THE ROLE OF RACE AND GENDER IN EFFECTIVELY PROMOTING ORGANIZATIONAL DIVERSITY INITIATIVES By Danielle M. Gardner A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Psychology – Master of Arts 2018 ABSTRACT THE ROLE OF RACE AND GENDER IN EFFECTIVELY PROMOTING ORGANIZATIONAL DIVERSITY INITIATIVES By Danielle M. Gardner As organizations continue to pursue implementing diversity initiatives, one question lies in how organizations can propose and present such efforts so as to most likely maximize support. The current thesis sought to address this question by exploring whether the demographics of an individual proposing a diversity initiative may impact subsequent attitudinal and behavioral initiative support. Using logic grounded in the Elaboration Likelihood Model and the Attributional Analysis of Persuasion, I hypothesized that individuals of majority-group identities proposing diversity initiatives may elicit more supportive initiative attitudes and behaviors through perceptions of participant personal relevance and promoter self-interest. To test these predictions, I conducted an experiment in which participants evaluated a diversity initiative proposal written by an employee who was either White or Black, and male or female. Results suggest that indeed, White promoters were perceived as less self-interested than Black promoters promoting a diversity initiative, which in turn predicted more positive initiative attitudes and behavioral support for the initiative; however, the hypothesized role of personal relevance was not supported. Theoretical and practical implications, as well as study limitations, are discussed. ACKNOWLEDGEMENTS I would like to thank my advisor, Dr. Ann Marie Ryan, for her developmental support on this project, in addition to her mentorship throughout my graduate career thus far. Additionally, I would like to thank my committee members, Dr. Kevin Ford and Dr. Fred Leong, for their helpful suggestions and insightful comments throughout this thesis process. Finally, I would like to thank my parents for their constant love and encouragement throughout the entirety of my education. I am endlessly appreciative of each of the sacrifices you have made towards developing me as a person, and I would certainly not be where I am today without your limitless love and support. iii TABLE OF CONTENTS LIST OF TABLES ..................................................................................................................... vi LIST OF FIGURES .................................................................................................................. vii CHAPTER 1 ............................................................................................................................... 1 INTRODUCTION ...................................................................................................................... 1 Diversity Initiatives ................................................................................................................. 2 Diverse Attraction and Recruiting ....................................................................................... 3 Diversity Retention and Climate .......................................................................................... 5 Diversity Training ............................................................................................................... 6 Reactions to Diversity Initiatives ......................................................................................... 8 CHAPTER 2 ............................................................................................................................. 10 THEORETICAL BACKGROUND........................................................................................... 10 Elaboration Likelihood Model ............................................................................................... 10 Attributional Analysis of Persuasion ..................................................................................... 15 Attitudes and Behaviors ........................................................................................................ 18 Proposed Model .................................................................................................................... 19 CHAPTER 3 ............................................................................................................................. 20 METHOD ................................................................................................................................. 20 Participants ........................................................................................................................... 20 Procedure .............................................................................................................................. 21 Primary Measures ................................................................................................................. 22 Additional Measures ............................................................................................................. 24 CHAPTER 4 ............................................................................................................................. 27 RESULTS ................................................................................................................................ 27 Hypothesis Testing:............................................................................................................... 27 Expanded Path-Model: .......................................................................................................... 31 Exploratory Analyses: ........................................................................................................... 32 CHAPTER 5 ............................................................................................................................. 35 DISCUSSION........................................................................................................................... 35 Theoretical and Practical Implications ................................................................................... 40 Limitations and Future Directions ......................................................................................... 