26* 18. Residual Positive Affect .23t .02 .27 .37 .18 .07 19. Average Negative Affect -.16 -.22t -.27* -.22t -.25* .04 20. Residual Negative Affect -.02 .03 -.34** -.02 -.25* .08 21. Average Skin Conductance -.07 .05 .24t .02 .25t -.11 22. Residual Skin Conductance -.18 .02 -.14 -.11 -.07 -.29* 23. Liking the Confederate .20t -.07 .48* .08 .61** .25* 24. Finding the Confed. Attractive .13 .08 .08 -.04 .27* .20 25. Enjoying the Task .34** .01 .40* .13 .43** .14 66 Table 2 (cont’d) 7 8 9 10 11 12 13 8. Observer-coded Participant DOM .17 9. Observer-coded Confed. WRM .19 -.20 10. Observer-coded Confed. DOM .02 -.62** .16 11. ISC WRM Vector -.12 -.19 .31* .20 12. ISC DOM Vector. .05 -.01 .15 -.06 -.10 13. ISC Elevation .11 .03 .16 .13 .18 .03 14. IIP WRM Vector .05 .09 -.22t -.09 -.37** .13 .05 15. IIP DOM Vector -.21 .06 -.05 -.08 -.02 -.43** -.03 16. IIP Elevation .13 -.24t .18 .32* .06 .26* .28* 17. Average Positive Affect .21 .20 .03 -.10 .06 .06 .29* 18. Residual Positive Affect .06 .03 -.05 -.06 -.14 .04 .01 19. Average Negative Affect -.10 -.19 .24t .18 .13 .06 .25* 20. Residual Negative Affect -.00 -.11 .03 .25* -.15 -.22t .08 21. Average Skin Conductance .08 .10 .18 -.17 -.13 -.12 .13 22. Residual Skin Conductance -.26t -.07 -.10 -.14 .23t .05 -.12 23. Liking the Confederate .17 .14 .10 .05 -.13 -.03 .01 24. Finding the Confed. Attractive -.02 .10 .08 .01 -.14 -.15 -.09 25. Enjoying the Task .15 .17 .19 .14 -.03 .05 .24* 67 Table 2 (cont’d) 14 15 16 17 18 19 20 15. IIP DOM Vector -.15 16. IIP Elevation .17 -.37** 17. Average Positive Affect -.03 .22t -.08 18. Residual Positive Affect .06 -.03 .21t .29* 19. Average Negative Affect -.01 -.05 .47** .12 -.08 20. Residual Negative Affect .11 .19 .22t -.12 -.12 .36** 21. Average Skin Conductance -.19 .03 .04 .08 -.03 -.07 -.07 22. Residual Skin Conductance -.17 .01 -.01 .09 .02 .07 -.15 23. Liking the Confederate .01 -.18 -.08 .08 .27* -.14 -.26* 24. Finding the Confed. Attractive -.05 -.05 -.21t -.22t -.03 -.13 -.05 25. Enjoying the Task .13 -.01 -.05 .21t .24 -.22t -.13 68 Table 2 (cont’d) 21 22 23 24 22. Residual Skin Conductance .19 23. Liking the Confederate .15 -.01 24. Finding the Confed. Attractive .01 -.01 .49** 25. Enjoying the Task .15 -.19 .56** .30* Note. WRM = Warmth, DOM = Dominance, Confed. = Confederate; ** = p < .01, * = p < .05, t = p < .10 69 Table 3. Summary of Hypothesis Tests Hypothesis Operationalization 1. Observer-coded interpersonal behaviors will be associated with participant-rated interpersonal behaviors 2. Interpersonal traits will be associated with interpersonal sensitivities 3a. Interpersonal coldness will be associated with anxiety during the interaction (Across Dyads) Finding Observer-coded participant warmth will correlate with participant self-reported warmth Observer-coded participant dominance will correlate with participant self-reported dominance Observer-coded confederate warmth will correlate with participant-rated confederate warmth Observer-coded confederate dominance will correlate with participant-rated confederate dominance r = .38 p < .01 SBI trait warmth will correlate negatively with ISC sensitivity to warmth SBI trait dominance will correlate negatively with ISC sensitivity to dominance IIP warm problems will correlate negatively with ISC sensitivity to warmth IIP dominant problems will correlate negatively with ISC sensitivity to dominance Liking the confederate will correlate with mean observer-coded confederate warm behaviors Liking the confederate will correlate with mean observer-coded participant warm behaviors Liking the confederate will correlate with participant-rated confederate warm behaviors Liking the confederate will correlate with participant self-reported warm behaviors Negative affect will be associated with mean observer-coded participant warm behaviors Negative affect will be associated with mean observer-coded confederate warm behaviors Negative affect will be associated with participant self-reported warm behaviors Negative affect will be associated with participant-rated confederate warm behaviors Skin conductance will be associated mean observer-coded confederate warmth 70 Hypothesis Supported Yes r = .35 p < .01 Yes r = .35 p < .01 Yes r = .21 p < .10 Inconclusive r = -.31 p = .01 r = -.14 p = .26 r = -.34 p < .01 r = -.43 p < .01 r = .10 p = .46 Yes r = .19 p = .14 No r = .61 p < .01 r = .48 p < .01 r = -.00 p = .98 r = .02 p = .82 r = -.34 p < .01 r = -.25 p < .05 r = -.10 p = .49 Yes No Yes Yes No Yes No No Yes Yes No Table 3 (cont’d) 3b. Interpersonal coldness will be associated with anxiety during the interaction (Within Dyads) Skin conductance will be associated with observer-coded participant warmth across dyads Skin conductance will be associated with participant-rated confederate warm behaviors Skin conductance will be associated with self-reported warm behaviors The mean correlation between momentary skin conductance and momentary confederate warmth will be significantly less than zero More than 50% of cases will exhibit this association The mean correlation between momentary skin conductance and momentary confederate warmth will be significantly less than zero after linear detrending More than 50% of cases will exhibit this association The mean correlation between momentary skin conductance