THE THEORY OF POLITE SUPPORT SEEKING By Ashley Ann Hanna Edwards A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Communication–Doctor of Philosophy 2014 ABSTRACT A THEORY OF POLITE SOCIAL SUPPORT SEEKING By Ashley Ann Hanna Edwards This dissertation proposes a Theory of Polite Support Seeking (TOPSS). Social support is a ubiquitous part of interpersonal relationships, but most research focuses on the role of the support provider and support recipient. However, TOPSS asserts that a support seeker can influence the support process by utilizing more or less effective support seeking messages. Drawing from face theory (Goffman, 1967), politeness theory (Brown & Levinson, 1987), and the attribution-emotion-intention model (i.e., attribution theory of motivation; e.g., Weiner, 1980), TOPSS posits that the dimensions of support seeking messages differentially threaten the positive and negative face of potential support providers (PSP), resulting in differences in PSP affect that influence support outcomes. Beyond formally proposing the theory, this dissertation serves as an initial test of the theoretical model. Findings provide partial support for the theory, revealing that the directness of support seeking messages indirectly influences support provision by decreasing perceived threat to positive face thereby increasing willingness to provide support. In combination, the proposal of TOPSS and findings of the initial theory test provide implications for social support theory, research, and practice. Copyright by ASHLEY ANN HANNA EDWARDS 2014 This dissertation is dedicated my Grandma and Grandpa, Mary and Earl Schoenecker, who have never missed an opportunity to provide me with high quality social support. iv ACKNOWLEDGEMENTS This dissertation would not have been possible without Dr. Amanda Holmstrom, my advisor and most effective social support provider. After working together on the Dream Team for one year, you took a big risk by allowing me to be your first advisee. You aided my development, both as an academic and a person. With your support I transformed from an easily intimidated graduate student to a confident theorist and professor. I am so blessed to call you my mentor, but more blessed to call you my friend. I look forward to our continued collaboration. I am also grateful to my esteemed committee: To Dr. Sandi Smith for her mentorship in articulating my ideas and helping me to identify my place as a message-centered interpersonal communication scholar. To Dr. Kayla Hales for reminding me to never overlook the details. To Dr. Bill Donohue for encouraging me to always consider the practical value of my projects. In addition to my dissertation committee, I am grateful to my earlier MSU mentors: To Dr. Tim Levine for encouraging exploration and valuing new perspectives. To Dr. Joe Walther for pushing me past what I thought I was capable of to show me my true abilities. Additionally, thank you to the MSU College of Communication Arts & Sciences for funding my research. Beyond my formal MSU mentors, I was blessed with a motley crew of social support providers in the form of fellow MSU doctoral students: To Dr. Dave Clare, my permanent statistical consultant, for teaching me about how face threatening friendship can truly be, how no office is complete without its own rock collection, and that tie-dye is still cool. To Dr. Jenny Cornacchione for countless football tailgates, Netflix marathons, and being a constant ally. To Bri DeAngelis for being the only other person to understand the ache for Duluth and its brutal v winters. To my academic little sister, Morgan Summers, for esteem support, coding, and cookies. Thank you to all of my other friends, near and far, for showing me the value of social support. I am also indebted to my earlier mentors: To Dr. Krystyna Aune, my Masters advisor, for providing esteem support at a critical time in my scholarly development. To Dr. Kelly Aune for giving this shy country mouse a voice. To Dr. Amy Ebesu Hubbard for preparing me for all levels of the academy. To Dr. Mike Sunnafrank and his wife Donna for broadening my horizons and instilling a sense of adventure. To Dr. Ryan Goei for introducing me to research. This project would also not have been possible without my family – from those who raised and grew up with me to those who welcomed me into the Edwards clan. To my parents: most kids only get two parents, but I’m blessed to have four! Thank you for providing every type of social support, including invisible support I failed to recognize and support I didn’t know I needed. To my grandparents for being my constant supporters. To my siblings, Emily, Angelica, Greg, Tori, Lucas, and Marcus: thanks for never being disappointed when my grad student salary and geographic distance prevented me from giving better presents and supporting you (in person) at every event. To my newest parents, Dougger and Nora, for welcoming me with open hearts. To my newest sister, Lora Hansen, for always providing a listening ear, thoughtful advice, and a partner in crime, as well as to Mark Hansen, for teaching me what it means to be an Edwards. Finally, substantial credit is due to my husband Jon for his unwavering support – probably the equivalent of an M.A. in Communication. You reminded me that my value extends beyond my academic identity, introduced me to craft beer, used cuddle couches to cheer me up on bad days, shared your name, listened to countless conversations about theory, and encouraged me when I forgot about the light at the end of the tunnel. I’m so blessed to be starting this brand new life together. Oh, and thanks for always doing the dishes. vi TABLE OF CONTENTS LIST OF TABLES ix LIST OF FIGURES x INTRODUCTION 1 CHAPTER ONE: INTRODUCING A THEORY OF SOCIAL SUPPORT SEEKING Seeking Social Support Existing Support Seeking Research Social Support Seeking: A Face Threatening Act Message Directness Message Valence Message Assertion The Influence of FTA on Support Outcomes Affect Additional Factors That Influence Support Seeking 4 4 6 10 13 16 17 18 19 21 CHAPTER TWO: AN INITIAL TEST OF THE THEORETICAL MODEL Overview Message Dimensions & Threat to PSP Face Threat to PSP Face, Negative Affect, & Support Outcomes 23 23 23 25 CHAPTER THREE: METHOD Participants Procedure Stimuli Instrumentation Perceived Directness Perceived Valence Perceived Assertion Perceived Threat to Positive Face Perceived Threat to Negative Face Negative Affect Willingness to Provide Support Support Provision Relational Closeness Demographics 28 28 28 29 30 30 31 31 31 32 32 33 33 34 34 vii CHAPTER FOUR: RESULTS OF AN INITIAL MODEL TEST Preliminary Analyses Testing the Measurement Model Testing the Structural Model Testing Alternative Models Alternative Model 1: Removing Negative Affect Alternative Model 2: Removing Threat to Negative Face Assessing the Theoretical Hypotheses Directness and Threat to Face Valence and Threat to Face Assertion and Threat to Face Threat to Face and Negative Affect Negative Affect and Willingness to Provide Support Willingness to Provide Support and Support Provision 35 35 36 36 37 37 38 38 38 38 39 39 39 39 CHAPTER FIVE: DISCUSSION Summary of Results Implications Limitations Future Directions Conclusion 40 40 40 47 52 54 APPENDICES Appendix A Tables and Figures Appendix B Message Set A: Roommate Problems Appendix C Message Set B: Internship Frustration Appendix D Manipulation Checks Appendix E Threat to Positive Face Appendix F Threat to Negative Face Appendix G Willingness to Provide Support Appendix H Relational Closeness 55 56 65 67 69 70 71 72 73 REFERENCES 74 viii LIST OF TABLES Table A1. Theory of Polite Support Seeking (TOPSS): Axioms & Propositions 56 Table A2. Means and Standard Deviations 57 Table A3. Correlation Matrix 58 Table A4. Unstandardized, Standardized, and Significance Levels for Predicted Model 59 Table A5. Unstandardized, Standardized, and Significance Levels for Final Model 60 ix LIST OF FIGURES Figure A1. A Theory of Polite Support Seeking (TOPSS) 61 Figure A2. Initial test of a Theory of Polite Support Seeking (TOPSS) 62 Figure A3. TOPSS Predicted Model 63 Figure A4. TOPSS Final Model 64 x INTRODUCTION Social support is a ubiquitous component of interpersonal relationships (Mortenson, 2009), with important implications for the wellbeing of individuals and their relationships (see for review Callaghan & Morrissey, 1993). Beyond relational benefits, such as increased closeness and satisfaction (Brunstein, Dangelmayer, & Schultheiss, 1996; Rook, 1987), social support impacts an individual’s physical and mental health (Ell, 1984). For example, a study by Frasure-Smith et al. (2000) demonstrated that social support reduces the link between depression and cardiac mortality, as well as decreases symptoms of depression for heart attack victims. Moreover, Kroenke and colleagues (2006) found that a lack of social support was correlated with increased mortality for women diagnosed with breast cancer. These examples demonstrate the gravity of accessing social support in times of need. Individuals may experience a deficit of social support for a range of issues, from daily challenges to life-threatening situations, yet little is known about the most effective methods to solicit support. Scholarly examination of the social support process has focused largely on understanding the qualities that make support provision messages effective (e.g., Burleson, 1994; Holmstrom & Burleson, 2011). Fewer studies have considered how support seeking messages influence the social support process (Trees, 2005). This lacuna represents an important limitation, as the characteristics of support seeking messages have important implications for the outcomes of social support exchanges (Hanna, 2013; Pasch & Bradbury, 1998). This dissertation proposes and empirically assesses a Theory of Polite Support Seeking (TOPSS), which explains how the features of support seeking messages impact support outcomes. The development of TOPSS relies on both deductive and inductive approaches. Drawing on earlier work by Hanna (2013) describing the dimensions of support seeking messages that 1 lead to the provision of quality support, TOPSS builds on existing theory to explain and predict why the construction of support seeking messages influences support outcomes. In doing so, TOPSS addresses empirical questions by attempting to explain the mechanisms between message characteristics and support outcomes, as well as practical questions such as how to obtain quality support. TOPSS seeks to explain, predict, and control outcomes in support seeking interactions across contexts and relationships. TOPSS has several theoretical and practical implications. First, TOPSS increases understanding of the full social support process. The current communication literature’s primary focus on support provision and reception largely ignores an important potential antecedent variable that systematically influences support outcomes: support seeking, or more specifically, the support seeking message. Previous studies demonstrate that the way support is solicited influences support receipt and support quality (e.g., Hanna, 2013; Hui-Jung, 2009; Pasch & Bradbury, 1998). Therefore, studies that examine support outcomes without consideration for support seeking messages overlook an additional explanation why support interactions may be more or less effective. By understanding how support seeking messages influence social support outcomes, researchers can develop a fuller understanding of the full social support process. Second, TOPSS adopts a communication orientation for understanding social support seeking. The limited existing research on social support seeking has primarily been conducted by psychologists and health professionals, who focus on understanding the optimal conditions and psychological barriers for seeking help (e.g., Barbee & Cunningham, 1995; Barbee et al., 1993; Barbee, Gulley, & Cunningham, 1990; Cauce et al., 2002; Cutrona, Suhr, & MacFarlane, 1990). However, the communication perspective suggests that message features may be influential factors impacting the social support process (e.g., Hanna, 2013). Support seeking research 2 conducted from this perspective can provide information about what individuals should say to effectively garner high quality support. Practically, TOPSS identifies ways that support seekers, including those with substantial or specific support needs, can maximize the benefits of support interactions. High-quality social support is beneficial for individuals’ physical health (Callaghan & Morrissey, 1993; Ell, 1984), psychological wellbeing (Siebert, Mutran, & Reitzes, 1999), and relationships (Rook, 1987; Sanderson, Rahm, Beigbeder, & Metts, 2005). TOPSS identifies how well-developed support seeking messages allow support seekers to capitalize on the benefits of support interactions. By learning to ask for support in effective ways, support seekers can increase the likelihood that they will receive high quality, well-matched support. In situations where support is valuable, a theory of support seeking can provide information about how to effectively solicit quality support to maximize its benefit. The primary goal of this dissertation is to articulate and test the Theory of Polite Support Seeking (TOPSS), which outlines how the dimensions of support seeking messages impact support outcomes. Face theory (Goffman, 1967), politeness theory (Brown & Levinson, 1987), and the attribution-emotion-intention model (i.e., attribution theory of motivation; e.g., Weiner, 1980) inform the current theoretical model. To begin, Chapter One describes the theoretical framework, identifying the assumptions and propositions of the TOPSS. Next, Chapter Two outlines the hypotheses and Chapter Three describes the methodology for an empirical study to test the theoretical model. Then, Chapter Four reports the findings of the initial test of TOPSS. Finally, Chapter Five discusses the implications and limitations of the study and provides recommendations for future research. 3 CHAPTER ONE: INTRODUCING A THEORY OF SOCIAL SUPPORT SEEKING Seeking Social Support Social support is a process that occurs through interpersonal transaction, in which individuals request and/or exchange emotional concern, instrumental aid, information sharing, companionship, and/or appraisal (House, 1981). As this definition highlights, support takes many forms. Cutrona and Suhr (1992) distinguish two broad categories of social support: actionfacilitating support and nurturant support. Action-facilitating support focuses on helping recipients solve an underlying problem or issue, whereas nurturant support focuses addressing the distress caused by the problem. More specifically, action-facilitating support includes informational support and tangible aid, whereas nurturant support includes emotional and esteem support, as well as companionship (i.e., network support). In all forms, social support is linked to psychological wellbeing, physical health, and relational benefits (Callaghan & Morrissey, 1993; Ell, 1984; Rook, 1987; Sanderson et al., 2005; Siebert et al., 1999). A preponderance of social support literature focuses on other phases of the support process, including the provision phase (Trees, 2005) and the reception phase (Bodie et al., 2011; Burleson et al., 2009; Holmstrom, 2012; Holmstrom & Burleson, 2011). Yet the support process encompasses more than providing and receiving support. In some support exchanges, support seeking is an antecedent phase [see Figure 1]. During the seeking phase, support seekers (SS) communicate a desire for social support to a potential support provider (PSP). A PSP who perceives the support request may meet, deny, or ignore the request. During the provision phase, support providers (SP) communicate comfort to support recipients (SR) by attempting to help solve a problem (i.e., action-facilitating support) or address distress caused by a problem (i.e., nurturant support). During the reception phase, a SR who perceives support provision may 4 engage in a number of behaviors, including accepting, acknowledging, utilizing, evaluating, or ignoring the support. The current project examines the larger support process by making predictions about how dimensions of support seeking messages may influence subsequent support provision. Support seeking is defined as an active, intentional process in which messages are used to solicit support from another in times of need (Hanna, 2013). This definition makes a number of assumptions. First, the definition assumes that support seeking is active and intentional. While individuals may communicate a desire for support unconsciously, the current definition focuses on instances where individuals consciously engage in communication in order to obtain support. Second, the definition assumes that support seeking is an interpersonal process. Support seeking may also occur within group contexts; however, this definition focuses on support seeking between two people. Third, this definition assumes that support seeking occurs in times of need, rather than constraining support seeking as a response to a problem or negative event. This assumption highlights the reality that support seeking does not require the occurrence of a problem (Cutrona, 2012), but rather, that individuals seek support for a variety of experiences, including positive (e.g., a promotion), neutral (e.g., a relocation), or negative (e.g., a job loss) events. Relying on this definition, the first theoretical assumption is: Axiom 1: Social support seeking may be conceptualized as an active, intentional, interpersonal process designed to elicit support through the exchange of messages. Additionally, Hanna (2013) situated support seeking as a social influence process. Hanna demonstrated that characteristics of effective support seeking messages (e.g., politeness, the use of positively valenced communication) were consistent with broader influence literature (Barry 5 & Shapiro, 1992; Levine & Boster, 2001). Moreover, Hanna’s conceptualization of support seeking as social influence is logically consistent with existing literature that frames making a request as social influence (e.g., Brown & Levinson, 1987; Wilson, Aleman, & Leatham, 1998). However, Hanna also identified important distinctions about the support