41 Conclusion ............................................................................................................................ 43 APPENDICES .......................................................................................................................... 45 APPENDIX A ....................................................................................................................... 46 APPENDIX B ....................................................................................................................... 53 iv APPENDIX C ....................................................................................................................... 62 REFERENCES ......................................................................................................................... 65 v LIST OF TABLES Table 1. Means, standard deviations, and bivariate correlations between manipulations and measured outcomes. ……………………………………………………………………………46 Table 2. Coefficients, standard errors, and significance values of personal relevance and promoter self-interest predicting initiative attitudes……………………………………………47 Table 3. Path estimates, standard errors, and p-values of path-model linking promoter race to voting behavior…………………………………………………………………………………48 Table 4. Path estimates, standard errors, and p-values for path-model linking promoter race to charity donation behavior………………………………………………………………………49 vi LIST OF FIGURES Figure 1. Proposed model linking promoter demographics to attitudes and behavioral support for proposed diversity initiative. …………………………………………………………………...50 Figure 2. Marginal interaction between promoter gender and promoter race on initiative attitudes…………………………………………………………………………………………51 Figure 3. Path-model tested in Mplus…………………………………………………………..52 vii CHAPTER 1 INTRODUCTION In the past five years, a number of industry giants have made large-scale commitments towards workplace diversity. Google spent $115 million on diversity initiatives in 2014, with $150 million dollars set aside for similar efforts in 2015 (Guynn, 2015). In that same year, Intel established a $300 million fund to increase the presence of diverse employees within their organization (Wingfield, 2015). Apple also joined in, committing over $50 million to diversity efforts in 2015 (Lev-Ram, 2015). While particularly large-scale, these examples are not unique, as nearly half of midsize companies and almost all Fortune 500 companies are currently investing in diversity initiatives (Dobbin & Kalev, 2016). Indeed, research suggests a growing number of organizations of all sizes are attempting to increase inclusiveness of under-represented individuals through the use of diversity management and diversity initiatives (Gilbert, Stead & Ivancevich, 1999). Such initiatives include efforts towards diverse recruitment, equitable selection methods, retention efforts for diverse employees, and diversity training (Ryan & Powers, 2012). However, the effectiveness of such efforts can often depend on employee reactions. Specifically, there is evidence of hesitance and resistance against diversity initiatives on the part of majority group members, most often whites and white males (Kravitz & Klineberg, 2000; Kravitz et al., 2000). Researchers have explored a number of methods to lessen this resistance, including program justification (Kidder, Lankau, Chrobot-Mason, Mollica & Friedman, 2004) and program framing (Kraviz & Klineberg, 2000). What has received less attention in the organizational literature are the potential effects of who is promoting the diversity initiatives on employee perceptions. 1 Research in social psychology has provided evidence suggesting the effectiveness of majority-group members in combating interpersonal discrimination of minorities (Petty, Fleming, Priester & Feinstein, 2001; Czopp & Monteith, 2003); accordingly, the purpose of the following study is to extend such thinking to organizational diversity initiatives. Specifically, the present study examines whether employee reactions to diversity initiatives differ depending on the demographics of the individual suggesting the diversity program. In line with a recent study finding that white male leaders are the least penalized for promoting diversity in the workplace (Hekman, Johnson, Foo & Yang, 2016), the current study explores the effectiveness of White and male individuals as diversity promoters. However, this thesis will go beyond simply identifying the effects of promoter demographics, and explore two potential mechanisms to explain the hypothesized differential outcomes. Such results can help those both in industry and academia understand how to more effectively promote and implement organizational diversity initiatives, while highlighting the importance and utility of majority-group allies. I will begin by providing a brief background describing the major types of diversity initiatives, as well as research outlining employee reactions to such efforts. Next, I will outline two major theories within the persuasion literature, which act as the primary theoretical grounding for my set of hypotheses. Finally, I will discuss the potential link between attitudes in support of organizational initiatives and supportive behaviors, prior to presenting my proposed model, methods, and results. Diversity Initiatives The present study focuses on increasing support for diversity initiatives by considering the effect of who is promoting the change effort. However, it is important to clarify the definition of “diversity initiatives,” and consider in what form these efforts may take place. Therefore, the 2 current section will outline literature related to common organizational diversity initiatives, including efforts related to diverse recruiting, retention of diverse employees, and diversity training. Diverse Attraction and Recruiting One type of organizational diversity initiative focuses on the attraction and recruitment of diverse applicants. With goals to increase diversity in organizations for practical, legal, and ethical reasons, many recruitment efforts focus on increasing the number of minorities and women applying for organizational positions (Newman & Lyon, 2009). Many of these efforts fall under targeted recruiting efforts, which