and momentary confederate warmth best-fit waves will be significantly less than zero More than 50% of cases will exhibit this association The mean correlation between momentary skin conductance and momentary participant warmth will be significantly less than zero More than 50% of cases will exhibit this association The mean correlation between momentary skin conductance and momentary participant warmth will be significantly less than zero after linear detrending More than 50% of cases will exhibit this association The mean correlation between momentary skin conductance and momentary participant warmth best-fit waves will be significantly less than zero More than 50% of cases will exhibit this association 71 r = -.26 p < .10 Inconclusive r = -.07 p = .61 No r = -.24 p < .10 M = 0.00 t = 0.09 p = .93 Inconclusive 50% p = .42 M = -0.03 t = -1.07 p = .29 No 57% P = .35 M = -.04 t = -0.85 p = .40 No 44% p = .42 M = -0.06 t = -1.76 p < .10 No 64% p < .05 M = -0.06 t = -2.41 p < .05 Yes 64% p < .05 M = -.07 t = -0.98 p = .33 Yes 58% p = .29 No No No No Inconclusive Yes No Table 3 (cont’d) 4a. Warmth complementarity will be associated with anxiety during the interaction (Within Dyads) 4b. Warmth complementarity will be associated with anxiety during the interaction (Across Dyads) The mean partial correlation between momentary skin conductance and deviation from warmth complementarity controlling for each individual’s warmth will be significantly greater than zero M = .03 t = 2.01 p < .05 Yes The mean partial correlation between momentary skin conductance and deviation from warmth complementarity controlling for each individual’s warmth will be significantly greater than zero after removing linear trends The mean correlation between momentary skin conductance and deviation from warmth complementarity best-fit waves will be significantly greater than zero Mean deviation from observer-coded warmth complementarity will increment each individual’s mean warmth and baseline skin conductance in predicting skin conductance during the interaction Mean deviation from observer-coded warmth complementarity will increment each individual’s mean warmth and baseline skin conductance in predicting skin conductance during the interaction after linear detrending Mean deviation of best-fit wave warmth complementarity waves will increment each individual’s mean warmth and baseline skin conductance in predicting skin conductance during the interaction Deviation from participant-rated warmth complementarity will increment participantrated warm behaviors and mean baseline skin conductance in predicting skin conductance during the interaction Mean deviation from observer-coded warmth complementarity will increment each individual’s mean warmth in predicting participant-ratings of liking the confederate Mean deviation from observer-coded warmth complementarity will increment each individual’s mean warmth in predicting participant-ratings of liking the confederate after linear detrending M = .01 t = 0.69 p = .50 No M = .03 t = 0.65 p = .52 No Beta = -.14 t = -1.38 p = .17 No Beta = -.13 t = -2.72 p < .01 No Beta = .09 t = 1.81 p < .10 No Beta = -.18 t = -0.74 p = .46 No Beta = -.37 t = -1.33 p = .19 No Beta = -01 t = -0.03 p = .97 No 72 Table 3 (cont’d) 5a. Dominance complementarity will be associated with anxiety during the interaction (Within Dyads) Mean deviation from observer-coded warmth complementarity best-fit waves will increment each individual’s mean warmth in predicting participant-ratings of liking the confederate Deviation from participant-rated warmth complementarity will increment participant-rated warm behaviors in predicting participant-ratings of liking the confederate Mean deviation from observer-coded warmth complementarity will increment each individual’s mean warmth and baseline negative affect in predicting negative affect during the interaction Mean deviation from observer-coded warmth complementarity will increment each individual’s mean warmth and baseline negative affect in predicting negative affect during the interaction after detrending Mean deviation of best-fit wave warmth complementarity waves will increment each individual’s mean warmth and baseline negative affect in predicting negative affect during the interaction Deviation from participant-rated warmth complementarity will increment participant-rated warm behaviors and mean baseline negative affect in predicting negative affect during the interaction The mean partial correlation between momentary skin conductance and deviation from dominance complementarity controlling for each individual’s dominance will be significantly greater than zero The mean partial correlation between momentary skin conductance and deviation from dominance complementarity controlling for each individual’s dominance will be significantly greater than zero after removing linear trends The mean correlation between momentary skin conductance and deviation from dominance complementarity best-fit waves will be significantly greater than zero 73 Beta = .01 t = 0.07 p = .95 No Beta = .29 t = 0.91 p = .37 No Beta = .20 t = 0.94 p = .35 No Beta = .21 t = 2.01 p < .05 Yes Beta = -.08 t = -.073 p = .47 No Beta = -.33 t = -1.20 p = .24 No M = -0.01 t = -0.86 p = .40 No M = -0.00 t = -0.17 p = .87 No M = -.02 t = -1.17 p = .25 No Table 3 (cont’d) 5b. Dominance complementarity will be associated with anxiety during the interaction (Across Dyads) Mean deviation from observer-coded dominance complementarity will increment each individual’s mean dominance and baseline skin conductance in predicting skin conductance during the interaction Mean deviation from observer-coded dominance complementarity will increment each individual’s mean dominance and baseline skin conductance in predicting skin conductance during the interaction after detrending Mean deviation of best-fit wave dominance complementarity waves will increment each individual’s mean dominance and baseline skin conductance in predicting skin conductance during the interaction Deviation from participant-rated dominance complementarity will increment participantrated dominant behaviors and mean baseline skin conductance in predicting skin conductance during the interaction Mean deviation from observer-coded dominance complementarity will increment each individual’s mean dominance in predicting participant-ratings of liking the confederate Mean deviation from observer-coded dominance complementarity will increment each individual’s mean dominance in predicting participant-ratings of liking the confederate after linear detrending Mean deviation from observer-coded dominance complementarity best-fit waves will increment each individual’s mean dominance in predicting participant-ratings of liking the confederate Deviation from participant-rated dominance complementarity will increment participantrated warm behaviors in predicting participant-ratings of liking the confederate 74 Beta = -.02 t = -0.24 p = .81 No Beta = -.05 t = -1.06 p = .92 No Beta = -.11 t = -0.88 p = -.38 No Beta = .26 t = 1.86 p < .10 Inconlusive Beta = -.07 t = -0.36 p = .72 No Beta = -.09 t = -0.71 p = .72 No Beta = .16 t = 1.18 p = .24 No Beta = .11 t = 0.31 p = .76 No Table 3 (cont’d) 6. Interpersonal sensitivities will moderate the effect of interpersonal behaviors on anxiety Mean deviation from observer-coded dominance complementarity will increment each individual’s mean dominance and baseline negative affect in predicting negative affect during the interaction Mean deviation from observer-coded dominance complementarity will increment each individual’s mean dominance and baseline negative affect in predicting negative affect during the interaction after detrending Mean deviation of best-fit wave dominance complementarity waves will increment each individual’s mean dominance and baseline negative affect in predicting negative affect during the interaction Deviation from participant-rated dominance complementarity will increment participant-rated warm behaviors and mean baseline negative affect in predicting negative affect during the interaction Sensitivity to warmth will correlate with the within-dyad association between momentary confederate warmth and momentary skin conductance Sensitivity to warmth will correlate with the within-dyad association between momentary confederate warmth and momentary skin conductance after removing linear trends Sensitivity to warmth will correlate with the within-dyad association between momentary confederate warmth and momentary skin conductance best-fit waves Sensitivity to dominance will correlate with the within-dyad association between momentary confederate dominance and momentary skin conductance Sensitivity to dominance will correlate with the within-dyad association between momentary confederate dominance and momentary skin conductance after removing linear trends Sensitivity to dominance will correlate with the within-dyad association between momentary confederate dominance and momentary skin conductance best-fit waves 75 Beta = .04 t = 0.28 p = .78 No Beta = .14 t = 1.40 p = .17 No Beta = -.04 t = -0.41 p = .69 No Beta = -.04 t = -0.13 p = .90 No r = -.08 p = .58 No r = .001 p = .997 No r = .08 p = .58 No r = -.14 p = .31 No r = -.03 p =.85 No r = -.16 p = .24 No Table 4. Regression Summaries Dependent Block Predictor Variables Variable Interaction 1 Intercept Skin Confederate Warmth Conductance Participant Warmth Baseline Skin Conductance 2 Deviation from Warmth Complementarity Interaction 1 Intercept Skin Confederate Warmth Conductance Participant Warmth Baseline Skin Conductance 2 Detrended Deviation from Warmth Complementarity Interaction 1 Intercept Skin Confederate Warmth Conductance Participant Warmth Baseline Skin Conductance 2 Deviation from Bestfit Wave Warmth Complementarity Interaction 1 Intercept Skin SBI Confederate Conductance Warmth SBI Participant Warmth Baseline Skin Conductance 2 SBI Deviation from Warmth Complementarity 2 R Change F change p value .89 .00 -.13 -.00 .95 Predictor p value .00 .17 .98 .00 -.14 .17 .00 .17 .00 .16 .13 .00 .89 .00 -.07 -.07 .95 -.13 .01 .02 .01 .36 .96 .17 .00 .89 -.00 -.07 .98 .09 .08 .01 .08 .03 .49 .87 .00 -.22 .11 .49 .93 .00 -.18 .46 .00 .46 Beta 76 Table 4 (cont’d) Liking the Confederate 1 2 Liking the Confederate 1 2 Liking the Confederate 1 2 Liking the Confederate 1 2 PostInteraction Negative Affect 1 2 Intercept Confederate Warmth Participant Warmth Deviation from Warmth Complementarity Intercept Confederate Warmth Participant Warmth Detrended Deviation from Warmth Complementarity Intercept Confederate Warmth