seeking process that necessitate its further study as a unique social influence sub-process. First, providing support is a relationally expected behavior (Argyle & Henderson, 1984; Roloff, 1987), whereas goals of other influence attempts may not be expected. Persuading an individual to engage in a socially normative behavior is likely to require different considerations than persuading an individual to engage in a behavior that is not expected or normative. For instance, an individual attempting to persuade a friend to engage in illicit drug use (i.e., a non-normative, stigmatized behavior; Flom et al., 2001) may be required to utilize a different kind of message than an individual attempting to persuade a romantic partner to provide emotional support (i.e., a socially expected relational behavior; Argyle & Henderson, 1984; Roloff, 1987). Second, the outcomes of support seeking attempts may have greater significance for wellbeing than the outcomes of other influence attempts (Callaghan & Morrissey, 1993; Ell, 1984; Rook, 1987; Sanderson et al., 2005; Siebert et al., 1999). The conceptualization of support seeking as an influence process yields a second theoretical assumption: Axiom 2: Social influence is a central goal of social support seeking, such that support seekers strive to influence the behavior of potential support providers. Existing Support Seeking Research Previous examinations of the support seeking process have occurred under many labels, including help-seeking (e.g., Gourash, 1978), social support elicitation (e.g., Jensen, 2001), and 6 social support solicitation (e.g., Lawrence et al., 2008; Pasch & Bradbury, 1998). Psychologists have conducted most existing research, which often focuses on identifying when support seeking occurs (e.g., Cauce et al., 2002; Liang, Goodman, Tummala-Narra, & Weintraub, 2005). For example, Cauce and colleagues (2002) observed that two conditions are optimal for seeking support: (a) an undesirable situation, and (b) the perception that help is necessary for the problem to be resolved. However, fewer studies focus on understanding how support seeking occurs. Other researchers study the methods individuals use to seek support. Beginning with observation, Cutrona and colleagues (e.g., Cutrona et al., 1990) monitored support seeking behavior in the laboratory and developed a typology of seeking and response behaviors in support interactions. The resulting typology, the Social Support Elicitation Behavior Code (SSEBC) primarily distinguishes between direct and indirect support seeking behaviors. Another approach by Barbee, Cunningham, and colleagues (Barbee & Cunningham, 1995; Barbee et al., 1993; Barbee et al., 1990) developed a theoretical model of interactive coping, sensitive interactions systems theory (SIST). The model focuses on the broader support process, beginning with the decision to seek support and ending with the long-term consequences of the support interaction. Although SIST does not focus exclusively on support seeking, it provides an additional dimension on which support seeking messages may vary: verbal – nonverbal. While the SSEBC and SIST provide a useful starting point for understanding how individuals solicit support, both are limited from the perspective of a communication scholar. First, neither approach allows for additional dimensions of support seeking messages. Second, neither approach fully examines the dimensional nature of direct-indirect or verbal-nonverbal; rather, both approaches examine these dimensions as dichotomies. Third, neither approach offers predictive utility about which messages will be most effective. 7 Hanna (2013) sought to address these limitations by examining existing literature and conducting an exploratory examination of support seeking messages to identify ways that verbal messages differ. Hanna acknowledged that nonverbal variance in support seeking may have important implications for support outcomes, but argued that differences in verbal dimensions of support seeking messages are complex enough to warrant individual study. To begin, Hanna conducted a review of existing literature on support seeking and related phenomena (e.g., helpseeking, support elicitation). From this review, Hanna articulated three verbal dimensions of support seeking messages: directness (i.e., the degree to which the message makes a straightforward request for support), valence of directed communication behavior (i.e., valence; the degree to which the message compliments or criticizes the potential support provider), and amount of self-disclosure (Barbee & Cunningham, 1995; Barbee et al., 1993; Barbee et al., 1990; Collins & Feeney, 2000; Cutrona, 2012; Cutrona, Shaffer, Wesner, & Gardner, 2007; Cutrona et al., 1990; Derlega, Winstead, Oldfield III, & Barbee, 2003; Hui-Jung, 2009; Norberg, Lindblad, & Boman, 2006; Trees, 2005). Next, Hanna collected over 800 support seeking messages across two studies using two methodologies. In Study 1, participants recalled effective (i.e., resulted in support receipt) and ineffective (i.e., did not result in support receipt) support seeking messages. In Study 2, participants produced support seeking messages in response to hypothetical scenarios in which support may be desired. Using an open and axial coding process (i.e., thematic analysis; see Hawker & Kerr, 2007), Hanna identified seven additional dimensions on which support seeking messages vary: emotional expression (i.e., the degree to which the message describes the support seeker’s emotional state), effort (i.e., the degree to which the message acknowledges the support seeker’s efforts to avoid or resolve the problem), assertions (i.e., the degree to which the message makes a request in the form of assertions), issue descriptiveness (i.e., the degree to 8 which the message describes the issue for which support is sought), locus (i.e., the degree to which the message acknowledges responsibility or places blame for the problem), future expectancy (i.e., the degree to which the message indicates the belief that things will get better or worse), and coping potential (i.e., the degree to which the message indicates that the support seeker has the strength or skill to cope with the problem). Second, Hanna (2013) investigated the relationship between variance in the dimensions of support seeking messages and support outcomes. Hanna developed quantitative coding schemes based on the thematic analysis and trained naïve assistants to code each support seeking message along the ten identified dimensions. According to the results of Study 1, directness, assertions, and valence all impact support receipt and support quality (i.e., the degree to which the message is perceived as helpful, supportive, and sensitive; Goldsmith, McDermott, & Alexander, 2000). Directness and valence positively influenced support receipt and quality. Conversely, assertions negatively influenced support receipt and quality. Hanna concluded that the directness and assertions findings reflect the importance of politeness: effective support seeking messages are less face-threatening than ineffective messages. Similarly, support seeking messages with positive valence may be less face-threatening than messages with negative valence, explaining why positively valenced messages are more effective. The current project takes Hanna’s (2013) research one step further, proposing a theory of support seeking in which degree of threat to the PSP’s face serves as a mechanism for the relationship between the dimensions of support seeking messages and support outcomes. Accordingly, Hanna (2013) provides the basis for the the third axiom: Axiom 3: Social support seeking messages vary in the degree of directness, valence, and assertion, which influences support outcomes. 9 Social Support Seeking: A Face Threatening Act Seeking support is a face-threatening process. Goffman (1967) introduced the concept of face, or “the positive social value a person effectively claims for himself [or herself] by the line others assume he [or she] has taken during a particular contact” (p. 5). In other words, face is an individual’s desired social persona or public image. Brown and Levinson (1987) extend Goffman’s notion of face by distinguishing between positive face and negative face. According to Brown and Levinson, positive face is the desire to be well-regarded, whereas negative face is the desire for autonomy. Goffman (1967) states that an individual’s face can be bolstered, maintained, or threatened in each social interaction. Face-threatening acts (FTA) occur when the desire to be well-regarded or autonomous is challenged. Politeness theory (Brown & Levinson, 1987) suggests that individuals attempt to minimize threats to face in conversation by systematically deviating from Grice’s (1989) Cooperative Principle (CP) and related conversational maxims. The CP states that individuals should, “Make your conversational contribution such as is required, at the stage which it occurs, by the accepted purpose or direction of the talk exchange in which you are engaged” (p. 45). Grice translates the principle into four conversational maxims (i.e., quality, quantity, manner, and relation) that influence how messages are constructed, such that cooperative messages are efficient. According to Brown and Levinson, individuals who deviate from Grice’s Cooperative Principle often do so in order to protect the positive or negative face of themselves or their conversational partners. For example, a speaker may violate the maxim of quality (i.e., truthfulness) by dishonestly stating, “I love this gift!” to a friend who gives an ill-chosen present in order to protect the positive face of the gift-giver. 10 Brown and Levinson propose four categories of politeness strategies: (a) bald on-record, (b) positive politeness, (c) negative politeness, and (d) off record. Speakers using bald on-record strategies do not explicitly attempt to minimize threats to face. Speakers using positive politeness strategies attempt to minimize threats to positive face, whereas speakers using negative politeness or off record strategies attempt to minimize threats to negative face. Brown and Levinson conceptualize the four politeness strategies as mutually exclusive, ordered from least to most polite. Yet a number of other scholars disagree, suggesting that the strategies are neither mutually exclusive, nor linear (Carson & Cupach, 2000; Craig, Tracey, & Spisak, 1986; Dillard, Wilson, Tusing, & Kinney, 1997; Lim & Bowers, 1991). For example, a speaker making a request may choose to utilize both positive politeness and negative politeness strategies: “You’re such a supportive friend (i.e., positive politeness). Do you think you could drive me to campus today? If not, it’s no big deal – I know how busy you are (i.e., negative politeness).” Based on this reasoning, the fourth theoretical assumption states: Axiom 4: Messages may simultaneously employ multiple politeness strategies. The social support process is fraught with implicit and potential FTA for support seekers, potential support providers, support providers, and support recipients. The support seeking phase is inherently face-threatening, consistent with other types of requests (e.g., Brown & Levinson, 1987; Craig et al., 1986; Wilson et al., 1998). It is loaded with implicit and potential threats to face for the support seeker and potential support provider. First, support seekers risk threats to positive and negative face. Seeking support may threaten seekers’ positive face by making them appear lazy and incapable of being self-reliant or lacking the foresight to manage their own problems (Craig et al., 1986; Goldsmith, 1992). Seekers also risk threats to negative face: by seeking support, support seekers may incur a debt they are expected to later reciprocate or may 11 constrain themselves to following the advice of the PSP (Goldsmith, 1992; Roloff, Janiszewski, McGrath, Burns, & Manrai, 1988). Second, the seeking phase risks threatening the face of PSPs. Potential support providers’ negative face may be threatened, as a request for support constrains their autonomy. Consequently, a PSP’s positive face may be threatened if he or she resists providing support, which may convey the inability to be counted on in times of need (Roloff et al., 1988). Furthermore, several types of face threats may concurrently occur during the support seeking phase. Although Brown and Levinson proposed that multiple face threats do not simultaneously occur, more recent research contradicts their claim (e.g., Johnson, Roloff, & Riffee, 2004a, 2004b; Wilson et al., 1998). For instance, support seekers who ask for advice may concomitantly threaten their own autonomy (i.e., SS negative face) and desire to be wellregarded (i.e., SS positive face), as well as the potential support provider’s autonomy (i.e., PSP negative face). Based on this reasoning, the fifth and sixth theoretical axioms are presented: Axiom 5: Social support seeking may threaten the positive face and negative face of both the support seeker and potential support provider. Axiom 6: In social support seeking episodes, multiple threats to face may simultaneously occur. However, the potential for a PSP’s positive face to be threatened remains controversial. In identity implications theory, Wilson, Aleman, and Leatham (1998) argue that requests for a favor should not threaten a recipient’s positive face because a request implies that the requestor believes the individual to be capable and willing to meet the request. Yet Roloff and colleagues (1988) found that being the recipient of a request might threaten positive face. To understand these contrasting perspectives, it is important to consider how the request is made. For example, 12 the request, “Will you help me move this weekend?” may not threaten positive face, but the request, “Will you help me move this weekend? I knew you wouldn’t think to volunteer, so I thought I’d ask” may threaten positive face by implying that the individual is not thoughtful. Accordingly, while threats to positive face may not be implicit in requests, there is certainly the potential for a request to threaten positive face depending on the content and style of the message used to make the request. Research by Hanna (2013) provides evidence that the directness, valence, and assertions of support seeking messages influence support outcomes. The current theory argues that the relationship between these message dimensions and support receipt centers on the degree to which each dimension threatens the face of the potential support provider. Based on this reasoning, the seventh theoretical axiom is advanced: Axiom 7: The degree to which support seeking messages are direct, valenced, and include assertions influences the degree to which the message threatens face. Drawing on existing research, TOPSS argues that message directness, valence, and assertion individually influence threat to face in predictable ways. Message directness and valence are content dimensions, which vary depending on what information is included, whereas assertion is a style dimension, which varies depending on how the message is delivered. The following sections outline the predicted impact of each dimension on perceived threat to positive and negative face. Message Directness The TOPSS maintains that support seeking messages vary in directness. Directness is a content dimension ranging from direct to indirect, which is conceptualized as the degree to 13 which the support seeking message makes a straightforward request for social support (Hanna, 2013). The most direct messages not only request support clearly from the PSP, but also request a specific type of support (e.g., esteem support), whereas with the least direct support seeking messages, it is ambiguous whether support of any kind is desired. Social support researchers consistently demonstrate that direct support seeking messages are more effective than indirect support seeking messages (Collins & Feeney, 2000; Derlega et al., 2003; Hanna, 2013; Hui-Jung, 2009). For instance, Hui-Jung (2009) established that individuals using online message boards prefer to respond to messages that ask for help (i.e., direct request) than hint at a desire for help (i.e., indirect request). Additionally, Collins and Feeney (2000) found that direct requests for support were associated with increased care and supporter responsiveness, whereas indirect requests for support were associated with less helpful support. Although direct support seeking messages have been demonstrated to be effective, direct messages may also be face-threatening. A direct request for support may be categorized as a task-oriented bald on-record strategy in politeness theory (Brown & Levinson, 1987). According to Brown and Levinson, bald on-record strategies do not focus on minimizing threat to the hearer’s face. The conceptualization of bald on-record strategies as impolite reflects the broader perspective that politeness and efficiency are bipolar; that is, a message may either be polite or efficient, but not polite and efficient. However, Kellerman and Shea (1996) argue that politeness and efficiency are meta-goals for message production that may be independent, compatible, or contradictory depending on situational constraints. Hanna (2013) asserts that these meta-goals are compatible in the support seeking process, such that the most effective support seeking messages are both polite and efficient. This perspective reflects the current project’s eighth theoretical axiom: 14 Axiom 8: Politeness and efficiency are compatible meta-goals in the support seeking process. The current theory proposes that the more direct the support seeking message, the more it is efficient and somewhat polite: Although highly direct support seeking messages may threaten a PSP’s negative face by constraining behavior, the same directness may minimize threat to a PSP’s positive face by taking the guesswork out of providing support. More specifically, direct support seeking messages threaten the PSP’s negative face by asking or telling the PSP to engage in a stated behavior. For example, a highly direct request such as, “Give me advice about studying for my exam,” threatens the PSP’s autonomy in two ways: (1) by requesting support, and (2) by requesting that the