encompass a wide-variety of activities to increase the attractiveness of the organization to a diverse group of individuals (Ryan & Powers, 2012). This is in contrast to generalized recruiting methods, which involve advertising an available position without reference to specific applicant demographics (Newman & Lyon, 2009). Therefore, targeted recruiting efforts may communicate to diverse applicants that they are particularly valued at one organization compared to another organization utilizing more generalized methods. These targeted recruiting efforts can take a number of forms. For instance, diverse candidates have been found to respond positively to images of diverse employees in recruitment advertisements and brochures (Avery, 2003; Avery, Hernandez & Hebl, 2004). Specifically, Avery (2003) conducted an experiment in which he manipulated organizational website advertisements to reflect varying levels of racial composition at different levels of the company (entry level vs. managerial level). He found that Black viewers were attracted to advertisement diversity particularly when it extended to the managerial level, and that White viewers were unaffected by varying levels of racial composition. Similar findings were obtained by Avery, Hernandez and Hebl (2004); upon manipulating organizational recruitment brochures to reflect 3 varying levels of diverse employee representation, the authors found that Black and Hispanic participants were more attracted to the organization when minorities were depicted, while White participants were unaffected by differing racial compositions. Beyond images, researchers have also examined the effectiveness of diversity statements in recruitment materials on diverse recruitment outcomes. For instance, Kim and Gelfand (2003) found that individuals with higher levels of ethnic identity had greater job pursuit intentions towards a company using a brochure featuring a diversity initiative statement than one using a brochure without such a statement. Further, Highouse and colleagues (1999) found Black applicants to be more attracted to organizations that advertise identity-conscious staffing policies emphasizing a commitment to equal opportunity than those not advertising such values. Broadly speaking, the combination of these findings suggests that the use of verbal and pictorial cues signaling pro-diversity values enhances female and minority applicants’ perceptions of organizations (Avery & McKay, 2006). It is important to note, however, that attraction of minority candidates does not necessarily ensure a diverse workforce. Although targeted recruiting may increase the number of minorities in an applicant pool, such efforts may simultaneously generate applications from candidates who are unqualified or uninterested in the position (Newman & Lyon, 2009). Consequently, only efforts resulting in an increase of qualified minority applicants able and willing to sustain a multistage hiring process will likely result in an increasingly diverse workforce. In fact, recruiting on demographics alone may potentially worsen an organization’s adverse impact ratio if minority- and majority-group applicants are disproportionately qualified (Newman & Lyon, 2009), as less qualified minority applicants may be more likely to withdraw earlier in the multistage selection processes (Schmit & Ryan, 2000; Ryan, Sacco, McFarland & 4 Kriska, 2000), resulting in more homogenous pool upon time for final decisions. Accordingly, efforts for diverse recruiting must focus not only on increasing the presence of diverse applicants, but also ensuring that these diverse applicants are appropriately qualified for the available positions. Diversity Retention and Climate After diverse employees are recruited and selected into an organization, organizational efforts may be taken to ensure that these employees are retained. Evidence suggests that turnover may be particularly costly for firms investing in minority recruitment, as minority retention rates tend to be significantly lower than those of White employees (McKay et al., 2007). Some studies suggest similar trends for women, such that their levels of turnover intention and turnover rates may be higher than their male counterparts (Xu, 2008; Hom, Roberson & Ellis, 2008). Given these trends, research has been devoted to identifying the causes of these turnover disparities, as well as the effectiveness of various diversity initiatives aimed at reducing such differences between groups. One line of inquiry has explored the relationship between diversity climate [defined by Kossek and Zonia (1993) as general perceptions towards diversity initiatives and their probable beneficiaries] and retention of diverse employees. For instance, McKay and colleagues (2007) related pro-diversity work climate perceptions with turnover intentions for a sample of 5370 Black, Hispanic and White managers. Ultimately, the authors found that compared to their White and Hispanic counterparts, Black managers’ diversity climate perceptions were significantly more associated with turnover intentions. Additionally, the authors found that this relationship was mediated by organizational commitment. Kaplan, Wiley and Maertz (2011) identified 5 similar relationships using a survey of 4184 workers, finding that positive perceptions of an organization’s diversity climate were related to decreased turnover intentions. Given the links between diversity climate and diverse employee retention, organizational efforts have been examined to create and strengthen an organization’s diversity climate. For example, Herdman and McMillan-Capehart (2010) surveyed 3578 employees across 163 organizations to better understand the determinants of diversity climate perceptions. The authors found that though presence of diversity programs (defined as “those which are overtly aimed at increasing and managing the diverse composition of an organization’s workforce”) was predictive of diversity climate, this relationship was moderated by the actual diversity of the organization, as well as the collective relational values of management teams. These findings suggest that though diversity programs are crucial to forming a strong diversity climate, their presence is not enough; in conjunction, organizations should