Participant Warmth Deviation from Bestfit Wave Warmth Complementarity Intercept SBI Confederate Warmth SBI Participant Warmth SBI Deviation from Warmth Complementarity Intercept Confederate Warmth Participant Warmth Pre-Interaction Negative Affect Deviation from Warmth Complementarity .03 .36 -.23 .37 -.36 .00 .37 .08 .19 .03 .19 .03 .36 .06 .16 -.01 .00 .67 .23 .97 .00 .97 .03 .41 .03 .17 .01 .00 .85 .22 .95 .00 .95 .38 .00 -.12 .00 .70 .87 .04 .29 .37 .01 .37 .47 .00 .18 -.12 .68 .43 .36 .44 .00 .20 .35 .01 .35 77 Table 4 (cont’d) PostInteraction Negative Affect 1 2 PostInteraction Negative Affect 1 2 PostInteraction Negative Affect 1 2 Intercept Confederate Warmth Participant Warmth Pre-Interaction Negative Affect Detrended Deviation from Warmth Complementarity Intercept Confederate Warmth Participant Warmth Pre-Interaction Negative Affect Deviation from Bestfit Wave Warmth Complementarity Intercept SBI Confederate Warmth SBI Participant Warmth Pre-Interaction Negative Affect SBI Deviation from Warmth Complementarity .47 .00 .11 -.01 .69 .81 .29 .91 .00 .21 .05 .04 .05 .45 .00 .01 -.02 .67 .08 .95 .82 .00 -.08 .47 .01 .47 .00 .21 .55 .00 -.45 .08 .78 .65 .00 -.33 .24 .01 .24 78 Table 4 (cont’d) Interaction 1 Skin Conductance 2 Interaction 1 Skin Conductance 2 Interaction 1 Skin Conductance 2 Intercept Confederate Dominance Participant Dominance Baseline Skin Conductance Deviation from Dominance Complementarity Intercept Confederate Dominance Participant Dominance Baseline Skin Conductance Detrended Deviation from Dominance Complementarity Intercept Confederate Dominance Participant Dominance Baseline Skin Conductance Deviation from Bestfit Wave Dominance Complementarity -.07 .04 .38 -.05 .60 .94 .00 -.04 .53 .01 .53 .89 .00 -.09 .03 .14 -.08 .20 .93 .00 -.05 .29 .00 .29 .89 .00 -.11 .02 .10 -.09 .17 .93 .00 -.05 .38 .00 .38 79 .89 .00 Table 4 (cont’d) Interaction 1 Skin Conductance 2 Liking the Confederate 1 2 Liking the Confederate 1 2 Liking the Confederate 1 2 Intercept SBI Confederate Dominance SBI Participant Dominance Baseline Skin Conductance SBI Deviation from Dominance Complementarity Intercept Confederate Dominance Participant Dominance Deviation from Dominance Complementarity Intercept Confederate Dominance Participant Dominance Detrended Deviation from Dominance Complementarity Intercept Confederate Dominance Participant Dominance Deviation from Bestfit Wave Dominance Complementarity -.32 .04 .01 -.27 .03 .93 .00 .26 .07 .01 .07 .05 .21 .28 .00 .21 .34 .16 -.07 .72 .00 .72 .05 .21 .24 .00 .14 .29 .08 -.09 .48 .01 .48 .06 .18 .29 .00 .10 .33 .05 .16 .24 .02 .24 80 .88 .00 Table 4 (cont’d) Liking the Confederate 1 2 PostInteraction Negative Affect 1 2 PostInteraction Negative Affect 1 2 Intercept SBI Confederate Dominance SBI Participant Dominance SBI Deviation from Dominance Complementarity Intercept Confederate Dominance Participant Dominance Pre-Interaction Negative Affect Deviation from Dominance Complementarity Intercept Confederate Dominance Participant Dominance Pre-Interaction Negative Affect Detrended Deviation from Dominance Complementarity .20 .05 .86 .11 .76 .00 .76 .67 .24 .51 .00 .19 .02 .90 .68 .00 .04 .78 .00 .78 .59 .08 .51 .00 .21 .05 .66 .68 .00 .05 81 .00 .50 .08 .05 .63 .00 .63 Table 4 (cont’d) PostInteraction Negative Affect 1 2 PostInteraction Negative Affect 1 2 Participant Warmth 1 2 Participant Dominance 1 2 Intercept Confederate Dominance Participant Dominance Pre-Interaction Negative Affect Deviation from Bestfit Wave Dominance Complementarity Intercept SBI Confederate Dominance SBI Participant Dominance Pre-Interaction Negative Affect SBI Deviation from Dominance Complementarity Intercept Gender SBI Participant Warmth Gender x Warmth Interaction Intercept Gender SBI Participant Dominance Gender x Dominance Interaction .20 .15 .11 .02 .89 .70 .00 -.04 .69 .00 .69 .49 .00 .09 .04 .71 .03 .92 .70 .00 -.04 .90 .00 .90 .15 .01 -.02 .39 .00 .86 .00 .01 .97 .00 .97 .05 .91 .01 .12 .02 -.01 .31 .23 .06 .05 .06 82 .53 .00 Table 4 (cont’d) Sensitivity to Warmth 1 2 Sensitivity to Warmth 1 2 Sensitivity to Dominance 1 2 Sensitivity to Dominance 1 2 Liking the Confederate 1 2 Intercept Gender SBI Trait Warmth Gender x Trait Warmth Interaction Intercept Gender IIP Warm Problems Gender x Warm Probelms Interaction Intercept Gender SBI Trait Dominance Gender x Trait Dominance Interaction Intercept Gender IIP Dominant Problems Gender x Trait Dominance Interaction Intercept Gender Confederate Warmth Gender x Confederate Warmth Interaction .10 .03 -.00 -.28 .20 .26 .97 .03 .10 .04 .10 .16 .00 -.04 -.41 .20 .11 .75 .00 .09 .04 .09 .18 .25 .20 .66 .04 .29 -.14 -.17 -.06 .00 .66 .19 .00 -.08 -.46 .12 .49 .00 -.12 .30 .01 .30 .03 .41 .14 .16 .25 .00 .26 .23 .06 .06 .06 83 Table 4 (cont’d) Liking the Confederate 1 2 3 Liking the Confederate 1 2 3 Within-dyad 1 Correlation between Detrended Warmth and 2 Detrended Skin 3 Conductance Intercept Confederate Warmth Participant Warmth Gender Detrended Deviation from Warmth Complementarity Gender x Detrended Deviation from Warmth Complementarity Interaction Intercept Confederate Dominance Participant Dominance Gender Detrended Deviation from Dominance Complementarity Gender x Detrended Deviation from Dominance Complementarity Interaction Intercept Confederate Warmth