provided support be informational support. Simultaneously, direct support seeking messages minimize threat to the PSP’s positive face by making it clear what the SS desires. Thus, direct support seeking messages minimize threats to PSPs’ desire to be wellregarded in two ways: (1) by ensuring that support is desired (i.e., that they are not butting-in to the SS’s business uninvited; Wardaugh, 1985), and (2) by ensuring they have not provided the wrong type of support (e.g., informational support when emotional support is desired). For example, the PSP who hears, “Give me advice about studying for my exam,” need not worry about appearing presumptuous for providing unwanted counsel or socially incompetent for providing the incorrect type of assistance. By following this logic, direct messages can be both polite and efficient by minimizing threats to the PSP’s positive face (i.e., positive politeness) while making specific interaction goals clear (i.e., efficient), in spite of the increased threats to the PSP’s negative face that are inherent in any request. Accordingly, the first core theoretical proposition states: 15 Proposition 1: The more direct the support seeking message, the greater the threat to the potential support provider’s negative face, but lesser the threat to the potential support provider’s positive face. Message Valence Support seeking messages also vary in valence. Valence (i.e., valence of directed communication behavior at the potential support provider) is a content dimension ranging from positive to negative, which is conceptualized as the degree to which the support seeking message (a) compliments or expresses gratitude toward the PSP, or (b) blames or criticizes the PSP (Hanna, 2013). Previous research in support seeking suggests that messages with positive valence are the most effective for soliciting support, whereas messages with negative valence are ineffective (Hanna, 2013; Pasch & Bradbury, 1998). These findings are consistent with research on effective social influence strategies, including flattery and ingratiation. For example, a study by Barry and Shapiro (1992) found that compliance-gaining messages that include flattery (i.e., positively valenced) were more effective than logical appeals or assertive requests. According to Goffman (1967), messages that imply respect, esteem, or standing bolster (i.e., give) face. Thus, support seeking messages with positive valence have implications for a PSP’s positive face. Independently, compliments or expressions of gratitude may bolster positive face; however, the effect may not be as strong when coupled with a request, as requests are intrinsically facethreatening (Brown & Levinson, 1987). Positive valence messages directed at the potential support provider may be consistent with positive politeness strategies identified by Brown and Levinson. Thus, instead of bolstering face in support seeking interactions, messages with positive valence may merely minimize inherent face threats. On the other end of the spectrum, support seeking messages with negative valence threaten a PSP’s positive face. Brett and 16 colleagues (2007) reported that conveying anger (e.g., “You should have done this – I’m pissed!”) threatens positive face, which may result in face threat reciprocation, such as request refusal (Johnson, 2007). Based on this logic, the second core theoretical proposition states: Proposition 2: The more positive the valence of the support seeking message, the more it minimizes the intrinsic face threat of a request, whereas the less positive the message, the greater the threat to the potential support provider’s positive face. Message Assertion Additionally, support seeking messages vary in degree of assertion. Assertion is a style dimension ranging from high in assertions to low in assertions, which is conceptualized as the degree to which the support seeking message occurs in the form of commands or demands (Hanna, 2013). High assertion messages are directive, whereas low assertion messages are inquisitive. According to Brown and Levinson (1987), directives intrinsically threaten autonomy (i.e., negative face). However, high assertion messages may also challenge positive face. Brett and colleagues report: “Commands may be perceived as a lack of respect, even contempt. By signaling the expectation of compliance, commands may attack face by conveying that the speaker is in a one-up position (Ridgeway & Berger, 1986). This signal may be reinforced by the language itself, which may be interpreted as signaling power held by the speaker (e.g., Bradac, 1990; Shapiro & Bies, 1994)” (2007, p. 90). In line with these arguments, the third core theoretical proposition states: Proposition 3: The lower the assertion of the support seeking message, the more effective, whereas the higher the assertion of the support seeking 17 message, the greater the threat to the potential support provider’s positive and negative face. The Influence of FTA on Support Outcomes The degree to which support seeking messages threaten the PSP’s face is likely to influence subsequent PSP behavior. Individuals respond to face-threatening acts in a number of different ways, including behaving aggressively and interpreting messages negatively rather than literally (Brett et al., 2007; Holtgraves, 1991). Broadly, individuals whose face has been threatened may engage in reparative facework (Goffman, 1967). Metts (1997) defines facework as, “a variety of communicative devices available to interactants for preventing face loss (both their own and others’), restoring face if lost, and facilitating the maintenance of poise in the advent of disrupted interactions” (p. 374). More specifically, corrective facework occurs after a FTA, whereas in preventative facework, an individual avoids or ignores a FTA. Johnson and colleagues (e.g., Johnson, 2007; Johnson, Roloff, & Riffee, 2004a, 2004b) identified one form of corrective facework: request refusal. By refusing a request, individuals have the opportunity to repair their own negative face by asserting their autonomy, while simultaneously delivering a reciprocal threat to the negative face of the requester (Johnson et al., 2004a). Thus, a PSP may refuse to provide support in order to repair negative face. Additionally, individuals who refuse a request may repair or enhance their own positive face (Folkes, 1982; Johnson et al., 2004a). For instance, an individual who refuses to provide emotional support to validate the behavior of a friend who engaged in infidelity may appear moral or conscientious toward the friend’s romantic partner. Based on this evidence, the fourth core theoretical proposition asserts: 18 Proposition 4: Increased degree of threat to the potential support provider’s positive face or negative face decreases the likelihood of support provision (i.e., request compliance). Affect Additional evidence suggests that affect may mediate the relationship between threat to face and request refusal (e.g., failure to provide support). Affect is a generalized motivational state, which is usually intense, short-lived, and directed at some external source (Nabi, 2010). Positive affect refers to “the extent to which a person feels enthusiastic, active, and alert” (Watson, Clark, & Tellegen, 1988, p. 1063). Conversely, negative affect refers to the extent to which a person experiences “subjective distress and unpleasurable engagement that subsumes a variety of aversive mood states, including anger, contempt, disgust, guilt, fear, and nervousness” (Watson et al., 1988, p. 1063). Abundant studies demonstrate that individuals may experience negative affect, including anger, hurt, embarrassment, defensiveness, and frustration upon encountering a threat to positive or negative face (Carson & Cupach, 2000; Cupach & Carson, 2002; Kennedy-Lightsey, 2010; Rains, 2013; Zhang & Stafford, 2008). Negative affect influences support outcomes, including support provision and willingness or intent to provide support, among others. For example, a study by Clare and colleagues (2014) examined how contextual factors influence the provision of support in the job-seeking environment. Relying on Weiner’s (1980) attribution-emotion-intention model, the study found that attributions about the support seeker and situation predicted a potential helper’s anger, which negatively influenced provision of support. More proximally, negative affect may influence willingness or intent to provide support. The attribution-affect-action theory (i.e., the attributionemotion-intention model) predicts that attributions about a support seeker influence support 19 provider anger and sympathy, which impacts a support provider’s help behavior (Weiner, 1995). In a series of six experiments, Weiner (1980) demonstrated that attributions in a variety of helpseeking situations influenced helper anger and disgust (i.e., negative affect), which resulted in unwillingness to help (i.e., avoidance behavior). Weiner’s findings have been replicated across several studies (e.g., Berkowitz, 1987; Meyer & Mulherin, 1980; Reisenzein, 1986; Schmidt & Weiner, 1988). Beyond influencing willingness to provide support and support provision, negative affect also impacts other support outcomes, such as the type, quality, or amount of support provided. For instance, MacGeorge (2001) investigated the relationship between support provider affect and goals to provide specific types of support. MacGeorge extended Weiner’s (1980) attributionemotion-intention model of helping by suggesting that beyond broadly influencing helping intentions, emotions influence support seekers’ specific goals for how to provide support. MacGeorge examined the impact of anger and sympathy on support interaction goals. Specifically, MacGeorge identified five categories of support interaction goals: (a) goals to provide problem-solving (i.e., action-facilitating) support, (b) goals to provide emotional (i.e., nurturant) support, (c) goals to distract, (d) goals to provide perspective, and (e) goals for determining responsibility. The study found that anger and sympathy differentially affected interaction goals. For instance, anger decreased goals to provide emotional support, whereas sympathy increased goals to provide problem-solving support. Further, MacGeorge argues that a support provider’s interaction goals impact the quality of support messages, although she does not directly test this assertion. In spite of a lack of empirical evidence, the notion that a support provider who experiences negative affect may be less motivated to develop appropriate interaction goals, resulting in lower quality support, makes 20 intuitive sense. In this way, negative affect may also help to explain Hanna’s (2013) finding that polite support messages lead to higher quality support than impolite messages. Thus, the relationship between negative affect and support quality warrants additional investigation. Conversely, social support outcomes can be influenced by positive affect. Several decades of research demonstrate a link between positive affect and helping (i.e., helping) behavior (e.g., Clark & Isen, 1982; Isen, 1987). For example, a study by George (1991) demonstrated that a salesperson’s state, not trait, positive affect predicted supervisor reports of helping others with work-related problems. Carlson, Charlin, and Miller (1988) identify two types of explanations for the positive affect – helpfulness link: First, they argue that positive affect can increase the salience of positive reinforcement. For example, positive affect may eliminate competing motivations, such as the desire to save face. Second, they argue that positive affect can increase perceptions that helping someone will be positively reinforcing. As an example, positive affect may improve one’s social outlook or perception of the goodness of human nature. For these reasons, positive affect is likely to increase positive support outcomes. Based on the arguments detailed above, the fifth core theoretical proposition states: Proposition 5: The potential support provider’s affect mediates the relationship between threat to face and support outcomes, including willingness to provide support, support provision, support type, and support quality, such that negative affect decreases support outcomes, whereas positive affect increases support outcomes. Additional Factors That Influence Social Support Seeking A number of additional contextual, relational, and individual factors are likely to influence the perceptions and behaviors of support seekers and potential support providers during 21 the social support process. First, support seekers’ construction of messages may be affected by a number of factors, including level of distress, relationship type, and the support seeker’s sex (Baxter, 1984; Collins & Feeney, 2000; Wellman & Wortley, 1990). For instance, Hanna (2013) demonstrated that a support seeker’s goals for specific forms of support influence message dimensions. More specifically, Study 1 demonstrated that the desire for nurturant support was negatively related to message politeness. Second, potential support providers’ perceptions and behaviors may be affected by multiple factors, including degree of imposition, relational characteristics, and face sensitivity (Argyle & Henderson, 1984; Brown & Levinson, 1987; White, Tynan, Galinsky, & Thompson, 2004). For example, power may moderate the relationship between negative affect and support outcomes (Brown & Levinson, 1987). In particular, a potential support provider who experiences negative affect as the result of facethreatening interaction with a supervisor may provide support despite feeling angry due to the power distance between the PSP and supervisor. Accordingly, the ninth core theoretical axiom asserts: Axiom 9: Additional contextual, relational, and individual factors related to the support seeker, potential support provider, and seeking interaction will influence the support process. 22 CHAPTER TWO: AN INITIAL TEST OF THE THEORETICAL MODEL Overview Chapter 3 proposes an initial test of TOPSS (see Figure 2 for full theoretical model). Drawing on the axioms and propositions of the theory, this chapter advances eight hypotheses to provide a partial test of the proposed theoretical model (see Figure 3 for model to be tested in the current study). More specifically, this study tests (a) the impact of support seeking message dimensions on threat to the potential support provider’s positive and negative face, (b) the impact of face threat on potential support provider affect, and (c) the influence of negative affect on willingness to provide support and (d) the effect of willingness to provide support on actual support provision. While the full theoretical model makes broad predictions about social support outcomes, the current study focuses on two specific support outcomes, willingness to provide support and support provision. Additionally, the current study does not examine the impact of any additional contextual, relational, or individual factors on the support process; rather, the current study focuses on testing the central tenants of TOPSS. Message Dimensions & Threat to PSP Face TOPSS proposes that support seeking messages vary in directness (Axiom 3) and that variance in directness threatens the positive and negative face of potential support providers in specific ways (Axiom 5, Axiom 6, Axiom 7, and Proposition 1). In light of the assumption that politeness and efficiency are not bipolar in support seeking (Axiom 8), Proposition 1 states that highly direct support seeking messages are theoretically sophisticated and effective, combining meta-goals for politeness and efficiency. Two hypotheses can be derived to test the combination of Axiom 3, Axiom 5, Axiom 6, Axiom 7, and Proposition 1: 23 Hypothesis 1: Message directness is negatively associated with perceived threat to the recipient’s positive face. Hypothesis 2: Message directness is positively associated with perceived threat to the recipient’s negative face. Additionally, TOPSS states that support seeking messages vary in valence (Axiom 3) and that variance in valence influences the positive face of potential support providers in specific ways (Axiom 7 and Proposition 2). Proposition 2 states that support seeking messages with high (i.e., more positive) valence are theoretically sophisticated and effective, minimizing the intrinsic face threat of a request, whereas messages with low (i.e., more negative) valence threaten the PSP’s positive face. One hypothesis can be derived to test this combination of Axiom 3, Axiom 7, and Proposition 2: Hypothesis 3: Message valence is negatively associated with perceived threat to the recipient’s positive face. TOPSS also predicts that support seeking messages vary in assertion (Axiom 3) and that variance in assertion threatens the positive and negative face of potential support providers in specific ways (Axiom 5, Axiom 6, Axiom 7, and Proposition 3). Proposition 3 states that low assertion support seeking messages are theoretically sophisticated and effective, whereas high assertion support seeking messages threaten the PSP’s positive and negative face. Two hypotheses can be derived to test the combination of Axiom 3, Axiom 5, Axiom 6, Axiom 7, and Proposition 3: Hypothesis 4: Message assertion is positively associated with perceived threat to the recipient’s positive face. 24 Hypothesis 5: Message assertion is positively associated with perceived threat to the recipient’s negative face. Threat to PSP Face, Negative Affect, & Support Outcomes TOPSS proposes that threats to face decrease the likelihood of support provision (Proposition 4) and other support outcomes due to negative affect are caused by threats to a potential support provider’s face (Proposition 5). Specifically, Proposition 5 states that a potential support provider’s affect mediates the relationship between threat to face and support outcomes, including willingness to provide support, support provision, support type, and support quality. The current test exclusively examines the impact of threat to face on negative affect, although future studies should also examine their impact on positive affect. To test the relationship between threats to face and negative affect articulated by Proposition 5, two hypotheses are derived: Hypothesis 6: Threat to