also pay attention to managerial values and levels of minority representation in management positions. Diversity Training A final common type of diversity initiative is diversity training. Organizations are increasingly turning to diversity training in an attempt to realize the full benefits of an increasingly diverse workforce, while minimizing the negative outcomes that can arise from dissimilarity (Holladay & Quinones, 2008). This type of training can be defined as one focused on developing the skills, knowledge, and motivation of employees to work productively and effectively alongside others of varying demographics and beliefs (Pendry, Driscoll & Field, 2007). These trainings can vary in their approaches and ultimate success. Consequently, Bendick, Egan and Lofhjelm (2001) sought to classify the forms in which these trainings were provided, and link these training differences to success outcomes. Through the surveying of a 6 sample of diversity trainers, the authors classified such trainings into three primary forms: those focused on individual attitudes, those focused on behavior and organizational systems, and those focused on changing workplace culture using organizational development approaches. Upon linking such trainings to effectiveness outcomes, the authors found that trainings focused on changing culture were the most effective, yet were only used by 25% of surveyed diversity trainers. Investigations into the antecedents of diversity training success have identified a variety of relevant factors. As an example, Rynes and Rosen (1995) found through their survey of 785 human resource professionals that perceived diversity training success was predicted by mandatory attendance for managers, long-term evaluations of results, managerial rewards for increasing diversity, and a broad inclusionary definition of organizational diversity. Further, Holladay, Knight, Paige and Quinones (2003) found that the way in which the training is framed affects attitudes and outcomes, such that broader frames (e.g. a training focused on multiple identity categories) elicited more positive outcomes than framings with narrower focuses (e.g. a training focused only on race). Finally, Combs and Luthans (2007) found improved outcomes for diversity trainings incorporating self-efficacy components, and identified diversity self-efficacy as a mediator for training and subsequent pro-diversity behavioral intentions. With these results in mind, Kalinoski and colleagues (2013) conducted a meta-analysis examining the overall effects of diversity training across 65 studies, while attempting to identify relevant boundary conditions and moderators. Ultimately, the authors found a consistent small- to medium-sized effect across all training outcomes, and larger effects on cognitive-based and skill-based outcomes relative to affective-based outcomes. Similarly, Berzukova, Spell, Perry and Jehn (2016) found in their meta-analysis of 260 samples that diversity training had the 7 largest effects on reactions to training and cognitive learning, with smaller effects found for behavioral and attitudinal learning. These authors ultimately concluded that although many diversity training programs fell short in effectively impacting certain training characteristics, examples of successful diversity training were indeed identified. In summary, diversity initiatives may span a variety of organizational efforts, from recruitment, to retention efforts, to training programs, and more. However, a review of relevant literature suggests that these initiatives may vary widely in their respective effectiveness and ability to impact important organizational outcomes. Consequently, researchers have sought to uncover how these initiatives are perceived by workers by exploring reactions to organizational diversity initiatives. The following section will review that literature, and discuss how previous research informs the current thesis. Reactions to Diversity Initiatives A number of studies have examined the way in which employees react to initiatives of this nature. For instance, Kossek and Zonia (1993) examined differences in reactions towards organizational diversity promotion efforts based on employee demographics. Based on intergroup theory, the authors predicted that given perceptions that organizational efforts to become increasingly multicultural would adversely impact current dominant groups (i.e. White men), individuals of these identities would be less supportive of such initiatives, to the detriment of the effort’s effectiveness. Through the surveying of 775 employees of a large public Midwestern university, the authors found that compared to White women and racial minorities, White men placed less value on organizational diversity initiatives, and further held less favorable attitudes towards the qualifications of women and racial minorities. The authors also found that gender heterogeneity of a working unit was predictive of support for diversity efforts. 8 Specifically, regardless of the respondents’ gender or race, increased ratios of women within a working unit were predictive of attitudes valuing diversity. Also basing their framing in social identity and intergroup theory, Mor Barak, Cherin and Berkman (1998) explored the relationship between diversity perceptions and employee demographics for a sample of 2686 employees of an electronics company. In line with the findings outlined by Kossek and Zonia (1993), the authors found that compared to White women and racial minorities, White men perceived the organization as more fair and inclusive. Further, White men reported seeing less value in and feeling more uncomfortable towards diversity than did women or minority samples. It is important to note that while much of the research has focused on the reactions of White men, individuals of other demographics may also vary in their reactions to diversity initiatives. For example, racial minorities may oppose such efforts in situations in which they wish to not highlight their minority status (Holladay, Knight, Paige & Quinones, 2003). Further, White women may withhold support for diversity efforts if they feel they are not explicitly included as potential beneficiaries (Kidder et al., 2004). Therefore, it is important to consider proactive employee reaction strategies intended for individuals of all demographics, not just White males. While research has made strides in identifying these differences in reactions, a gap remains in understanding the mechanisms behind how these attitudes are formed. As such, the current study will explore two potential mechanisms to better understand what causes varying reactions, and how positive attitudes in favor of diversity initiatives may be maximize. 