Participant Warmth Gender Sensitivity to Warmth Gender x Sensitivity to Warmth Interaction .05 .37 .09 .16 .10 -.01 .00 .53 .19 .40 .94 .00 .84 -.36 .00 .13 .00 .08 .20 .17 .00 .30 .36 .03 .14 -.12 .25 .34 .00 .55 -.27 .06 .06 .06 .16 .89 .51 .98 .42 .01 .88 -.02 .09 .00 .12 .00 .78 .36 .02 .11 .02 84 Appendix B Figures 85 Figure 1. Interpersonal Circle 86 Appendix C Materials 87 MEASURES OF LIKING a. How much do you like the other person? b. How much would you like to see the other person again? c. How much would you like to be friends with the other person? d. How attractive do you find the other person? e. How enjoyable did you find this task? 88 PROMPTING QUESTIONS a. What would constitute a ‘perfect’ day for you? b. If you could wake up tomorrow having gained any one quality or ability, what would it be? c. If you could change anything about the way you were raised, what would it be? d. Is there something that you’ve dreamed of doing for a long time? e. If you knew what in one year you would die suddenly, would you change anything about the way you are living now? Why? f. What is your most treasured memory? g. When is the last time you cried? 89 REQUIRED BEHAVIORS Friendly Behaviors Agree , nod (JK) Lean Towards,Smile (LM) Advise, Invite, Encourage (NO) Unfriendly Behaviors Criticize, disagree (BC) Lean Away, Ignore (DE) Avoid Eye Contact, self-efface (FG) Other Behaviors Interrupt, Ask the prompting question first (PA) A period of silence (HI) 90 REFERENCES 91 REFERENCES Aron, A., Melinat, E., Aron, E. N., Vallone, R. D., & Bator, R. J. (1997) The experimental generation of interpersonal closeness: A procedure and some preliminary findings. Personality and Social Psychology Bulletin, 23, 363-377. Adams, R. S., & Tracey, T. J. G. (2004). Three versions of the interpersonal adjective scales and their fit to the circumplex model. Assessment, 11(3), 263-270. Alden, L. E., Wiggins, J. S., & Pincus, A. L. (1990). Construction of circumplex scales for the inventory of interpersonal problems. Journal of Personality Assessment, 55(3-4), 521536. Ambady, N., Hallahan, M., & Rosenthal, R. (1995). On judging and being judged accurately in zero-acquaintance situations. Journal of Personality and Social Psychology, 69, 518–529. Ansell, E. B., Kurtz, J. E., & Markey, P. M. (2008). Gender differences in interpersonal complementarity within roommate dyads. Personality and Social Psychology Bulletin, 34(4), 502-512. Ansell, E. B., Thomas, K. M., Hopwood, C. J., Chaplin, T. M. (2012). Interpersonal Complementarity and Stress in a Parent-Adolescent Interaction Task. Talk presented at the Society for Personality Assessment Annual Meeting, Chicago, IL. Anusic, I., Schimmack, U., Pinkus, R. T., & Lockwood, P. (2009). The nature and structure of correlations among big five ratings: The halo-alpha-beta model. Journal of Personality and Social Psychology, 6, 1142-1156. Assor, A., Aronoff, J., & Messe, L. A. (1986). An experimental test of defensive processes in impression formation. Journal of Personality and Social Psychology, 50(3), 644-650. Back, M. D., Schmukle, S. C., & Egloff, B. (2009). Predicting actual behavior from the explicit and implicit self-concept of personality. Journal of Personality and Social Psychology, 97(3), 533-548. Barelds, D. P. H. (2005). Self and partner personality in intimate relationships. European Journal of Personality, 19, 501–518. Bluhm, C., Widiger, T. A., & Miele, G. M. (1990). Interpersonal complementarity and individual differences. Journal of Personality and Social Psychology, 58(3), 464-471. Borkenau, P. & Liebler, A. (1992). Trait inferences: Sources of validity at zero acquaintance. Journal of Personality and Social Psychology, 62(4), 645-657. 92 Borkenau, P., Mauer, N., Reimann, R., Spinath, F. M., & Angleitner, A. (2004). Thin slices of behavior as cues of personality and intelligence. Personality Processes and Individual Differences, 86(4), 599-614. Buss, D. M., & Craik, K. H. (1983a). Act prediction and the conceptual analysis of personality scales: Indices of act density, bipolarity, and extensity. Journal of Personality and Social Psychology, 45(5), 1081-1095. Byrne, D. (1971) The Attraction Paradigm, New York: Academic Press Cappella, J. N. (1996). Dynamic coordination of vocal and kinesic behavior in dyadic interaction: Methods, problems, and interpersonal outcomes. In J. H. Watt & C. A. VanLear (Eds.), Dynamic patterns in communication processes (pp. 353–386). Thousand Oaks, CA: Sage Carson, R. C. (1969). Interaction concepts of personality. Chicago: Aldine. Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd. Ed.). Hillsdale, NJ, England: Lawrence Erlbaum Associates, Inc. Dodge, K. A., & Crick, N. R. (1990). Social Information-Processing Bases of Aggressive Behavior in Children. Personality and Social Psychology Bulletin, 16(1), 8-22. Dodge, K. A., & Somberg, D. R. (1987). Hostile attributional biases among aggressive boys are exacerbated under conditions of threats to the self. Child Development, 58(1), 213-224. Dryer, D. C., & Horowitz, L. M. (1997). When do opposites attract? Interpersonal complementarity versus similarity. Journal of Personality and Social Psychology, 72(3), 592-603. Dryenforth, P. S., Kashy, D. A., Donnellan, M. B., Lucas, R. E. (2010). Predicting relationship and life satisfaction from personality in nationally representative samples from three countries: The relative importance of actor, partner, and similarity effects. Journal of Personality and Social Psychology, 99, 690-702. Fleeson, W., & Gallagher, P. (2009). The implications of big five standing for the distribution of trait manifestation in behavior: Fifteen experience-sampling studies and a meta-analysis. Journal of Personality and Social Psychology, 97(6), 1097-1114. Fournier, M. A., Moskowitz, D. S., & Zuroff, D. C. (2008). Integrating dispositions, signatures, and the interpersonal domain.Journal of Personality and Social Psychology, 94(3), 531545. Friedman, J.N.W., Oltmanns, T.F., & Turkheimer, E. (2007). Interpersonal perception and personality disorders: Utilization of a thin slice approach. Journal of Research in Personality, 41, 667-688. 93 Fritz, H. L., Nagurney, A. J., Helgeson, V. S. (2003). Social interactions and cardiovascular reactivity during problem disclosure among friends. Personality and Social Psychology Bulletin, 29, 713-725. Gallo, L. C., Smith, T. W., & Kircher, J. C. (2000). Cardiovascular and electrodermal responses to support and provocation: Interpersonal methods in the study of psychophysiological reactivity. Psychophysiology, 37, 289-301. Gifford, R., & O'Connor, B. (1987). The interpersonal circumplex as a behavior map. Journal of Personality and Social Psychology, 52(5), 1019-1026. Gormly, J. (1974). A comparison of predictions from consistency and affect theories for arousal during interpersonal disagreement. Journal of Personality and Social Psychology, 40(5), 658-663. Gottman, J. M. (1979). Detecting cyclicity in social interaction. Psychological Bulletin, 86(2), 338-348. Graziano, W. G., Bruce, J., Sheese, B. E., & Tobin, R. M. (2007). Attraction, personality, and prejudice: Liking none of the people most of the time. Interpersonal Relations and Group Processes, 93(4), 565-582. Gurtman, M. B. (1992). Construct validity of interpersonal personality measures: The interpersonal circumplex as a nomological net. Journal of Personality and Social Psychology, 63(1), 105-118. Gurtman, M. B. (1996). Interpersonal problems and the psychotherapy context: The construct validity of the Inventory of Interpersonal Problems. Psychological Assessment, 8, 241255. Hardy, J. D., & Smith, T. W. (1988). Cynical hostility and vulnerability to disease: Social support, life stress, and physiological response to conflict. Health Psychology, 7(5), 447459. Henry, W. P., Schacht, T. E., & Strupp, H. H. (1986). Structural analysis of social behavior: Application to a study of interpersonal process in differential psychotherapeutic outcome. Journal of Consulting and Clinical Psychology, 54(1), 27-31. Hill, C. R., & Safran, J. D. (1994) Assessing interpersonal schemas: Anticipated responses of significant others. Journal of Social and Clinical Psychology, 13(4), 366-379. Hopwood, C. J., Ansell, E. B., Pincus, A. L., Wright, A. G. C., Lukowitsky, M. R., & Roche, M. J. (2011). The circumplex structure of interpersonal sensitivities. Journal of Personality, 79(4), 707-740. 94 Hopwood, C. J., Pincus, A. L., DeMoor, R. M., & Koonce, E. A. (2008). Psychometric characteristics of the inventory of interpersonal problems-short circumplex (IIP-SC) with college students. Journal of Personality Assessment, 90(6). Horowitz, L. M., Alden, L. E., Wiggins, J. S., & Pincus, A. L. (2000) Inventory of Interpersonal Manual. Odessa, FL: The Psychological Corporation. Horowitz, L. M., Wilson, K. R., Turan, B., Zolotsev, P., Constantino, M. J., & Henderson, L. (2006). How interpersonal motives clarify the meaning of interpersonal behavior: A revised circumplex model. Personality and Social Psychology Review, 10(1), 67-86. Humbad, M. N., Donnellan, M. B., Iacono, W. G., McGue, M., & Burt, S. A. (2010). Is spousal similarity for personality a matter of convergence or selection? Personality and Individual Differences, 49, 827-830. Kiesler, D. J. (1983). The 1982 Interpersonal Circle: A taxonomy for complementarity in human transactions. Psychological Review, 90(3), 185-214. Kiesler, D. J. (1996). Contemporary interpersonal theory and research. New York: Wiley. Kiesler, D. J., Schmidt, J. A., & Wagner, C. C. (1997). A circumplex inventory of impact messages: An operational bridge between emotion and interpersonal behavior. In R. Plutchik & H. Contes (Eds.), Circumplex models of personality and emotions (pp. 221244). Washington, DC: American Psychological Association. Kivlighan, D. M., & Angelone, E. O. (1992). Interpersonal problems: Variables influencing participants’ perception of group climate. Journal of Counseling Psychology, 39(4), 468472. Kivlighan, D. M., Marsh-Angelone, M., & Angelone, E. O. (1994). Projection in group counseling: The relationship between members’ interpersonal problems and their perception of the group leader. Journal of Counseling Psychology, 41(1), 99-104. Krueger, J. & Clement, R. W. (1994). The truly false consensus effect: An ineradicable and egocentric bias in social perception. Journal of Personality and Social Psychology, 67(4), 596-610. Lackner, H. K., Goswami, N., Hinghofer-Szalkay, H., Papousek, I., Scharfetter, H., Furlan, Rafaello, & Schwaberger, G. (2010). Effects of stimuli on cardiovascular reactivity occuring at