positive face increases a potential support provider’s negative affect. Hypothesis 7: Threat to negative face increases a potential support provider’s negative affect. Although TOPSS predicts that negative affect will influence support outcomes broadly, the current theoretical test focuses on two related support outcomes: The current study tests the impact of negative affect directly on attitude toward willingness to provide support and indirectly on support provision (i.e., mediated by willingness to provide support). Separately, studies demonstrate that negative affect can decrease willingness to provide support and support provision (e.g., Clare et al., 2014; Weiner, 1980), yet no studies examine the relationship between those outcomes. Even Weiner’s tests of attribution-affect-action theory stops short of 25 determining whether a helper’s self-reported likelihood or intent to help leads to increased helping behavior. Nonetheless, Weiner (1985) states, “Expectancy and affect, in turn, are presumed to guide motivated behavior,” (p. 548), suggesting a correlation between willingness to help (i.e., an attitude) and helping (i.e., a behavior). The relationship between attitude and behavior is well studied in the broader social influence literature. An examination of that literature demonstrates a complex, controversial relationship between behaviors (e.g., providing support), attitudes (e.g., a willingness to provide support), and behavioral intentions (e.g., the intention to provide support). One of the earliest studies by LaPiere (1934) argued for a negative relationship between attitudes and behaviors after observing that reported attitudes did not match actual behaviors for serving Chinese travelers at dining and lodging establishments. Conversely, a meta-analysis by Kim and Hunter (1993a) examined studies containing 138 reported attitude-behavior correlations that included over 90,000 subjects. The results indicated that the average observed correlation between attitudes and behavior was r = .47. After correcting for error of measurement and restriction of range, the correlation increased to r = .79, indicating a substantial positive relationship between attitudes and behaviors. Moreover, a later meta-analysis by Kim and Hunter (1993b) probing the attitude-behavior link as mediated by behavioral intention reported data consistent with a strong link between attitudes, behavioral intentions, and behavior. Based upon this reasoning, the current study tests TOPSS by focusing on two related outcomes: (1) willingness to provide support and (2) support provision. Two hypotheses are derived from Proposition 5 and social influence literature on the attitude-behavior link: Hypothesis 8: A potential support provider’s negative affect decreases willingness to provide support. 26 Hypothesis 9: A potential support provider’s willingness to provide support predicts support provision. 27 CHAPTER THREE: METHOD The current study relies on a 2 (direct or indirect) x 3 (positive, neutral, or negative valence) x 2 (high assertions or low assertions) between-subjects online survey design to examine the direct and indirect impact of the manipulated message dimensions onto (a) perceived threat to PSP’s positive face, (b) perceived threat to PSP’s negative face, (c) PSP’s negative affect, (d) PSP’s willingness to provide support, and (e) PSP’s decision to provide a supportive message. The dimensions of support seeking messages were categorically examined in this initial test of TOPSS; however, future research is required to examine the role of the full message dimensions in the theoretical model. Participants A total of 249 participants were recruited from undergraduate communication courses at a large Midwestern university, through a departmental research pool. Participants received research credit toward applicable courses. Five participants’ responses were removed due to missing data, resulting in 244 participants remaining in analyses. The participant sample was 73.1% White (n = 178), 10.3% Black or African American (n = 25), 5.4% Asian (n = 13), 5.4% Multiracial (n = 13), 2.0% Hispanic or Latino/Latina (n = 5), 0.9% American Indian (n = 2), and 2.0% indicated they were part of a race not listed above (n = 5). Two participants (0.9%) elected not to answer the question. Participants were 52.5% female (n = 127) and ranged in age from 18 to 33 (M = 20.10, SD = 3.47). Procedure Following recruitment, participants were directed to an online survey hosted at Qualtrics.com. The landing page presented an electronic consent form; progression to the next page constituted consent. Next, participants read one of 12 support seeking messages created by 28 the 2 (direct or indirect) x 3 (positive, neutral, or negative) x 2 (high assertion or low assertion) design. The instructions began, “The current study is designed to examine friendship between college-aged adults. Throughout this study, please think of a specific close friend. Enter this friend’s initials below.” After entering a friend’s initials and responding to items measuring the closeness of the relationship, the instructions continued: “This survey asks you to imagine you are talking to the close friend you mentioned on the last page. Please carefully read the following message, as the remainder of this survey will ask you questions related to the message.” Next, the survey presented one of 12 support-seeking messages. In the following section, participants were asked to evaluate the directness, valence, and assertion of the message, which served as a manipulation check of the messages. Next, participants responded to several sections of a questionnaire: (a) threat to positive and negative face, (b) affect, and (c) willingness to provide social support. In the next section, the instructions read: “Below, imagine you have the opportunity to respond to your friend’s message. Your message can be as long or short as you would like. Please try to avoid spelling errors and try to be clear in your wording. What would you say to your friend?” Finally, participants were asked to provide demographic information before being thanked for their participation and directed to a separate page to provide information necessary to receive survey credit. Stimuli Two sets of 12 support seeking messages were created for this study. Each set of messages was developed by drawing on a scenario from Hanna (2013), who pretested the degree to which support may be desired in fifteen situations. Hanna demonstrated that a wide variety of support might be wanted in both of the scenarios selected (i.e., frustration with a roommate and problems with an internship). For each scenario, the researcher created twelve unique messages, 29 varying in directness (i.e., high or low), valence (i.e., positive, neutral, or negative), and assertions (i.e., high or low). A full list of messages is available in Appendices A & B. The researcher pilot tested the two sets of messages for realism and consistency between manipulations and manipulation checks to select one scenario for full data collection. Participants (N = 95) were drawn from the same department as participants for the full study. Participants received research or extra credit for participation. The participant sample was 78% White (n = 74), 12.7% Black or African American (n = 12), 6.3% Asian (n = 6), 1.0% Hispanic or Latino/Latina (n = 1), 1.0% Multiracial (n = 1), and 1.0% indicated they were part of a race not listed above, specifying Middle-Eastern (n = 1). Participants were predominantly female (60.4%, n = 58) and ranged in age from 18 to 30 (M = 20.92, SD = 1.83). The participants in the pilot study were demographically similar to the students in the full study; however, students who participated in the pilot study were ineligible to participate in the full study. In the pilot study, there was no difference in realism between Message Set A (n = 46, M = 5.76, SD = 1.17) and Message Set B (n = 49, M = 5.57, SD = 1.19), t (93) = 0.95, p = .42. However, message set B demonstrated greater consistency between message dimensions (i.e., manipulations) and perceived message dimensions (i.e., manipulation checks). Based on these analyses, the researcher selected Message Set B about problems with an internship for use in the current study. Instrumentation Perceived Directness. Six items developed from Hanna’s (2013) directness definition and coding scheme were pretested (a = .86) and used as a manipulation check for message directness, a = .88. As an example, participants were asked, “To what extent does the message 30 make a clear request for social support?” where 1 = Not at all, 3 = Somewhat, and 5 = Very much so. A full list of items is included in Appendix C. Perceived Valence. Six items developed from Hanna’s (2013) valence definition and coding scheme were pretested (a = .79) and used as a manipulation check for message valence, a = .89. For instance, participants were asked, “To what extent does the message compliment or express gratitude toward you?” and “To what extent does the message criticize, blame, or accuse you?” where 1 = Not at all, 3 = Somewhat, and 5 = Very much so. A full list of items is included in Appendix C. Perceived Assertion. Six items developed from Hanna’s (2013) assertion definition and coding scheme were pretested (a = .93) and used as a manipulation check for message assertions, a = .93. For example, participants were asked, “To what extent does the message make a demanding request?” where 1 = Not at all, 3 = Somewhat, and 5 = Very much so. A full list of items is included in Appendix C. Perceived Threat to Positive Face. To assess threat to the potential support provider’s positive face, the researcher utilized a measure created by Cupach and Carson (2002). The original measure includes 10 items; however, two items not appropriate for the current context (i.e., focused on the relationship between sender and receiver) were excluded. Two additional items were constructed and pretested (a = .89) to replace the excluded items, a = .87. Each 7point Likert-type (1 = Strongly disagree, 7 = Strongly agree) item begins with the root “This person’s actions…” (i.e., modified from the original “My partner’s actions…”) and describes the hypothetical message sender’s action. For example, one item states, “This person’s actions were rude,” while a reverse-coded items states, “This person’s actions were tactful.” A full list of items is included in Appendix D. 31 Perceived Threat to Negative Face. To assess threat to the potential support provider’s negative face, the researcher utilized a 4-item measure created by Cupach and Carson (2002). Six additional items were constructed and pretested in an attempt to increase scale reliability, a = .93, .95. Each 7-point Likert-type (1 = Strongly disagree, 7 = Strongly agree) item begins with the root “This person’s actions…” (i.e., modified from the original “My partner’s actions…”) and describes the hypothetical message sender’s action. For example, one item states, “This person’s actions took away some of my independence.” A full list of items is included in Appendix E. Negative Affect. Negative affect was measured using the 10-item negative affect subscale of Watson, Clark, and Tellegen’s (1988) abbreviated Positive and Negative Affect Schedule (PANAS) , a = .95, .96. Watson, Clark, and Tellegen conducted a study establishing that the abbreviated PANAS is internally consistent, stable, and uncorrelated, as well as demonstrates convergent and discriminant validity. Negative affect is conceptualized as subjective distress consisting of various negative mood states. Low negative affect is calm and serene, whereas high negative affect represents high levels of distress. The directions for the scale read: “This scale consists of a number of words that describe different feelings and emotions. Read each item and then mark the appropriate answer. Indicate to what extent you feel this way right now, as a result of the support seeking message.” These directions were modified from the original scale, which asked participants to indicate “the extent to which you feel this way right now, that is, at the present moment.” Each item was assessed on a 5-point scale where 1 = Very slightly or not at all, 2 = A little, 3 = Moderately, 4 = Quite a bit, and 5 = Extremely. The negative affect words include: distressed, upset, guilty, scared, hostile, irritable, ashamed, nervous, jittery, and afraid. However, only scared, hostile, irritable, ashamed, jittery, and afraid were retained following confirmatory factor analysis. 32 In addition, the 10 items from the positive affect subscale of PANAS (Watson, Clark, & Tellegen, 1988) were included a = .92, .92. The positive affect subscale was consistently formatted with the negative affect subscale and included the following items: attentive, interested, alert, excited, enthusiastic, inspired, proud, determined, strong, and active. However, only excited, enthusiastic, inspired, and proud were retained following confirmatory factor analysis. Willingness to Provide Support. This study measured general willingness to provide support, as well as willingness to provide each of Cutrona and Suhr’s (1992) types of support. Seven items were created and pretested to assess the participant’s general willingness to provide support to the support seeker (a = .94, .97). Prior research on willingness to provide support (i.e., help) utilizes mixed approaches, from behavioral data (e.g., Berkowitz, 1987) to items created for a specific context (e.g., Clare et al., 2013). The items were designed by modifying items from previous studies to fit the broader support context (Clare, Hanna, Holmstrom, & Stolz, 2013; Clare et al., 2014; Weiner, 1980, 2014). For example, one item states, “I would like to provide support to this person.” A full list of items is available in Appendix F. Support Provision. Two trained coders who were naïve to the study’s hypotheses dichotomously coded messages as support or non-support. Social support was defined to coders as a process that occurs through interpersonal transaction, in which individuals request and/or exchange emotional concern, instrumental aid, information sharing, companionship, and/or appraisal. Coders agreed for 97% of messages. Coder disagreements were resolved through discussion. Instances where the participant did not provide a message or provided a message that was categorized as non-support were coded as 0 (n = 20), whereas messages that were categorized as support were coded as 1 (n = 224). 33 Relational Closeness. Relational closeness between the potential support provider and hypothetical support seeker was assessed using Dibble, Levine, and Park’s (2011) relational closeness measure. The scale consists of 12 7-point Likert-type items that load on a unidimensional factor. For example, one item states, “I think about my friend a lot.” Items were averaged to provide a score for relational closeness. Four items were dropped from further analyses based on the results of confirmatory factor analyses, α = .93, .95. On a 7-point scale, the friend participants selected to consider for the hypothetical scenario was close (M = 6.04, SD = 0.94). See Appendix G for a full list of items and to see which items were retained. Demographics. In addition, the questionnaire measured the age, sex, and race of each participant. 34 CHAPTER FOUR: RESULTS OF AN INITIAL MODEL TEST This chapter reports the results of an initial test of the Theory of Polite Support Seeking (TOPSS). First, this chapter presents preliminary analyses, including manipulation check analyses. Second, it outlines a test of the measurement model using confirmatory factor analysis. Third, it describes the results of a structural equation model of the hypothesized structural model. Fourth, this chapter details the post-hoc exploration of alternate structural models. Preliminary Analyses The researcher conducted several analyses to ensure that the manipulated variables (i.e., message directness, message valence, and message assertions) were perceived as intended. First, a t-test confirmed that participants perceived direct messages (M = 5.80, SD = 1.00) as more direct than indirect messages (M = 5.39, SD = 0.97), t (242) = -3.29, p < .001. Second, an ANOVA demonstrated that participants perceived positive valence messages (M = 5.02, SD = 1.24) as more positive than neutral valence messages (M = 4.18, SD = 1.44) or negative valence messages (M = 4.51, SD = 1.32), F (2, 241) = 8.16, p < .001. However, a t-test showed that participants perceived no difference in valence between neutral and negative valence messages, t (160) = 1.57, p = .12. Third, a t-test confirmed that participants perceived high assertion messages (M = 3.67, SD = 1.54) as stylistically more assertive than low assertion messages (M = 3.23, SD = 1.39), t (242) = -2.35, p < .05. Overall, these analyses demonstrated that the manipulations were effective, with the exception of the distinction between neutral and negative valence messages However, the manipulation checks also reveal that although the manipulated dimensions significantly varied, they did not represent the full range to which support seeking messages can vary in directness, assertions, and valence. 