9 CHAPTER 2 THEORETICAL BACKGROUND In this next section, I will introduce two theoretical frames to help explain the potential relationship between promoter demographics and differential attitudinal outcomes. First, I will discuss the Elaboration Likelihood Model, which can inform how promoter demographics may relate to persuasive outcomes through perceived argument relevance for the message recipient. After, I will outline the Attributional Analysis of Persuasion, which explains how perceptions of promoter self-interest may be key in understanding the potential relationship between promoter demographics and differences in persuasive outcomes. Elaboration Likelihood Model One explanatory tactic through which diversity initiatives may be more effectively argued can be found in the persuasion literature. Specifically, Petty and Cacioppo’s (1986) Elaboration Likelihood Model describes persuasion as a process that can take one of two paths. The central route represents purposeful and thoughtful consideration of an argument, while the peripheral route is more reliant on simple cues rather than scrutiny of the presented information. In the case of the current study, it is possible that promoter demographics may elicit differential persuasive outcomes through either route. What determines which route is followed is primarily based on three factors: ability to process the information, the nature of cognitive processing, and motivation to process the information. Particularly relevant to the central route is the motivation determinant. As Petty and Cacioppo (1986) outline, factors that make one individual more motivated to process a persuasive message than another include personal relevance, need for cognition, and personal responsibility. Therefore, one possible lever to increase majority-group members’ motivation to 10 use the central processing route when considering a diversity initiative proposal may be to make the message increasingly personally relevant. Numerous studies have examined the relationship between increasing personal relevance and subsequent persuasive message processing. For instance, Cacioppo, Petty and Sidera (1982) conducted an experiment in which they manipulated message type and relevance to participant self-schema, such that some participants were in conditions where the message matched their self-schema, while others were in incongruent conditions. Results suggested that message recipients tended to generate more topic-specific and favorable thoughts when the message was congruent to their self-schema than when it was incongruent. Additionally, the experiment conducted by Burnkrant and Unnava (1989) hypothesized that under situations of high personal relevance, people are more motivated to engage in issue-relevant elaboration than when the issue has little personal relevance. To test this, the authors examined the effects of manipulated levels of participant self-referencing and argument strength on message recall and attitudes, ultimately finding greater message recall in the self-referencing condition. Further, the authors found that participants were more responsive to message arguments in the high self-referencing conditions than in the low self-referencing conditions. These collective findings therefore suggest that increasing relevance for message recipients can lead to improved persuasive outcomes. While the majority of the referenced self- relevance manipulations were related to attitude congruence between recipient and message, other research has looked at the effects of demographic congruence between message provider and recipient. In line with the argument that demographics such as race and gender are salient dimensions of the self (McGuire & Padawer-Singer, 1976), researchers have examined whether congruence on such aspects elicits similar effects found in studies on attitude congruence. For 11 instance, Lee, Fernandez and Martin (2002) found evidence that an advertisement featuring a model that was demographically congruent with a participant’s race led to increased instances of participant self-reference, which in turn led to more favorable thoughts and attitudes regarding the advertised product. Similar results linking demographic congruence between model and participant to improved product attitudes via identity referencing have been replicated in a number of studies (Martin, Lee & Yang, 2004; Sierra, Hyman & Torres, 2009). Therefore, the Elaboration Likelihood Model (Petty & Cacioppo, 1986) would seem to suggest that one could increase a person’s likelihood of thoughtful consideration of an argument by leveraging the message’s personal relevance. As demographics including race and gender are salient aspects of the self, having an individual of similar identities promote a message may increase the chance the argument receives purposeful reflection and ultimately positive attitude endorsement. Such demographic congruence may be particularly useful in the case of majority- group members evaluating a diversity-related proposal; as the content of the message may not seem immediately personally relevant for majority-group members, the presence of an individual of similar demographics delivering the argument may increase perceived personal relevance. Consequently, I anticipated that the relationship between promoter demographics and perceived argument personal relevance would differ depending on the demographics of the participant. Given potential differences in baseline levels of attitudes towards diversity initiatives between majority- and minority-group participants (Kossek & Zonia, 1993; Cherin & Berkman, 1998; Holladay et al., 2003), there may be a greater opportunity to improve attitudes of majority- group members compared to those of minority-group members, who may already have comparatively more positive attitudes towards