regular intervals during mental stress. Journal of Psychophysiology, 24(1), 48-60. Leary, T. (1957). Interpersonal diagnosis of personality: A functional theory and methodology for personality evaluation. New York: Ronaid Press. Lewin, K. (1936). Principles of topological psychology. New York, NY, US: McGraw-Hill. 95 Lizdek, I., Sadler, P., Woody, E., Ethier, N., & Malet, G. (2012). Capturing the stream of behavior: A computer-joystick method for coding interpersonal behavior continuously over time. Social Science Computer Review, 30, 513-521. Locke, K. D. (2000). Circumplex scales of interpersonal values: Reliability, validity, and applicability to interpersonal problems and personality disorders. Journal of Personality Assessment, 75(2), 249-267. Locke, K. D. (2005). Interpersonal problems and interpersonal expectations in everyday life. Journal of Social and Clinical Psychology, 24(7), 915-391. Locke, K. D., & Sadler, P. (2007). Self-efficacy, values, and complementarity in dyadic interactions: Integrating interpersonal and social-cognitive theory. Personality and Social Psychology Bulletin, 33(1), 94-109. Markey, P. M., Funder, D. C., & Ozer, D. J. (2003). Complementarity of interpersonal behaviors in dyadic interactions.Personality and Social Psychology Bulletin, 29(9), 1082-1090. Markey, P. M., & Kurtz, J. E. (2006). Increasing acquaintanceship and complementarity of behavioral styles and personality traits among college roommates. Personality and Social Psychology Bulletin, 32(7), 907-916. Markey, P., Lowmaster, S., & Eichler, W. (2010). A real-time assessment of interpersonal complementarity. Personal Relationships, 17(1), 13-25. Markey, P. M., & Markey, C. N. (2009). A brief assessment of the interpersonal circumplex: The IPIP-IPC. Assessment 16(4), 352-361. McCrae, R. R., Martin, T. A., Hrebickova, M., Urbanek, T., Boomsma, D. I., Willemsen,G., et al. (2008). Personality trait similarity between spouses in four cultures. Journal of Personality, 76, 1137–1163. Mischel, W. (1977). “The interaction of person and situation”. In Magnusson, D., & Endler, N.S., Personality at a crossroads: Current issues in interactional psychology (pp 333352). Hillsdale, New Jersey: Lawrence Erlbaum Associates. Moskowitz, D. S. (1994). Cross-situational generality and the interpersonal circumplex. Journal of Personality and Social Psychology, 66(5), 921-933. Moskowitz, D. S., Ho, M.-H. R., & Turcotte-Tremblay, A.M. (2007). Contextual influences on interpersonal complementarity. Personality and Social Psychology Bulletin, 33, 10511063. Newton, T. L., & Bane, C. M. H. (2001). Cardiovascular correlates of behavioral dominance and hostility during dyadic interaction. International Journal of Psychophysiology, 40, 33-46. 96 Newton, T. L., Watters, C. A., Philhower, C. L., & Weigel, R. A. (2005). Cardiovascular reactivity during dyadic social interaction: The roles of gender and dominance. International Journal of Psychophysiology, 57, 219-228. O’Connor, B. P., & Dyce, J. (1997). Interpersonal rigidity, hostility, and complementarity in musical bands. Journal of Personality and Social Psychology, 72(2), 362-372. Orford, J. (1986). The rules of interpersonal complementarity: Does hostility beget hostility and dominance, submission? Psychological Review, 93(3), 365-377. Oltmanns, T.F., Friedman, J.N., Fiedler, E.R., & Turkheimer, E. (2004). Perceptions of people with personality disorders based on thin slices of behavior. Journal of Research in Personality, 38, 216-229. Paunonen, S. V., & Ashton, M. C. (2001). Big Five Factors and Facets and the Prediction of Behavior. Journal of Personality and Social Psychology, 81(3), 524-539. Pincus, A. L., Lukowitsky, M. R., & Wright, A. G. C. (2010). The interpersonal nexus of personality and psychopathology. In Millon T., Krueger R. F. and Simonsen E. (Eds.), Contemporary directions in psychopathology: Scientific foundations of the DSM-V and ICD-11. New York, NY, US: Guilford Press. Pope, M. K., & Smith, T. W. (1991). Cortisol Excretion in High and Low Cynically Hostile Men. Psychosomatic Medicine, 53, 386-392. Prokasy, W. F., & Ebel, H. C. (1976). Three components of the classically conditioned GSR in human subjects. Journal of Experimental Psychology, 73, 247-256. Raffalovich, L. E. (1994). Detrending time series: A cautionary note. Sociological Methods and Research, 22, 492-519. Riemann, R., & Angleitner, A. (1993). Inferring interpersonal traits from behavior: Act prototypicality versus conceptual similarity of trait concepts. Journal of Personality and Social Psychology, 64(3), 356-364. Sadler, P., Ethier, N., Gunn, G. R., Duong, D., & Woody, E. (2009). Are we on the same wavelength? interpersonal complementarity as shared cyclical patterns during interactions. Journal of Personality and Social Psychology, 97(6), 1005-1020. Sadler, P., & Woody, E. (2003). Is who you are who you're talking to? Interpersonal style and complementarily in mixed-sex interactions. Journal of Personality and Social Psychology, 84(1), 80-96. Schade, N., Thomas, K. M., Nelson, A. B., Moran, T. P., Moser, J. S., & Hopwood, C. J. (2012). A pilot study of physiological anxiety responses to interpersonal behaviors. Poster 97 presented at 15th Annual Meeting of the Society for Interpersonal Theory and