35 Testing the Measurement Model The researcher employed the two-step method advocated by Kline (2005) to test the theoretical model and its specific hypotheses: (1) test the measurement model using confirmatory factor analysis (CFA) and (2) test the structural model using structural equation modeling (SEM). The researcher tested goodness of fit for each model with a chi-square, root mean square error approximation (RMSEA), and comparative fit index (CFI), indices that are commonly reported by communication scholars (e.g., Afifi & Afifi, 2009; Afifi & Weiner, 2006; Ledbetter et al., 2011; Southwell & Torres, 2006). Models demonstrating goodness of fit have a nonsignificant chi-square, a RMSEA of 0.06 or less, and a CFI of 0.9 or greater (Bentler & Bonett, 1980; Hu & Bentler, 1999). However, the significance of chi-square is a less important indicator of fit than others, as it is heavily influenced by sample size (Bentler & Bonett, 1980). To begin, the researcher performed confirmatory factor analysis (CFA) using AMOS version 21 to assess the fit of the measurement model. The original model demonstrated inadequate fit, χ² (2813) = 6508.691, p < .01, RMSEA = 0.07 (90% CI = 0.06 – 0.09), CFI = 0.79. Items were removed iteratively based on residual error until a model with adequate fit could be identified, χ² (1188) = 2426.38, p < .01, RMSEA = 0.06 (90% CI = 0.06 – 0.07), CFI = 0.90. The final measurement model included reduced items for most factors [see appendices for retained and removed items]. Testing the Structural Model The researcher also used AMOS version 21 to test the hypothesized regression paths of the structural model. The structural equation model used maximum likelihood estimation and contained eight observed constructs (i.e., message directness, message valence, message assertions, threat to positive face, threat to negative face, negative affect, willingness to provide 36 support, and support provision). The hypothesized model demonstrated poor fit, χ² (16) = 252.77, p < .01, RMSEA = 0.25 (90% CI = 0.22 – 0.27), CFI = 0.39 (Hu & Bentler, 1999) [see Table 3 & Figure 4]. However, examination of the individual regression paths indicated several significant pathways, suggesting that although the full theoretical model could not be supported, additional investigation of the data was warranted. Testing Alternative Models Although the full theoretical model failed, the researcher conducted post-hoc exploration in an attempt to determine which parts of the model were faulty. The researcher examined two alternative models in an attempt to find a structural model to fit the data. Alternative Model 1: Removing Negative Affect. For the first alternative model, the researcher removed negative affect as a mediator between threats to face and willingness to provide support. Both conceptual and statistical rationale exists for this model: Conceptually, there may be several problems with the use of affect in the current study. First, it may be difficult to induce negative affect with an online survey employing a hypothetical methodology. Second, the measurement of negative affect may have been too broad or focused on the wrong aspects of negative affect. The literature primarily focuses on anger, disgust, and hurt as responses to face threats (e.g., Cupach & Carson, 2002; Rains, 2013; Zhang & Stafford, 2008), whereas the negative affect scale PANAS includes feelings of guilt, shame, and nervousness. Statistically, a correlation matrix [see Table 2] revealed that the correlations (a) threat to positive face with willingness to provide support (r = - 0.49) and (b) threat to negative face with willingness to provide support (r = - 0.47) were both greater than the correlation between negative affect and willingness to provide support (r = - 0.28), indicating that negative affect does not mediate the relationship between threat to face and willingness to provide support. After 37 removing negative affect from the model and adding directional effects between threats to positive and negative face with willingness to provide support, the model continued to demonstrate poor fit, χ² (10) = 188.48, p < .01, RMSEA = 0.27 (90% CI = 0.24 – 0.31), CFI = 0.42. Alternative Model 2: Removing Threat to Negative Face. Building on the first alternative model, the researcher removed threat to negative face from the model based on an examination of the initial model. None of the manipulated message dimensions significantly predicted threat to negative face in the theoretical model or correlated with threat to negative face in the correlation matrix [see Table 2]. After removing threat to negative face from the model, the model demonstrated excellent fit, χ² (7) = 3.76, p = 0.81, RMSEA = 0.00 (90% CI = 0.00 – 0.05), CFI = 1.00. Based on the goodness of fit, the researcher selected the second alternative model as the final model [see Table 4 & Figure 5]. . Assessing the Theoretical Hypotheses To assess the study’s hypotheses, the researcher examined the final alternative model that demonstrated fit, as SEM indicated that the original theoretical model did not fit the data. Directness and Threat to Face. The first two hypotheses predict that message directness influences threat to positive and negative face. H1 states, “Message directness is negatively associated with perceived threat to the recipient’s positive face.” The final model supports H1, B = -.47, ß = -.20, p < .01. H2 states, “Message directness is positively associated with perceived threat to the recipient’s negative face.” However, the final model does not support H2, B = -.16, ß = -.06, p = .32. Valence and Threat to Face. H3 states, “Message valence is negatively associated with perceived threat to the recipient’s positive face.” The final model does not support H3, B = -.16, 38 ß = -.11, p = .08, yet the pathway between valence and perceived threat to the recipient’s positive face is retained in the model. Assertion and Threat to Face. The fourth and fifth hypotheses predict that message assertions threaten both positive and negative face. The final model does not support H4 (B = .19, ß = .08, p = .22) or H5 (B = -.13, ß = -.05, p = .41); however, the pathway between assertions and threat to positive face (i.e., H4) is in the predicted direction and is included in the final model. Threat to Face and Negative Affect. H6 states, “Threat to positive face increases a potential support provider’s negative affect.” H6 is not supported, as it does not appear in the final model. H7 states, “Threat to negative face increases a potential support provider’s negative affect.” H7 is also not supported and does not appear in the final model. Negative Affect and Willingness to Provide Support. H8 states, “A potential support provider’s negative affect decreases willingness to provide support.” The final model does not support H8, as negative affect is dropped entirely from the final model. Willingness to Provide Support and Support Provision. H9 states, “A potential support provider’s willingness to provide support predicts support provision.” The final model provides support for H9, B = .12, ß = .39, p < .01. 39 CHAPTER FIVE: DISCUSSION The current study tested the Theory of Polite Support Seeking (TOPSS), which predicts how the construction of support seeking messages influences support outcomes. The results of this initial test provide partial support for the theory. This chapter provides a discussion divided into five parts: (1) a summary of results, (2) a discussion of theoretical and pragmatic implications, (3) the identification of several research limitations, (4) an agenda for future research, and (5) a conclusion. Summary of Results The current study provides partial support for TOPSS. The current study did not support the predicted model. However, a modified alterative model (i.e., final model) demonstrated strong fit, providing support for several components of the theory. First, message directness reduces threat to positive face. Second, willingness to provide support predicts support provision. Overall, the final model showed that the directness of support seeking messages indirectly influences support provision by decreasing perceived threat to positive face thereby increasing willingness to provide support. However, neither threat to negative face nor negative affect fit in the final model. Moreover, two predicted pathways were included in the final model, despite lacking statistical significance: (a) message valence reduces threat to positive face and (b) message assertions increase threat to positive face. Implications As an initial test of the Theory of Polite Support Seeking, the current study has important theoretical, as well as practical, implications. This section begins by providing a discussion of the findings of the current study and the implications of those findings for TOPSS. Next, it presents an evaluation of TOPSS as a post-positivistic theory. Then, it identifies implications for 40 other support theory and research. Finally, this section highlights several practical implications of the current study’s findings. To begin, it is important to understand the findings of the current study and how those findings impact the future of the theory. First, the study supports the prediction that more direct support seeking messages result in lesser threat to positive face, leading to greater willingness to provide support, which predicts support provision. These findings are consistent with prior literature, showing that direct support seeking messages are more effective than indirect messages (Collins & Feeney, 2000; Derlega et al., 2003; Hanna, 2013; Hui-Jung, 2009) and that attitudes predict behaviors (Kim & Hunter, 1993a; 1993b). Second, both valence and assertions are included in the final model, although the pathways between these message dimensions and threat to face are non-significant. One explanation is that these dimensions are less influential than message directness and that the study was underpowered to identify these relationships. Another possibility is that the message inductions were weak. For example, the manipulation check demonstrates that although participants differentiated between high and low assertion messages, they did not perceive the high assertion message as high in assertion (see Table 2). Rather, on average the high assertion message was only perceived as a 3.67 on a 7-point scale. Similarly, participants did not perceive any difference between neutral valence and negative valence messages. It may be a result of these weak inductions that neither valence nor assertions influenced threat to face as predicted. For future studies, researchers should seek to develop stronger inductions by pretesting messages to ensure greater range along each dimension. Still, the appearance of valence and assertions in the final model suggests the need for further study. 41 Third, both negative affect and threat to negative face drop from the final model. An examination of the correlation matrix indicated that threat to positive face and threat to negative face were both related more closely to willingness to provide support than negative affect was related to willingness to provide support. One explanation is that the items retained from the negative PANAS measure may have focused on the wrong negative emotions. PANAS measured guilt and shame, but prior research found anger and frustration as responses to face threat that predict willingness to help (e.g., Clare et al., 2014; Rains, 2013; Weiner, 1980; Weiner, 1995). Another possibility is that the relationship between threat to face and willingness to provide support is not mediated by emotion. Rather, threat to face may directly predict willingness to provide support, as suggested by the current study. Alternatively, it may be that negative affect is difficult to induce in an online hypothetical survey, such that participants were not emotionally affected because they recognized the situation and message were not real. Similarly, the hypothetical methodology provides one explanation for why threat to negative face is not retained in the final model. Participants may not have felt their freedom was threatened by the hypothetical support seeking request. Although participants may feel like they look bad (i.e., threat to positive face) as a result of the hypothetical support seeking message, they may not feel like their freedom is threatened (i.e., threat to negative face), as the consent form indicated that the participant could end participation at any point without penalty. Another possibility is that the manipulations, which did not represent the full spectrum of message assertions and valence, were not strong enough to threaten participant’s negative face. Future tests of TOPSS should continue to examine the role of negative affect and threat to negative face, employing new methodologies and measurement to ensure that this study’s failure to support the theoretical predictions is the result of flawed theory rather than faulty methods. 42 The implications of the current study for the evaluation of TOPSS can be examined using criteria for evaluating post-positivistic theory (Berger & Chaffee, 1987; Kuhn, 1977). According to this perspective, theories should be accurate in explanatory and predictive power, internally and externally consistent, parsimonious, falsifiable, broad in scope, provocative, and fruitful. While the criteria are useful for evaluating a theory, the findings of the current study cannot assess the scope, provocation, or fruitfulness of TOPSS, as these characteristics are difficult to examine through a single empirical test. First, the findings of the current study provide a test that is useful for evaluating the accuracy of TOPSS’s explanatory and predictive power. The results of this study indicate the current theoretical framework is inaccurate. Statistically, the data did not match the theoretical model, revealing problems with the theory. However, post-hoc analyses indicate that the problems may center on the inclusion of negative affect as a mediator between threat to face and willingness to provide support, as well as the influence of message dimensions on perceived threat to negative face. Additional tests of the theory will be required to substantiate that these problems are not artifacts of the current study and its methodology, but rather inaccuracies in the underlying theory. However, although the current test suggests that TOPSS is inaccurate as a whole, many of the theorized relationships were supported. As an example, the directness of the support seeking message predicted threat to positive face, which influenced willingness to provide support and indirectly support provision. Second, the findings of the current study are useful for evaluating the consistency of TOPSS with other theories and research related to social support (i.e., external consistency). The current study provides evidence that TOPSS is consistent with existing research in establishing a request for social support as a face-threatening act (Brown & Levinson, 1987; Roloff et al., 43 1988). Additionally, findings of the current study support the relationship theorized in TOPSS between direct support seeking messages and positive support outcomes, including willingness to provide support and support provision (e.g., Collins & Feeney, 2000; Derlega et al., 2003; Hanna, 2013; Hui-Jung, 2009). Also as theorized, direct support seeking messages increase perceived threat to the potential support provider’s face, consistent with research by Wardaugh (1985). Finally, the findings of the current study show that TOPSS is consistent with research by Johnson and colleagues establishing request refusal as a response to perceived face threats (e.g., Johnson, 2007; Johnson et al., 2004a, 2004b). Although, the findings of the current study seem inconsistent with previous research linking threats to face with negative affect and negative affect with support provision, the measurement of negative affect in this study may be to blame. Accordingly, further research is needed to determine whether the relationship between threat to face and support provision is direct, mediated by negative affect, or mediated by something else. Overall, TOPSS and the findings of the study are consistent with other research. Third, the findings of the current study are useful for evaluating the parsimony of TOPSS. The failure of negative affect and threat to negative face to fit in the model suggests that TOPSS can be more parsimonious. Although these results require further substantiation, reformulations of the theory should consider that the association between support seeking message dimensions and support outcomes may be simpler than TOPSS currently predicts. For instance, continued research may demonstrate that negative affect is not an important mediator between threat to face and willingness to provide support. Finally, the findings of the current study display the ability of TOPSS to be tested and falsified. The derivation of the theory’s axioms and propositions into hypotheses demonstrates 44 the testability of TOPSS, while the failure of the theorized structural equation model validates the theory’s falsifiability. Beyond providing an opportunity to evaluate TOPSS, the current study has implications for broader theory and research. Broadly, TOPSS helps to further develop a message-centered approach to the larger social support process, consistent with Burleson’s (2009) definition of interpersonal communication, in which interpersonal communication is “a complex, situated social process in which people who have established a communicative relationship exchange messages in an effort to generate shared meanings and accomplish social goals” (p. 151). Although individually, this study does not quite accomplish Burleson’s goals for examining message exchange, the theoretical perspective contributes to a fuller vision of the social support process at large. By considering TOPSS in combination with existing social support theory, which is focused on different stages of the support process, social support researchers as a collective can embody Burleson’s definition. Moreover, TOPSS and the current study highlight the importance of messages in social support seeking efforts, which motivates the re-examination of existing theory and research. For example, Weiner’s attribution-affect-action model examines how the attributions potential support providers (i.e., helpers) form based on situational characteristics of the problem influence willingness to help. In light of the current study, it may be valuable to consider how the way a support seeker describes a problem impacts a PSP’s attributions. For instance, it may be that support seekers can verbally emphasize their efforts to alter the attributions made by the PSP, in turning influencing the PSP’s helping behavior. Considering the influence of the support seeker’s message, rather than focusing exclusively on the potential support provider’s thoughts 45 and actions, favors an interaction or transaction model of communication, rather than an overly simplified action model. TOPSS and the findings of the current study also have implications for theories and models focused on the reception phase of social support, such as the cognitive-emotional theory of esteem support messages (CETESM; Holmstrom & Burleson, 2011). CETESM predicts that highly inductive