this topic. Further, in considering participant perceptions of argument relevance, it may be that those with a minority-group identity already 12 feel that a proposed diversity initiative is relevant to them, and therefore the effect of who is delivering the message may not be particularly important for this group; however, for majority- group participants who may not initially see personal relevance in a diversity initiative, the effect of promoter demographics may be particularly strong. As such, I hypothesized the following: Hypothesis 1: Promoter demographics will predict perceived personal relevance as moderated by recipient demographics, such that the effect will be stronger for majority- group members compared to minority-group members. Hypothesis 2: Perceived personal relevance of a diversity initiative will be positively related to positive attitudes towards the proposed initiative. Hypothesis 3: Perceived personal relevance will mediate the relationship between promoter demographics and attitudes towards the proposed initiative, as moderated by participant demographics. While the discussed hypotheses are based in logic surrounding use of the central route of persuasion, it is also possible that promoter demographics may lead to differential persuasive outcomes through the peripheral route. As outlined by Petty and Cacioppo (1986), the peripheral route of persuasion involves interpretation of simple cues rather than scrutinized consideration of the argument. Such cues may include perceived promoter credibility, production quality of the message, and promoter attractiveness. Indeed, a number of studies have demonstrated a link between communicator attractiveness and subsequent persuasion (Chaiken, 1979; Kahle & Homer, 1985), such that physically attractive communicators tended to be more persuasive than communicators who were less attractive. As such, it would seem that communicator visible characteristics may directly influence persuasive outcomes; consequently, as promoter race and 13 gender are visible and salient identities, it is possible that such characteristics could similarly affect persuasion. This possibility seems particularly likely given differences in perceived credibility and power based on demographics. As White males continue to represent the majority of membership in the highest tiers of industry (Zweigenhaft & Domhoff, 2011), it follows that individuals of these demographics are likely to hold power and influence. Further, research suggests that beyond actual power, males may hold greater informal influence as they are viewed as more competent than their female counterparts (Goldberg, 1968; Gruber & Gaebelein, 1979). As the peripheral route operates on quick evaluations based on simple cues rather than true argument consideration, it follows that promoter demographics may elicit differential persuasive outcomes based on heuristic judgments of credibility and power. Therefore, I hypothesized the following: Hypothesis 4: Promoter demographics will be related to positive attitudes towards the proposed initiative, such that majority-group promoters will elicit more positive attitudes than minority-group promoters. It is worth noting that this proposed direct path between promoter demographics and persuasive outcomes may also result from implicit or explicit biases against women and non- White individuals. Indeed, research suggests that both explicit and implicit measures of bias are systematically related to a number of negative behaviors, ranging from subtle, interpersonal discrimination to more overt discriminatory acts (Dovidio, Kawakami & Gaertner, 2002; Hebl, Foster, Mannix & Dovidio, 2002). Consequently, the possibility of bias as the root cause of potential differential reactions dependent on promoter demographics must be considered as an 14 alternative explanation, and is addressed in the current study through the statistical control of participant biases. Attributional Analysis of Persuasion While the Elaboration Likelihood Model can shed light on how promoter demographics can affect attitudes towards diversity initiatives either directly or through perceptions of message relevance, another model can help to understand a process involving perceptions of promoter intentions. Specifically, the Attributional Analysis of Persuasion (AAP; Eagly & Chaiken, 1975; Eagly, Wood & Chaiken, 1978; Wood & Eagly, 1981) suggests the importance of understanding the role of the perceiver’s causal inferences concerning why a communicator is advocating a particular position. The AAP posits that message expectancy is related to subsequent perceiver persuasion, such that the less expected a communicator’s position given his/her personal characteristics and situational pressures, the stronger the perceiver’s inference that the message corresponds to reality (Eagly & Chaiken, 1975). Alternatively, when a communicator advocates for an expected position, persuasion is less likely as perceivers question the communicator’s honesty (Priester & Petty, 1995) and bias (Eagly, Wood & Chaiken, 1978). Thus, majority-group individuals proposing a diversity initiative may be more effective than women or racial minorities in subsequent persuasion, as the proposal may be perceived as unexpected given the communicator’s characteristics. A number of studies have tested this theory, ultimately finding evidence in support of its assertions. For example, Priester and Petty (1995) conducted an experiment in which argument quality and argument expectancy were manipulated, such that participants were either in conditions in which a communicator argued for an expected or unexpected position based on previously provided information. In line with AAP predictions, their results suggest that subjects 15 exposed to the expectancy-disconfirming source rated the communicator as more honest than those in the expectancy-confirming condition. Similarly, Eagly, Wood and Chaiken (1978) manipulated message expectancy via background information provided prior to the delivery of a message. The authors found that subjects changed their opinions more when the message disconfirmed rather than confirmed their expectancies, and that communicators who delivered an expected message were rated as more biased. Finally, Wood and Eagly (1981) tested