Research, Montreal, Quebec, Canada. Shechtman, N. & Horowitz, L. M. (2006). Interpersonal and non-interpersonal interactions, interpersonal motives, and the effects of frustrated motives. Personality and Social Psychology Bulletin, 32(8), 1126-1139. Smith, T. W., & Allred, K. D. (1989). Blood-pressure responses during social interaction in highand low-cynically hostile males. Journal of Behavioral Medicine, 12(2), 135-143. Smith, T. W., Gallo, L. C., & Ruiz, J. M. (2003). In Suls J., Wallston K. A. (Eds.), Toward a social psychophysiology of cardiovascular reactivity: Interpersonal concepts and methods in the study of stress and coronary disease. Malden: Blackwell Publishing. Smith, J. L., & Ruiz, J. M. (2007). Interpersonal orientation in context: Correlates and effects of interpersonal complementarity on subjective and cardiovascular experiences. Journal of Personality, 75(4), 679-708. Snyder, M, & Ickes, W. (1985). Personality and social behavior. In G. Lindzey and E. Aronson (Eds.), Handbook of social psychology: Third edition. Vol. 2 (pp. 883-947). New York: Random House. Sober, E. (2001). Venetian sea levels, british bread prices, and the principle of the common cause. The British Journal for the Philosophy of Science, 52, 331-346. Soldz, S., Budman, S., Demby, A., & Merry, J. (1995). A short form of the inventory of interpersonal problems circumplex scales. Assessment, 2(1), 53-63. Sullivan, H. S. (1953). The interpersonal theory of psychiatry. New York: Norton. Strong, S. R., & Hills, H. (1986). Interpersonal Communication Rating Scale. Richmond, VA: Virginia Commonwealth University. Strong, S. R., Hills, H. I., Kilmartin, C. T., DeVries, H., Lanier, K., Nelson, B. N., . . . Meyer, C. W., (1988). The dynamic relations among interpersonal behaviors: A test of complementarity and anticomplementarity. Journal of Personality and Social Psychology,54(5), 798-810. Stroud, L. R., Tanofsky-Kraff, M., Wilfey, D. E., & Salovey, P. (2000). The Yale interpersonal stressor (YIPS): Affective, physiological, and behavioral responses to a novel interpersonal rejection paradigm. Annals of Behavioral Medicine, 22(3), 204-213. Tasca, G. A., & McMullen, L. M. (1992). Interpersonal complementarity and antitheses within a stage model of psychotherapy. Psychotherapy: Theory, Research, Practice, Training, 29(4), 515-523. 98 Tracey, T. J. (1994). An examination of the complementarity of interpersonal behavior. Journal of Personality and Social Psychology, 67(5), 864-878. Tracey, T. J. G. (2004). Levels of interpersonal complementarity: A simplex representation. Personality and Social Psychology Bulletin, 30(9), 1211-1225. Tracey, T. J., Ryan, J. M., & Jaschik-Herman, B. (2001). Complementarity of interpersonal circumplex traits. Personality and Social Psychology Bulletin, 27(7), 786-797. Tracey, T. J. G., Sherry, P., & Albright, J. M. (1999). The interpersonal process of cognitive– behavioral therapy: An examination of complementarity over the course of treatment. Journal of Counseling Psychology, 46(1), 80-91. Tiedens, L. Z., & Fragale, A. R. (2003). Power moves: Complementarity in dominant and submissive nonverbal behavior. Journal of Personality and Social Psychology, 84(3), 558-568. Vandenberg, S. G. (1972). Assortative mating, or who marries whom? Behavior Genetics, 2, 127-158. Vella, E. J., & Friedman, B. H. (2009). Hostility and anger in: Cardiovascualr reactivity and recovery to mental arithmetic stress. International Journal of Psychophysiology, 72, 253259. Wagner, C. C., Kiesler, D. J., & Schmidt, J. A. (1995). Assessing the interpersonal transaction cycle: Convergence of action and reaction interpersonal circumplex measures. Journal of Personality and Social Psychology, 69, 938-949 Warner, Rebecca M. Spectral Analysis of Time-Series Data. New York, NY, US: Guilford Press, 1998. Watson, D., Clark, L. A., & Carey, G. (1988). Positive and negative affectivity and their relation to anxiety and depressive disorder. Journal of Abnormal Psychology, 97(3), 346-353. Wiggins, J. S. (1991). Agency and communion as conceptual coordinates for the understanding and measurement of interpersonal behavior. In W. M. Grove & D. Ciccetti (Eds.), Thinking clearly about psychology: Vol. 2. Personality and psychopathology (pp. 89-113). Minneapolis, MN: University of Minnesota Press. Wiggins, J. S. (2003). Paradigms of personality assessment. New York: Guilford Press. Wiggins, J. S., Phillips, N., & Trapnell, P. D. (1989). Circular reasoning about interpersonal behavior: Evidence concerning some untested assumptions underlying diagnostic classification. Journal of Personality and Social Psychology, 56, 296–305. 99 Wiggins, J. S., Trapnell, P., & Phillips, N. (1988). Psychometric and geometric characteristics of the revised interpersonal adjective scales (IAS—R). Multivariate Behavioral Research, 23(4), 517-530. Witvliet, C. V., & Vrana, S. R. (1995). Physiological responses as indices of affective dimensions. Psychophysiology, 32, 436-443. Wood, D., Harms, P., & Vazire, S. (2010). Perceiver effects as projective tests: What your perceptions of others say about you. Journal of Personality and Social Psychology, 99(1), 174-190. Yaughn, E., & Nowicki, S. (1999). Close relationships and complementary interpersonal styles among men and women. The Journal of Social Psychology, 139(4), 473-478. 100