esteem support messages (i.e., messages that use questions to encourage support seekers to engage in appraisal) will be more effective than highly assertive esteem support messages (i.e., messages that use assertions to tell the support seeker how to think, feel, or behave). Yet support seekers who have directly requested advice about how to improve their self-esteem may be unsatisfied by an inductive support message. For example, a support seeker who says, “Mom, just tell me what I need to do to improve my self-esteem!” may be frustrated if the mother responds by suggesting, “What are your strongest qualities?” That support seeker might perceive that Mom did not listen to the initial request and may not perceive her message as supportive. Accordingly, although the CETESM may predict which styles of esteem support are most effective overall, one boundary condition may be support episodes when the style of support provision is a mismatch for the style of support requested. Additionally, the current study may have implications for MacGeorge’s (2001) work on the connection between support provider affect and goals to provide support. MacGeorge found that anger and sympathy influenced the types of social support individuals were willing to provide. However, the findings of the current study, which demonstrated that affect was an ineffective mediator between threat to face and willingness to provide support, suggest that it may not be negative affect, but rather face threats that drive support goals. Overall, the findings 46 of the current study provide new perspectives for researchers interested in understanding the social support process. Practically, the current study reveals that (1) the directness of support seeking messages influences support provision, and (2) close friends are likely to provide support. Most importantly, this study suggests that individuals can influence whether they receive support by using support seeking messages that are direct, clearly asking a PSP for a specific type of support. Therefore, individuals seeking support should use a direct request to increase the likelihood they will receive it. Moreover, therapists or others counseling individuals who desire support may advise using direct messages as a method of increasing support receipt. For example, a marital therapist may suggest that the spouses use clear, direct requests for social support in order to reduce threat to the other’s face and increase the likelihood of support receipt. Or, a doctor may recommend that an elderly patient struggling with daily tasks directly request assistance from family members. In addition, the current study demonstrated that most (92%) of close friends were willing to provide support in a hypothetical interaction. The willingness of close friends to provide social support is unsurprising, as Argyle and Henderson (1984) identified providing social support as an expectation within friendships. Yet the implication is important: Close friends are likely sources of social support. Limitations Despite their important theoretical and pragmatic significance, these findings are qualified by a number of study limitations. First, this study is limited by its methodology, including the hypothetical design, restriction to a single support-seeking scenario, and use of close friends as support providers. Hypothetical methodologies have limited ecological validity 47 because participants may be unable or unwilling to report how they would feel or behave when faced with the described scenario. For example, a participant may report he or she is willing to provide support due to a desire to appear likeable to the researcher (i.e., social desirability; Crowne & Marlowe, 1960; Podsakoff, MacKenzie, Lee, & Podsakoff, 2003), even though she or he might be too busy to provide support. Additionally, the use of a hypothetical scenario may be problematic for inducing perceived threats to negative face and negative affect, resulting in the failure to fit each variable in the final model. Moreover, the use of a singular scenario (i.e., seeking support for an internship) is problematic, as existing research demonstrates that characteristics about the problem for which an individual is seeking support influence whether or not a potential support provider is willing to help. As an example, the attribution-affect-action model stipulates that a PSP is more likely to provide support when the support seeker is not the cause of the problem (Weiner, 1980, 1985). The current study focuses on a hypothetical problem with the support seeker’s internship, yet seeking support for interpersonal problems may be more complicated. If the support seeker is instead seeking support about conflict with a friend who the potential support provider also knows, the degree of threat to face may increase. For example, the PSP may experience a dual threat to negative face if he or she perceives that providing support requires the PSP to choose sides in the conflict. Additionally, purporting a close friend is the hypothetical support seeker may be a limitation of the current study. As previously indicated, existing research demonstrates the expectation that friends provide one another with social support, including advice and emotional support (Argyle & Henderson, 1984). In the current study, relational closeness was high (M = 6.04 on a 7-point scale) and did not vary much between participants (SD = 0.94). Hypothetical 48 support provision also did not vary much in the current study: 91.8% (n = 224) of participants provided support to their hypothetical close friend. Although it is encouraging that directness, threat to positive face, and willingness to provide support all directly or indirectly influenced support provision in spite of the low variance of support provision, TOPSS may have greater or lesser predictive utility between support seekers and potential support providers whose relationships are less close. Still, one might ask: If close friends are apt to provide support anyway, how much does the support seeking message matter? Support seeking messages remain an important consideration in the support process for several reasons. First, there are likely differences between whether participants indicate they are willing to provide support in response to a hypothetical request and how often they actually provide support when faced with a true request. Social desirability provides one motivation for this discrepancy: Because providing support to close friends is normative, participants may be more likely to provide hypothetical support than actual support (Crowne & Marlowe, 1960). Additionally, hypothetical support is less costly to PSPs, as it requires minimal time and no opportunity for rejection. Moreover, the instructions for the study may have prompted participants to write a supportive message to complete the study, rather than to provide support. The instructions read, “Below, imagine you have the opportunity to respond to your friend’s message. Your message can be as long or short as you would like. Please try to avoid spelling errors and try to be clear in your wording. What would you say to your friend?” Participants may have interpreted the instructions as a command or request, rather than an opportunity to provide support if desired. Second, the current study relied on a broad conceptualization of social support for coding messages. It is possible that a support seeker would be more discerning. Third, although the provision of support may be high, supportive 49 messages vary in quality. The current study did not examine support quality, but TOPSS predicts that highly direct, high valence, low assertions messages will elicit better support, which may be more important to consider than merely whether support was offered. Fourth, individuals seek support within a multitude of other relationships, including from family, co-workers, physical and mental health professionals, spiritual leaders, and romantic partners, among others (Collins & Feeney, 2000; Pasch & Bradbury, 1998; Wellman & Wortley, 1990). The current study found that 92% of close friends provided hypothetical support, but further research is needed to determine whether support provision is as normative in other relationships. Finally, the opportunity to increase likelihood of support receipt, even if only in a small percentage of seeking episodes, is worthwhile. Research documenting the value of social support for one’s health, wellbeing, and relationships is abundant (Callaghan & Morrissey, 1993; Ell, 1984; Rook, 1987; Sanderson et al., 2005; Siebert et al., 1999). Accordingly, developing strategies to increase or improve the social support one receives is valuable. However, these methodological limitations suggest that future research should employ broader methodologies, including diverse scenarios and relationships, in order to better assess TOPSS. Second, this study is limited by its sample. More specifically, participants were homogenous in age and race. This homogeneity is problematic because either demographic factor could influence which support seeking behaviors are prevalent or effective. For example, research indicates that cultures deal with politeness (i.e., face) in distinct ways (Oetzel & TingToomey, 2003; Oetzel et al., 2001; Ting-Toomey et al., 1991). Therefore, future research should endeavor to include diverse participants in order to examine the consistency of these findings across broader populations. 50 Third, this study is limited because the study used dichotomized manipulations of theoretical message dimensions of directness, valence, and assertions. The researcher utilized dichotomized manipulations out of consideration for methodological parsimony; however, using dichotomized message dimensions artificially simplified the variety of messages used to seek support. This artificial simplification prevented the researcher from testing some alternative hypotheses. For example, it may be that moderately direct messages are more effective for soliciting support than either highly direct or highly indirect messages. Therefore, although the dichotomization of message dimensions allowed for the testing of the theoretical predictions, it precluded testing some other possibilities. Future research should include greater diversity of messages along each message dimension to provide a richer test of TOPSS. Finally, this study is limited by examining support seeking as a single interaction, whereas true support seeking may be an ongoing transaction, limiting the ecological validity of the study. In the current study, participants received a single support seeking message about a situation for which they had no additional background knowledge. However, support seeking may in actuality occur over time, support may be solicited multiple times for the same problem, the PSP may have additional information about the support seeker or situation, and support provision may be ongoing instead of provided through a single message. Therefore, although TOPSS aims to predict the effectiveness of support seeking messages as they occur in a natural environment, the current study tests only a sliver of that larger transaction. Future research should work towards developing a transactional perspective that encompasses the larger phenomena of social support. 51 Future Directions The purpose of the current study was to provide an initial test of the Theory of Polite Support Seeking, but the aforementioned limitations provide clear evidence there is much additional work needed to test the theory. To begin, TOPSS needs new studies to retest the initial theory using new populations, methodologies, and situations. Continued studies should also test the boundaries of TOPSS by examining relational, contextual, and individual factors that may influence the social support process. For example, Weiner’s (1980) attribution-emotion-intention model identifies sympathy as an emotion influencing whether a potential support provider provides support. Future studies may choose to examine affect more broadly to determine which aspects of affect best explain the connection between threat to PSP face and support outcomes. Moreover, TOPSS requires additional theory tests to overcome the limitations of the current study. A better test of TOPSS accomplish several goals: First, a better study would address the measurement issues with the current study, primarily by measuring negative affect in a clearer manner. Second, it would utilize the full range of message dimensions, instead of restricting messages to two- or three-levels of variance. Third, a better test would rely on a nonhypothetical paradigm to increase the perceived cost of providing support, as well as providing increased external validity. Fourth, it would allow the support seeker and potential support provider to interact face-to-face, in order to increase the potential for face threat. Fifth, a better study would examine a broader range of support seeking topics, including problems that vary in severity and issue. Finally, a better study would test the full theory, including examining the role of positive affect and whether TOPSS predicts support quality in addition to willingness to provide support and support provision. 52 These additional tests will help the theorist to determine whether the failure of the model in the current study lies in problems with the current data set or problems within the theory itself. Should the model continue to fail, an important future direction will be revisiting the problematic elements to determine if modifications can be made to produce a theory of support seeking that provides an explanatory, predictive, or control function. First, TOPSS can serve an explanatory function by explaining why some messages yield support while others are ineffective. Second, TOPSS can serve a predictive function by identifying which messages are most likely to threaten face, resulting in reduced likelihood of receiving support. Third, TOPSS can serve a control function by revealing ways to modify support seeking messages to increase effectiveness. Additionally, researchers should continue to test the remaining components of the model, namely testing the impact of support seeking messages on received support quality. Next, researchers should begin to test the boundary conditions of the theory, including variations in contextual and relational variables. For example, researchers can examine whether the construction of support seeking messages differentially impacts provision of different types of support or whether norms in certain relationships overpower the influence of support seeking message characteristics. Moreover, future research should consider whether a theory that predicts which messages will lead to the receipt of quality support could be applied to broader social influence interactions. Although support seeking is a distinct form of social influence, the mechanism by which the message characteristics directness, valence, and assertion impact support outcomes (i.e., politeness) may similarly operate in other persuasive attempts, such as convincing someone to stop smoking or to purchase a new car. 53 Conclusion The Theory of Polite Support Seeking provides a new lens through which to examine social support interactions, drawing from existing theory and research, situating support seeking as a process of social influence. Whereas many existing models focus on social support providers and recipients, TOPSS acknowledges the important role that support seekers may play in the process. Furthermore, TOPSS’s focus on support seekers reinforces the message-centered approach to interpersonal communication within the domain of social support research. Although the current study does not fully support the Theory of Polite Support Seeking, the study indicates that the theory has potential. Many of the theoretically predicted relationships significantly influence outcomes, and a post-hoc model suggests that a simplified model may have greater predictive utility. Most importantly, the study provides a starting point and many suggestions for additional theoretical tests of TOPSS. 54 APPENDICES 55 APPENDIX A Tables and Figures Table A1. Theory of Polite Support Seeking (TOPSS): Axioms & Propositions Axiom 1 Social support seeking may be conceptualized as an active, intentional, interpersonal process designed to elicit support through the exchange of messages. Axiom 2 Social influence is a central goal of social support seeking, such that support seekers strive to influence the behavior of potential support providers. Axiom 3 Social support seeking messages vary in the degree of directness, valence, and assertion, which influences support outcomes. Axiom 4 Messages may simultaneously employ multiple politeness strategies. Axiom 5 Social support seeking may threaten the positive face and negative face of both the support seeker and potential support provider. Axiom 6 In social support seeking episodes, multiple threats to face may simultaneously occur. Axiom 7 The degree to which support seeking messages are direct, valenced, and include assertions influences the degree to which the message threatens face. Axiom 8 Politeness and efficiency are compatible meta-goals in the support seeking process. Axiom 9 Additional contextual, relational, and individual factors related to the support seeker, potential support provider, and seeking interaction will influence the support process. Proposition 1 The more direct the support seeking message, the greater the threat to the potential support provider’s negative face, but lesser the threat to the potential support provider’s positive face. Proposition 2 The more positive the valence of the support seeking message, the more it minimizes the intrinsic face threat of a request, whereas the less positive the message, the greater the threat to the potential support provider’s positive face. Proposition 3 The lower the assertion of the support seeking message, the more effective, whereas the higher the assertion of the support seeking message, the greater the threat to the potential support provider’s positive and negative face. Proposition 4 Increased degree of threat to the potential support provider’s positive face or negative face decreases the likelihood of support provision (i.e., request compliance). Proposition 5 The potential support provider’s affect mediates the relationship between threat to face and support outcomes, including willingness to provide support, support provision, support type, and support quality, such that negative affect decreases support outcomes, whereas positive affect increases support outcomes. 