a model examining the processing stages of recipients interpreting a message. In the first stage, recipients were thought to infer the causes for the position the communicator took in the message as either due to personal characteristics or factual evidence. Next, recipients hypothetically determined the degree to which the communicator is biased in his/her understanding of the issue. Lastly, the recipient’s perception of communicator bias was expected to be related to message persuasiveness. The authors tested this model experimentally and ultimately identified evidence in its support. Researchers such as Petty and colleagues (2001) have attempted to extend this model of persuasion with relation to issues involving diversity and discrimination. Specifically, Petty et al. (2001) conducted an experiment in which participants read an essay written by a first-generation black student. Within the essay, the student either advocated for a new scholarship to benefit racial minorities (confirmed expectation, given perceived self-interest) or children of alumni (disconfirmed expectation, given writer could not benefit). Results showed that when the student disconfirmed individual interest by advocating for a cause not benefiting himself, he was rated as more trustworthy. Further, participants were more surprised when their expectations were disconfirmed than were those in the expectancy-confirmation conditions. Related research has focused on the differential effectiveness of discrimination confrontations based on demographics 16 of the confronter. For example, Czopp and Monteith (2003) manipulated the race and gender of a confederate confronting participants regarding a sexist or racist statement. The authors ultimately found that non-target confrontations elicited more participant guilt, less participant discomfort, and were viewed as less of an overreaction than confrontations by target individuals. The researchers hypothesized that these outcome differences were rooted in differential perceptions of confronter self-interest, with recent research providing evidence with this line of thought (Gulker, Mark and Monteith, 2013). Therefore, the Attributional Analysis of Persuasion would appear to assert that communicator characteristics relate to message expectancy through perceptions of self-interest. To the extent that the communicator presents an argument that is expected based on his/her characteristics and interests, subsequent persuasion will be less likely than if the communicator advocates for an unexpected argument. As research on source-position expectancies has found that people expect communicators to take positions in their own self-interest (Petty et al., 2001), it would follow that communicators advocating an unexpected message unrelated to personal interest would subsequently be more persuasive. Related to the present study, I predicted that participants would more likely expect a diversity initiative proposal to be delivered by a female or racial minority, given perceptions of the communicator’s status as a beneficiary under the proposed policies; thus, the message delivered by a majority-group member (thought to not benefit from the proposed initiative) may be perceived as unexpected, and subsequently elicit increased positive participant attitudes. Given these predictions, I hypothesized the following relationships: 17 Hypothesis 5: Promoter demographics will be related to participant perceptions of promoter self-interest, such that majority-group promoters will be rated as less self- interested as compared to minority-group members. Hypothesis 6: Perceptions of promoter self-interest will be negatively related to attitudes towards the proposed initiative. Hypothesis 7: Perceptions of promoter self-interest will mediate the relationship between promoter demographics and attitudes towards the proposed initiative. Attitudes and Behaviors My final hypothesis proposed a link between attitudes towards the proposed initiative and behavioral support for the proposed initiative. This link between attitudes and behaviors has been examined by a wide variety of research over multiple decades (Ajzen & Fishbein, 1977; Regan & Fazio, 1977; Fazio, Chen, McDonel & Sherman, 1982). Attitudes can be defined as being held in relation to some aspect of the individual’s world, while behavioral criteria are defined as consisting of one or more observable actions (Ajzen & Fishbein, 1977). One theory that can help explain this attitude-behavior link is the theory of planned behavior. A central factor in the theory of planned behavior is an individual’s intention to perform a given behavior (Ajzen, 1991), such that stronger intentions to engage in a behavior should be associated with greater likelihood that the behavior will be performed. The theory suggests that a crucial antecedent for behavioral intention is the individual’s attitude towards the behavior, such that positive attitudes will lead to greater intention to perform the behavior. Therefore, the theory of planned behavior would suggest that attitudes are linked to behaviors through one’s intention to perform that behavior. Given this theoretical framing, I investigated the following hypothesis: 18 Hypothesis 8: Attitudes towards the proposed initiative will be related to behavioral support for the proposed initiative. Proposed Model Given the logic grounding the previously discussed hypotheses, I proposed the following model (Figure 1) explaining the processes through which promoter demographics may be linked to positive attitudes and behavioral support for the proposed diversity initiative. As outlined by the Elaboration Likelihood Model, I expected that the demographics of an individual promoting a diversity initiative may be linked to participant attitudes towards the initiative through perceived relevance of the message to the participant. I expected this relationship to be moderated by the participant’s own demographics, such that the effect of promoter demographics may be particularly relevant for majority group participants, while this effect may be less pronounced for women and racial minority participants. Further, as suggested by the Attributional Analysis of Persuasion, I expected that promoter demographics may be linked to attitudes toward the proposed initiative through perceived promoter self-interest. Finally, I predicted that participant attitudes towards the proposed initiative will be related to behavioral support for the initiative. 