56 Table A2. Means and Standard Deviations P. Directness P. Valence P. Assertions Negative Affect Threat to Pos. Face Threat to Neg. Face Willingness Relational Closeness All 5.59 (1.00)* 4.57 (1.37)* 3.45 (1.48)* 1.89 (1.31)* 2.77 (1.21)* 2.53 (1.27)* 5.92 (0.96)* 6.04 (0.94)* Direct 5.39 (0.97)* 4.23 (1.39) 3.30 (1.46)* 1.93 (1.27)* 3.00 (1.24)* 2.60 (1.29)* 5.76 (0.97)* 6.01 (0.79)* Indirect 5.80 (1.00)* 4.93 (1.26)* 3.61 (1.50)* 1.85 (1.35)* 2.53 (1.13)* 2.45 (1.25)* 6.08 (0.93)* 5.98 (1.08)* Pos. Valence 5.57 (1.16)* 5.02 (1.24)* 3.67 (1.49)* 1.85 (1.35)* 2.60 (1.25)* 2.50 (1.35)* 5.95 (0.99)* 5.94 (1.14)* Neutral Valence 5.63 (0.91)* 4.18 (1.44) 3.38 (1.48)* 2.01 (1.29)* 2.78 (1.05)* 2.43 (1.17)* 5.99 (0.92)* 6.12 (0.81)* Neg. Valence 5.55 (0.93)* 4.52 (1.32)* 3.30 (1.47)* 1.82 (1.30)* 2.92 (1.27)* 2.65 (1.27)* 5.80 (0.98)* 6.07 (0.83)* High Assertion 5.54 (1.11)* 4.51 (1.47)* 3.67 (1.54)* 1.78 (1.27)* 2.87 (1.17)* 2.46 (1.30)* 5.91 (0.98)* 6.03 (1.02)* Low Assertion 5.64 (0.89)* 4.63 (1.28)* 3.23 (1.39)* 2.01 (1.34)* 2.67 (1.24)* 2.59 (1.24)* 5.92 (0.95)* 6.06 (0.86)* Notes: Rows represent measured variables, whereas columns represent categorized message dimensions. An asterisk (*) indicates that the mean varied significantly from the midpoint of the scale (i.e., 4). 57 Table A3. Correlation Matrix D Directness (D) Valence (V) Assertions (A) Threat to Positive Face (P) Threat to Negative Face (N) Negative Affect (NA) Willingness to Provide Support (W) Support Provision (SP) Relational Closeness (RC) V A 1 .01 -.01 -.20** -.06 -.03 .16* P N NA 1 -.01 -.11 -.05 .01 .06 .09 -.06 .04 -.05 W SP 1 .08 -.05 -.09 -.01 1 .72** .50** -.50** 1 .49** -.47** 1 -.28** 1 .03 -.03 -.23** -.23** -.26** -.30** -.20** -.28** .41** .51** 1 .24** Note: All reported values represent a Pearson’s product moment correlation (r). Items with * are significant at p < .05, two-tailed. Items with ** are significant at p < .01, two-tailed. 58 RC 1 Table A4. Unstandardized, Standardized, and Significance Levels for Predicted Model Parameter Estimate Unstandardized Standardized Directness  Threat to Positive Face -0.47 -0.20 Directness  Threat to Negative Face -0.16 -0.06 Valence  Threat to Positive Face -0.16 -0.11 Assertions  Threat to Positive Face 0.19 0.08 Assertions  Threat to Negative Face -0.13 -0.05 Threat to Positive Face  Negative Affect 0.32 0.31 Threat to Negative Face  Negative Affect 0.29 0.30 Negative Affect  Willingness to Provide Support -0.20 -0.26 Willingness to Provide Support  Support Provision .11 0.39 2 Note: χ (16) = 252.77, p < .01; RMSEA = 0.25 (90% CI = 0.22 – 0.27); CFI = 0.39 59 p .002 .33 .08 .21 .41 .001 .001 .001 .001 Table A5. Unstandardized, Standardized, and Significance Levels for Final Model Parameter Estimate Unstandardized Standardized Directness  Threat to Positive Face -0.47 -0.20 Valence  Threat to Positive Face -0.16 -0.11 Assertions  Threat to Positive Face 0.19 0.08 Threat to P. Face  Willingness to Provide Support -.39 -.49 Willingness to Provide Support  Support Provision .11 0.39 Note: χ2 (7) = 3.76, p = .81; RMSEA = 0.00 (90% CI = 0.00 – 0.05); CFI = 1.00 60 p .002 .08 .21 .001 .001 Figure A1. A Theory of Polite Support Seeking (TOPSS) 61 Figure A2. Initial test of a Theory of Polite Support Seeking (TOPSS) 62 Figure A3. TOPSS Predicted Model Notes: All parameter estimates are standardized; unstandardized estimates and significance levels are presented in Table 3. * indicates pathways that are significant at p < .01, whereas ** indicates pathways that are significant at p < .001. 63 Figure A4. TOPSS Final Model Notes: All parameter estimates are standardized; unstandardized estimates and significance levels are presented in Table 4. * indicates pathways that are significant at p < .01, whereas ** indicates pathways that are significant at p < .001. 64 APPENDIX B Message Set A: Roommate Problems Positive valence/highly direct/high assertion I’m so frustrated with my roommate Casey. Casey never does the dishes or pays rent on time and invites people over on nights when I’m trying to study. You’re taking a communication class and must have strong interpersonal skills. Tell me what I should do! Neutral valence/highly direct/high assertion I’m so frustrated with my roommate Casey. Casey never does the dishes or pays rent on time and invites people over on nights when I’m trying to study. Tell me what I should do! Negative valence/highly direct/high assertion I’m so frustrated with my roommate Casey. Casey never does the dishes or pays rent on time and invites people over on nights when I’m trying to study. It’s not like you have any expertise in this area, but tell me what I should do! Positive valence/lowly direct/high assertion I’m so frustrated with my roommate Casey. Casey never does the dishes or pays rent on time and invites people over on nights when I’m trying to study. You’re taking a communication class and must have strong interpersonal skills. Don’t be a bad roommate! Neutral valence/lowly direct/high assertion I’m so frustrated with my roommate Casey. Casey never does the dishes or pays rent on time and invites people over on nights when I’m trying to study. Don’t be a bad roommate! Negative valence/lowly direct/high assertion I’m so frustrated with my roommate Casey. Casey never does the dishes or pays rent on time and invites people over on nights when I’m trying to study. It’s not like you have any expertise in this area. Don’t be a bad roommate! Positive valence/highly direct/low assertion I’m so frustrated with my roommate Casey. Casey never does the dishes or pays rent on time and invites people over on nights when I’m trying to study. You’re taking a communication class and must have strong interpersonal skills. What do you think I should do? Neutral valence/highly direct/low assertion I’m so frustrated with my roommate Casey. Casey never does the dishes or pays rent on time and invites people over on nights when I’m trying to study. What do you think I should do? Negative valence/highly direct/low assertion I’m so frustrated with my roommate Casey. Casey never does the dishes or pays rent on time and invites people over on nights when I’m trying to study. It’s not like you have any expertise in this area, but what do you think I should do? 65 Positive valence/lowly direct/low assertion I’m so frustrated with my roommate Casey. Casey never does the dishes or pays rent on time and invites people over on nights when I’m trying to study. You’re taking a communication class and must have strong interpersonal skills. Would you be upset if you were in my position? Neutral valence/lowly direct/low assertion I’m so frustrated with my roommate Casey. Casey never does the dishes or pays rent on time and invites people over on nights when I’m trying to study. Would you be upset if you were in my position? Negative valence/lowly direct/low assertion I’m so frustrated with my roommate Casey. Casey never does the dishes or pays rent on time and invites people over on nights when I’m trying to study. It’s not like you have any expertise in this area, but would you be upset if you were in my position? 66 APPENDIX C Message Set B: Internship Frustration Positive valence/highly direct/high assertion I hate my internship. It seemed like a dream job when I started, but I’m having a hard time getting along with the other interns. Plus, my boss is disappointed in my performance. You’re taking a communication class, so you must have strong professional skills. Tell me what I should do! Neutral valence/highly direct/high assertion I hate my internship. It seemed like a dream job when I started, but I’m having a hard time getting along with the other interns. Plus, my boss is disappointed in my performance. Tell me what I should do! Negative valence/highly direct/high assertion I hate my internship. It seemed like a dream job when I started, but I’m having a hard time getting along with the other interns. Plus, my boss is disappointed in my performance. It’s not like you have any expertise in this area, but tell me what I should do! Positive valence/lowly direct/high assertion I hate my internship. It seemed like a dream job when I started, but I’m having a hard time getting along with the other interns. Plus, my boss is disappointed in my performance. You’re taking a communication class, so you must have strong professional skills. Think outside the box! Neutral valence/lowly direct/high assertion I hate my internship. It seemed like a dream job when I started, but I’m having a hard time getting along with the other interns. Plus, my boss is disappointed in my performance. Think outside the box! Negative valence/lowly direct/high assertion I hate my internship. It seemed like a dream job when I started, but I’m having a hard time getting along with the other interns. Plus, my boss is disappointed in my performance. It’s not like you have any expertise in this area. Think outside the box! Positive valence/highly direct/low assertion I hate my internship. It seemed like a dream job when I started, but I’m having a hard time getting along with the other interns. Plus, my boss is disappointed in my performance. You’re taking a communication class, so you must have strong professional skills. What do you think I should do? Neutral valence/highly direct/low assertion I hate my internship. It seemed like a dream job when I started, but I’m having a hard time getting along with the other interns. Plus, my boss is disappointed in my performance. What do you think I should do? 67 Negative valence/highly direct/low assertion I hate my internship. It seemed like a dream job when I started, but I’m having a hard time getting along with the other interns. Plus, my boss is disappointed in my performance. It’s not like you have any expertise in this area, but what do you think I should do? Positive valence/lowly direct/low assertion I hate my internship. It seemed like a dream job when I started, but I’m having a hard time getting along with the other interns. Plus, my boss is disappointed in my performance. You’re taking a communication class, so you must have strong professional skills. Are you applying for internships? Neutral valence/lowly direct/low assertion I hate my internship. It seemed like a dream job when I started, but I’m having a hard time getting along with the other interns. Plus, my boss is disappointed in my performance. Are you applying for internships? Negative valence/lowly direct/low assertion I hate my internship. It seemed like a dream job when I started, but I’m having a hard time getting along with the other interns. Plus, my boss is disappointed in my performance. It’s not like you have any expertise in this area, but are you applying for internships? 68 APPENDIX D Manipulation Checks Notes: Each item is anchored on a 7-point Likert-type scale where 1 = Strongly disagree, 4 = Neither agree nor disagree, and 7 = Strongly agree. * indicates a reverse-coded item. Bolded items were retained for final analyses. Message Directness [4 items included in SEM analyses] 1. To what extent does the message make a clear request for social support? 2. To what extent does the message tell you what this person wants? 3. To what extent does the message suggest that you help this person? 4. To what extent does the message indicate this person wants your assistance? 5. To what extent does the message make it obvious what this person is hoping for? 6. To what extent does the message have a clear purpose? Message Valence [4 items included in SEM analyses] 1. To what extent does the message compliment you? 2. To what extent does the message express gratitude toward you? 3. To what extent does the message make you feel valued or appreciated? 4. To what extent does the message criticize you?* 5. To what extent does the message praise you? 6. To what extent does the message indicate that this person thinks highly of you or your skills? Message Assertions [6 items included in SEM analyses] 1. To what extent does the message make a demanding request? 2. To what extent does the message tell you what to do? 3. To what extent does the message insist you behave a specific way? 4. To what extent does the message make assert that you do something? 5. To what extent does the message instruct you how to behave? 6. To what extent does the message command you to do something? 69 APPENDIX E Threat to Positive Face Notes: Each item is anchored on a 7-point Likert-type scale where 1 = Strongly disagree, 4 = Neither agree nor disagree, and 7 = Strongly agree. * indicates a reverse-coded item, whereas ~ indicates an item created by the researcher, independent of Cupach & Carson’s original items. Bolded items were retained for final analyses. [4 items included in SEM analyses] 1. This person’s actions were polite. * 2. This person’s actions were rude. 3. This person’s actions were insensitive. 4. This person’s actions showed disrespect toward me. 5. This person’s actions were justified. * 6. This person’s actions were hostile. 7. This person’s actions showed contempt towards me. 8. This person’s actions were tactful. * 9. This person’s actions showed disregard for me. ~ 10. This person’s actions were respectful. *~ 70 APPENDIX F Threat to Negative Face Notes: Each item is anchored on a 7-point Likert-type scale where 1 = Strongly disagree, 4 = Neither agree nor disagree, and 7 = Strongly agree. ~ indicates the researcher created this item, independent of Cupach & Carson’s original items. Bolded items were retained for final analyses. [7 items included in SEM analyses] 1. This person’s actions constrained my choices. 2. This person’s actions took away some of my independence. 3. This person’s actions made me look bad in the eyes of others. 4. This person’s actions invaded my privacy. 5. This person’s actions prevented me from doing what I wanted to do. ~ 6. This person’s actions made me feel trapped. ~ 7. This person’s actions stopped me from making my own choices. ~ 8. This person’s actions caused me to behave a specific way. ~ 9. This person’s actions allowed me to do whatever I wanted. ~ 10. This person’s actions made me feel powerless. ~ 71 APPENDIX G Willingness to Provide Support Note: Each item is anchored on a 7-point Likert-type scale where 1 = Strongly disagree, 4 = Neither agree nor disagree, and 7 = Strongly agree. Bolded items were retained for final analyses. General Willingness to Provide Support [7 items included in SEM analyses] 1. I would like to help this person. 2. I would like to provide support to this person. 3. I would like to send a message to this person. 4. I would like to have the opportunity to say something positive to this person. 5. I would like to spend some time helping this person. 6. I would like to help this person any way I am able. 7. I would like to do what I could to help this person. 72 APPENDIX H Relational Closeness Note: Each item is anchored on a 7-point Likert-type scale where 1 = Strongly disagree, 4 = Neither agree nor disagree, and 7 = Strongly agree. Bolded items were retained for final analyses. [8 items included in SEM analyses] 1. My relationship with my friend is close. 2. When we are apart, I miss my friend a great deal. 3. My friend and I disclose important personal things to each other. 4. My friend and I have a strong connection. 5. My friend and I want to spend time together. 6. I’m sure of my relationship with my friend. 7. My friend is a priority in my life. 8. My friend and I do a lot of things together. 9. When I have free time I choose to spend it alone with my friend. 10. I think about my friend a lot. 11. My relationship with my friend is important in my life. 12. I consider my friend when making important decisions. 73 REFERENCES 74 REFERENCES Afifi, W. A., & Afifi, T. D. (2009). Avoidance among adolescents in conversations about their parents' relationship: Applying the theory of motivated information management. Journal of Social and Personal Relationships, 26, 488-511. doi: 10.1177/0265407509350869 Afifi, W. A., & Weiner, J. L. (2006). Seeking information about sexual health: Applying the theory of motivated information management. Human Communication Research, 32, 3557. doi: 10.1111/j.1468-2958.2006.00002.x Argyle, M., & Henderson, M. (1984). The rules of friendship. Journal of Social and Personal Relationships, 1, 211-237. doi: 10.1177/0265407584012005 Barbee, A. P., & Cunningham, M .R. (1995). An experimental approach to social support communications: Interactive coping in close relationships. In B. R. Burleson (Ed.), Communication Yearbook, 18 (pp. 381-413). Thousand Oaks, CA: Sage. Barbee, A. P., Cunningham, M. R, Winstead, B. A., Derlega, V. J., Gulley, M. R., Yankeelov, P. A., & Druen, P. B. (1993). Effects of gender role expectations on the social support process. Journal of Social Issues, 49(3), 175-190. doi: 10.1111/j.15404560.1993.tb01175.x Barbee, A. P., Gulley, M. R., & Cunningham, M. R. (1990). Support seeking in personal relationships. Journal of Social and Personal Relationships, 7, 531-540. doi: 10.1177/0265407590074009 Barry, B., & Shapiro, D. L. (1992). Influence tactics in combination: The interactive effects of soft versus hard tactics and rational exchange. Journal of Applied Social Psychology, 22(18), 1429-1441. doi: 10.1111/j.1559-1816.1992.tb00958.x Baxter, L. A. (1984). An investigation of compliance-gaining as politeness. Human Communication Research, 10, 427-456. doi: 10.1111/j.1468-2958.1984.tb00026.x Bentler, P. M., & Bonett, D. G. (1980). Significance tests and goodness of fit in the analysis of covariance structures. Psychological Bulletin, 88, 588-606. doi: 10.1037/00332909.88.3.588 Berger, C., & Chaffee, S. H. (1987). What communication scientists do. In C. Berger & S. H. Chaffee (Eds.), Handbook of communication science (pp. pp 99-122). Newbury Park, CA: Sage. Berkowitz, L. (1987). Mood, self-awareness, and willingness to help. Journal of Personality and Social Psychology, 52, 721-729. doi: 10.1037/0022-3514.52.4.721 75 Bodie, G. D., Burleson, B. R., Holmstrom, A. J., McCullough, J. D., Rack, J. J., Hanasono, L. K., & Rosier, J. G. (2011). Effects of cognitive complexity and emotional upset on processing supportive messages: Two tests of a dual-process theory of supportive communication outcomes. Human Communication Research, 37, 350-376. doi: 10.1111/j.1468-2958.2011.01405.x Brett, J. M., Olekalns, M., Friedman, R., Goates, N., Anderson, C., & Lisco, C. C. (2007). Sticks and stones: Language, face, and online dispute resolution. Academy of Management, 50, 85-99. doi: 10.5465/AMJ.2007.24161853 Brown, P., & Levinson, S. (1987). Politeness: Some universals in language usage. New York: Cambridge University Press. Brunstein, J. C., Dangelmayer, G., & Schultheiss, O. C. (1996). Personal goals and social support in close relationships: Effects on relationship mood and marital satisfaction. Journal of Personality and Social Psychology, 71, 1006-1019. doi: 10.1037/00223514.71.5.1006 Burleson, B. R. (1994). Comforting messages: Features, functions, and outcomes. In J. A. Daly & J. M. Wiemann (Eds.), Strategic interpersonal communication (pp. 135-161). Hillsdale, NJ: Erlbaum. Burleson, B. R. (2009). The nature of interpersonal communication: A message-centered approach. In C. R. Berger, M. E. Roloff & D. R. Roskos-Ewoldson (Eds.), The handbook of communication science (2nd ed., pp. 143-163). Thousand Oaks, CA: Sage. Burleson, B. R., Hanasono, L. K., Bodie, G. D., Holmstrom, A. J., Rack, J. J., Rosier, J. G., & McCullough, J. D. (2009). Explaining gender differences in responses to supportive messages: Two tests of a dual-process approach. Sex Roles, 61, 265-280. doi: 10.1007/s11199-009-9623-7 Callaghan, P., & Morrissey, J. (1993). Social support and health: A review. Journal of Advanced Nursing, 18, 203-210. doi: 10.1046/j.1365-2648.1993.18020203.x Carlson, M., Charlin, V., & Miller, N. (1988). Positive mood and helping behavior: A test of six hypotheses. Interpersonal Relations and Group Processes, 55, 211-229. doi: 10.1037/0022-3514.55.2.211 Carson, C. L., & Cupach, W. R. (2000). Facing corrections in the workplace: The influence of perceived face threat on the consequences of managerial approaches. Journal of Applied Communication, 28, 215-234. doi: 10.1080/00909880009365572 Cauce, A. M., Domenech-Rodriguez, M., Paradise, M., Cochran, B. N., Munyi Shea, J., Srebnik, D., & Baydar, N. (2002). Culutural and contextual influences in mental health help seeking: A focus on ethnic minority youth. Journal of Consulting and Clinical Psychology, 70, 44-55. doi: 10.1037//0022-006X.70.1.44 76 Clare, D., Hanna, A. A., Holmstrom, A. J., & Stolz, E. (2013). Job seekers' job search behaviors and the willingness of others to provide social support. Paper presented at the National Communication Association Annual Meeting, Washington, D. C. Clare, D., Hanna, A. A., Holmstrom, A. J., Summers, M., & Musatics, C. (2014). Job seekers' behaviors and the willingness of others to provide esteem support: The role of attributions, emotions, and efficacy. Paper presented at the Annual Meeting of the International Communication Association, Seattle, WA. Clark, M. S., & Isen, A. M. (1982). Toward understanding the relationship between feeling states and social behavior. In A. Hastorf & A. M. Isen (Eds.), Cognitive social psychology (pp. 73-108). New York, NY: Elsevier. Collins, N. L., & Feeney, B. C. (2000). A safe haven: An attachment theory perspective on support seeking and caregiving in intimate relationships. Journal of Personality and Social Psychology, 78, 1053-1073. doi: 10.1037//0022-3514.78.6.1053 Craig, R. T., Tracey, K., & Spisak, F. (1986). The discourse of requests: Assessment of a politeness approach. Human Communication Research, 12, 437-468. doi: 10.1111/j.14682958.1986.tb00087.x Cupach, W. R., & Carson, C. L. (2002). Characteristics and consequences of interpersonal compliants associated with perceived face threat. Journal of Social and Personal Relationships, 19, 443-462. doi: 10.1177/0265407502019004047 Cutrona, C. E. (2012). Recent advances in research on social support in couples: Insights from event-sampling and daily diary research. In P. Noller & G. C. Karantzas (Eds.), The Wiley-Blackwell handbook of couples and family relationships. Oxford, UK: WileyBlackwell. Cutrona, C. E., Shaffer, P. A., Wesner, K. A., & Gardner, K. A. (2007). Optimally matching support and perceived spousal sensitivity. Journal of Family Psychology, 21, 754-758. doi: 10.1037/0893-3200.21.4.754 Cutrona, C. E., & Suhr, J. A. (1992). Controllability of stressful events and satisfaction with spouse support behaviors. Communication Research, 19, 154-174. doi: 10.1177/009365092019002002 Cutrona, C. E., Suhr, J. A., & MacFarlane, R. (1990). Interpersonal transactions and the psychological sense of support. In S. Duck & R. Silver (Eds.), Personal relationships and social support (pp. 30-45). London: Sage. Derlega, V. J., Winstead, B. A., Oldfield III, E. C., & Barbee, A. P. (2003). Close relationships and social support in coping with HIV: A test of sensitive interaction systems theory. AIDS and Behavior, 7, 119-129. doi: 1090-7165/03/0600-0119/0 77 Dibble, J. L., Levine, T. R., & Park, H. S. (2011). The unidimensional relationship closeness scale (URCS): Reliability and validity evidence for a new measure of relationship closeness Psychological Assessment, 24, 565-572. doi: 10.1037/a0026265 Dillard, J. P., Wilson, S. R., Tusing, K. J., & Kinney, T. A. (1997). Politeness judgments in personal relationships. Journal of Language and Social Psychology, 16, 297-325. doi: 10.1177/0261927X970163003 Ell, K. (1984). Social networks, social support, and health status: A review. Social Service Review, 58(1), 133-149. doi: Retrieved from http://www.jstor.org/stable/30011713 Flom, P. L., Friedman, S. R., Kottiri, B. J., Neaigus, A., Curtis, R., Des Jarlais, D. C., . . . Zenilman, J. M. (2001). Stigmatized drug use, sexual partner concurrency, and other sex risk network and behavior characteristics of 18- to 24-year-old youth in a high-risk neighborhood. Sexually Transmitted Diseases, 28, 598-607. doi: 10.1097/00007435200110000-00006 Folkes, V. S. (1982). Communicating the reasons for social rejection. Journal of Experimental Social Psychology, 18, 618-628. doi: 10.1016/0022-1031(82)90052-X Frasure-Smith, N., Lespérance, F., Gravel, G., Masson, A., Juneau, M., Talajic, M., & Bourassa, M. G. (2000). Social support, depression, and mortality during the first year after myocardial infarction. Circulation, 101, 1919-1924. doi: 10.1161/01.CIR.101.16.1919 George, J. M. (1991). State or trait: Effects of positive mood on prosocial behavior at work. Journal of Applied Psychology, 76, 299-307. doi: 10.1037/0021-9010.76.2.299 Goffman, E. (1967). 'On face-work.' Interaction ritual: Essays on face-to-face behavior (pp. 545). New York, NY: Pantheon Books. Goldsmith, D. J. (1992). Managing conflict goals in supportive interaction: An integrative theoretical framework. Communication Research, 19, 264-286. doi: 10.1177/009365092019002007 Goldsmith, D. J., McDermott, V. M., & Alexander, S. C. (2000). Helpful, supportive and sensitive: Measuring the evaluation of enacted social support in personal relationships. Journal of Social and Personal Relationships, 17, 369-391. doi: 10.1177/0265407500173004 Gourash, N. (1978). Help-seeking: A review of the literature. American Journal of Community Psychology, 6(5), 413-423. doi: 10.1007/BF00941418 Grice, H. P. (1989). Studies in the way of words. Cambridge, MA: Harvard University Press. 78 Hanna, A. A. (2013). Social support seeking: The dimensions and goals of support seeking messages. Preliminary dissertation. Department of Communication. Michigan State University. Hawker, S., & Kerr, C. (2007). Doing grounded theory. In E. Lyons & A. Coyle (Eds.), Analysing qualitative data in psychology (pp. 87-97). Thousand Oaks, CA: SAGE. Holmstrom, A. J. (2012). What helps - and what doesn't - when self-esteem is threatened?: Retrospective reports of esteem support. Communication Studies, 63, 77-98. doi: 10.1080/10510974.2011.586399 Holmstrom, A. J., & Burleson, B. R. (2011). An initial test of a cognitive-emotional theory of esteem support messages. Communication Research, 38, 326-355. doi: 10.1177/0093650210376191 Holtgraves, T. (1991). Interpreting questions and replies: Effects of face-threat, question form, and gender. Social Psychology Quarterly, 54, 15-24. doi: 10.2307/2786785 House, J. S. (1981). Work stress and social support. Reading, MA: Addison Wesley Publishing Company. Hu, L., & Bentler, P. M. (1999). Cutoff criteria for fit indices in covariance structure analysis: Conventional criteria versus new alternative. Structural Equation Modeling: A Multidisciplinary Journal, 6, 1-55. doi: 10.1080/10705519909540118 Hui-Jung, C. (2009). Online social support: Which posts were answered? Journal of Contemporary Eastern Asia, 8, 31-46. doi: Retrieved from http://eastasia.yu.ac.kr/documents/Chang_Hui_Jung.pdf Isen, A. M. (1987). Positive affect, cognitive processes, and social behavior. In L. Berkowitz (Ed.), Advances in experimental social psychology (Vol. 20, pp. 203-253). New York, NY: Academic. Jensen, S. L. (2001). Development of a scale to code the elicitation of social support. (Doctor of Philosophy), Iowa State University, Ames, Iowa. (3034191) Johnson, D. I. (2007). Politeness theory and conversational refusals: Associations between various types of face threat and perceived competence. Western Journal of Communication, 71, 196-215. doi: 10.1080/10570310701518427 Johnson, D. I., Roloff, M. E., & Riffee, M. A. (2004a). Politeness theory and refusals of requests: Face threat as a function of expressed obstacles. Communication Studies, 55, 227-238. doi: 10.1080/10510980409388616 79 Johnson, D. I., Roloff, M. E., & Riffee, M. A. (2004b). Responses to refusals of requests: Face threat and persistence, persuasion, and forgiving statements. Communication Quarterly, 52, 347-356. doi: 10.1080/01463370409370205 Kellermann, K., & Shea, B. C. (1996). Threats, suggestions, hints, and promises: Gaining compliance efficiently and politely. Communication Quarterly, 44, 145-165. doi: 10.1080/01463379609370007 Kennedy-Lightsey, C. D. (2010). Recognizing contributions: Face-support and face-threat influences students' emotional and communicative responses. Communication Research Reports, 27, 20-29. doi: 10.1080/08824090903526588 Kim, M.-S., & Hunter, J. E. (1993a). Attitude-behavior relations: A meta-analysis of attitudinal relevance and topic. Journal of Communication, 43, 101-142. doi: 10.1111/j.14602466.1993.tb01251.x Kim, M.-S., & Hunter, J. E. (1993b). Relationships among attitudes, behavioral intentions, and behavior: A meta-analysis of past research, part 2. Communication Research, 20, 331364. doi: 10.1177/009365093020003001 Kroenke, C. H., Kubzansky, L. D., Schernhammer, E. S., Holmes, M. D., & Kawachi, I. (2006). Social networks, social support, and survival after breast cancer diagnosis. Journal of Clinical Oncology, 24, 1105-1111. doi: 10.1200/JCO.2005.04.2846 Kuhn, T. (1977). The essential tension. Chicago, IL: Chicago University Press. LaPiere, R. T. (1934). Attitudes vs. actions. Social Forces, 13, 230-237. doi: 10.1093/ije/dyp398 Lawrence, E., Bunde, M., Barry, R. A., Brock, R. L., Sullivan, K. T., Pasch, L. A., . . . Adams, E. E. (2008). Partner support and marital satisfaction: Support amount, adequacy, provision, and solicitation. Personal Relationships, 15, 445-463. doi: j.1475-6811.2008.00209.x Ledbetter, A. M., Mazer, J. P., DeGroot, J. M., Meyer, K. R., Mao, Y., & Swafford, B. (2011). Attitudes toward online social connection and self-disclosure as predictors of Facebook communication and relational closeness. Communication Research, 38, 27-53. doi: 10.1177/0093650210365537 Levine, T. R., & Boster, F. J. (2001). The effects of power and message variables on compliance. Communication Monographs, 68, 28-48. doi: 10.1080/03637750128049 Liang, B., Goodman, L., Tummala-Narra, P., & Weintraub, S. (2005). A theoretical framework for understanding help-seeking processes among survivors of intimate partner violence. American Journal of Community Psychology, 36, 71-84. doi: 10.1007/s10464-005-62336 80 Lim, T.-S., & Bowers, J. W. (1991). Facework solidarity, approbation, and tact. Human Communication Research, 17, 415-450. doi: 10.1111/j.1468-2958.1991.tb00239.x MacGeorge, E. L. (2001). Support providers' interaction goals: The influence of attributions and emotions. Communication Monographs, 68, 72-97. doi: 10.1080/03637750128050 Metts, S. (1997). Face and facework: Implications for the study of personal relationships. In S. Duck (Ed.), Handbook of personal relationships (2nd ed., pp. 373-390). Chichester, NY: John Wiley & Sons. Meyer, J. P., & Mulherin, A. (1980). From attribution to helping: An analysis of the mediating effects of affect and expectancy. Journal of Personality and Social Psychology, 39(201210). doi: 10.1037/0022-3514.39.2.201 Mortenson, S. T. (2009). Interpersonal trust and social skill in seeking social support among Chinese and Americans. Communication Research, 36, 32-53. doi: 10.1177/093650208326460 Nabi, R. L. (2010). The case for emphasizing discrete emotions in communication research. Communication Monographs, 77, 153-159. doi: 10.1080/03637751003790444 Norberg, A. L., Lindblad, F., & Boman, K. K. (2006). Support-seeking, perceived support, and anxiety in mothers and father after children's cancert treatment. Psycho-Oncology, 15, 335-343. doi: 10.1002/pon.960 Oetzel, J. G., & Ting-Toomey, S. (2003). Face concerns in interpersonal conflict: A crosscultural empirical test of the face negotiation theory. Communication Research, 30, 599624. doi: 10.1177/0093650203257841 Oetzel, J. G., Ting-Toomey, S., Masumoto, T., Yokochi, Y., Pan, X., Takai, J., & WIlcox, R. (2001). Face and facework in conflict: A cross-cultural comparison of China, Germany, Japan, and the United States. Communication Monographs, 68, 235-258. doi: 10.1080/03637750128061 Pasch, L. A., & Bradbury, T. N. (1998). Social support, conflict, and the development of marital dysfunction. Journal of Consulting and Clinical Psychology, 66, 219-230. doi: 10.1037/0022-006X.66.2.219 Rains, S. A. (2013). The nature of psychological reactance revisited: A meta-analytic review. Human Communication Research, 39, 47-73. doi: 10.1111/j.1468-2958.2012.01443.x Reisenzein, R. (1986). A structural equation analysis of Weiner's attribution-affect model of helping behavior. Journal of Personality and Social Psychology, 38, 291-300. doi: 10.1037/0022-3514.50.6.1123 81 Roloff, M. E. (1987). Communication and reciprocity within romantic relationships. In M. E. Roloff & G. R. Miller (Eds.), Interpersonal processes: New directions in communication research (pp. 11-38). Newbury Park, CA: Sage. Roloff, M. E., Janiszewski, C. A., McGrath, M. A., Burns, C. S., & Manrai, L. A. (1988). Acquiring resources from intimates: When obligation substitutes for persuasion. Human Communication Research, 14, 364-396. doi: 10.1111/j.1468-2958.1988.tb00161.x Rook, K. S. (1987). Social support versus companionship: Effects on life stress, loneliness, and evaluation by others. Journal of Personality and Social Psychology, 52, 1132-1147. Sanderson, C. A., Rahm, K. B., Beigbeder, S. A., & Metts, S. (2005). The link between the pursuit of intimacy goals and satisfaction in close same-sex friendships: An examination of the underlying processes. Journal of Social and Personal Relationships, 22, 75-98. doi: 10.1177/0265407505049322 Schmidt, G., & Weiner, B. (1988). An attribution-affect-action theory of behavior: Replications of judgments of help-giving. Personality and Social Psychology Bulletin, 14, 610-621. doi: 10.1177/0146167288143021 Siebert, D. C., Mutran, E. J., & Reitzes, D. C. (1999). Friendship and social support: The importance of role identity to aging adults. Social Work, 44, 522-533. doi: 10.1093/sw/44.6.522 Southwell, B. G., & Torres, A. (2006). Connecting interpersonal and mass communication: Science news exposure, perceived ability to understand science, and conversation. Communication Monographs, 73, 334-350. doi: 10.1080/03637750600889518 Ting-Toomey, S., Gao, G., Trubisky, P., Yang, Z., Kim, H. S., Lin, S.-L., & Nishida, T. (1991). Culture, face maintenance, and styles of handling interpersonal conflict: A study in five cultures. The International Journal of Conflict Management, 2, 275-296. doi: 10.1108/eb022702 Trees, A. R. (2005). Nonverbal and verbal approach behaviors related to reported and perceived support seeking goals in conversations. Western Journal of Communication, 69, 319-337. doi: 10.1080/10570310500305190 Wardaugh, R. (1985). How conversation works. Oxford, UK: Basil Blackwell. Watson, D., Clark, L. A., & Tellegen, A. (1988). Development and validation of brief measures of positive and negative affect: The PANAS scales. Journal of Personality and Social Psychology, 54, 1063-1070. doi: 10.1037/0022-3514.54.6.1063 Weiner, B. (1980). A cognitive (attribution)-emotion-action model of motivated behavior: An analysis of judgments of help-giving. Journal of Personality and Social Psychology, 39, 186-200. doi: 10.1037/0022-3514.39.2.186 82 Weiner, B. (1985). An attribution theory of achievement motivation and emotion. Psychological Review, 92, 548-573. doi: 10.1037/0033-295X.92.4.548 Weiner, B. (1995). Judgements of responsibility: A foundation for a theory of social conduct. New York, NY: Guilford. Weiner, B. (2014, Feb. 27). [RE: Likelihood to help scale request]. Wellman, B., & Wortley, S. (1990). Different strokes from different folks: Community ties and social support. American journal of Sociology, 558-588. doi: 10.2307/2781064 White, J. B., Tynan, R., Galinsky, A. D., & Thompson, L. (2004). Face threat sensitivity in negotiation: Roadblock to agreement and joint gain. Organizational Behavior and Human Decision Processes, 94, 102-124. doi: 10.1016/j.obhdp.2004.03.005 Wilson, S. R., Aleman, C. G., & Leatham, G. B. (1998). Identity implications of influence goals: A revised analysis of face-threatening acts and application to seeking compliance with same-sex friends. Human Communication Research, 25, 64-96. doi: 10.1111/j.14682958.1998.tb00437.x Zhang, S., & Stafford, L. (2008). Perceived face threat of honest but hurtful evaluative messages in romantic relationships. Western Journal of Communication, 72, 19-39. doi: 10.1080/10570310701828628 83