19 CHAPTER 3 METHOD To test this model, I conducted an experiment in which the race and gender of an individual proposing a diversity initiative was manipulated. Race was manipulated such that the promoter was either White or Black, and gender was manipulated such that the promoter was either male or female. Therefore, the study was a 2 (Promoter race: White, Black) x 2 (Promoter gender: male, female) factorial experimental design. Participants Data from 420 participants was collected from Amazon Mechanical Turk (mTurk). In order to be eligible for the study, participants were required to work full-time outside of mTurk and be located within the United States. Only data from those who passed the three manipulation checks (available in Appendix B) was analyzed, leaving a final sample of 352 participants. Of the remaining participants, 194 (55.0%) identified as male, 155 (43.9%) identified as female, and 3 (0.1%) preferred not to disclose gender. The average age of participants was 34.81 years (SD = 9.97), with an average of 14.63 years of work experience (SD = 9.84). The sample was majority White (74.9%), with 11.7% identifying as Asian, 8.3% identifying as Black, 3.4% identifying as Latino/Hispanic, and 1.7% identifying as either Native Hawaiian/Pacific Islander, American Indian/Alaska Native, or other. More specifically, 40.8% of our sample were White men, with the remaining 59.2% comprised of women and racial minorities, as these groups were purposely sampled to test proposed participant moderation in Hypothesis 1. With respect to previous exposure to organizational diversity initiatives, 48.6% reporting having worked for an organization in which a diversity initiative was enacted. Participants were compensated $1.30 in exchange for their time. 20 Procedure Participants were asked to imagine that they worked for a fictional organization (Stone & Steel) that has solicited proposals from workers with regards to what sorts of initiatives they would like the organization to pursue. Participants were tasked with reviewing and evaluating the written proposal of a coworker, on the topic of establishing a Diversity Task Force. Participants first read a short introduction discussing the author of the proposal. It is in this introduction that the experimental manipulations regarding promoter demographics were expressed. Specifically, the author introduction featured the name and photo of the person proposing the diversity initiative, featuring either a White male, White female, Black male, or Black female (all stimuli available in Appendix C). Photo stimuli were pilot tested for equivalence on attractiveness and perceived age. Participants then read the short proposal written by the described author, advocating for the establishment of a task force devoted to diversity within their organization. More specifically, the proposal described plans for the task force to enact diversity training, diverse recruitment strategies, and mentoring programs for diverse employees within the organization. This proposal was pilot tested to ensure sufficient variance in participant reactions and responses prior to final data collection. Upon reflecting on the proposal, participants completed a series of measures, both directly applicable to study hypotheses as well as a number of additional exploratory measures. These measures were administered in the following order: perceived persuasiveness, attitudes towards the proposed initiative, behavioral support of the proposed initiative, perceived personal relevance, perceived promoter self-interest, perceived promoter credibility, Social Dominance Orientation, Attitudes Towards Workplace Diversity, manipulation checks, participant demographics, charity donation. More information about each 21 of these measures is discussed in the next section, organized by whether the measure was primary and central to the discusses hypotheses, or exploratory and additional in nature. Primary Measures Attitudes towards the proposed initiative. To measure participant attitudes towards the proposed initiative, participants completed a nine-item bipolar scale created for the purposes of this study. Participants were asked to indicate the extent to which a pair of opposite adjectives reflected their opinion towards the proposed initiative, by marking on a seven-point scale which adjective more accurately reflected their attitudes. An example item is “Low Quality/High Quality,” with the entire scale available in Appendix B. This scale was pilot tested for internal reliability and sufficient variance using a sample of undergraduate students at a large Midwestern university. Results of the pilot suggested strong internal consistency (α = .94), with sufficient variance and range to proceed with using the measure as constructed (SD = 1.00). This measure demonstrated further internal consistency within the final study (α = .96). Behavioral support of proposed initiative. To assess behavioral support of the proposal, participants responded to a single item asking if they would vote in support of the proposed initiative (i.e. Knowing that there are other proposals to consider and limited funding such that not all proposals will be implemented, would you vote in support for or against the proposed initiative?). Behavioral support additionally was measured via an item assessing the amount of money the participant would allocate to the initiative (i.e. How much money would you choose to allot to the proposed initiative out of the $5000 available, knowing that there are other proposals to consider and accommodate?). As a continuous rather than dichotomous behavioral indicator, this question was included to ensure the outcome demonstrated adequate 22 variability to examine potential differences. Accordingly, participants were asked to allot anywhere between $0-$5000 via a sliding scale. Perceived personal relevance. To measure perceived personal relevance, participants completed a measure comprised of items adapted from two personal-relevance scales, one by Wells, Leavitt and McConville (1971) and the other by Zhao and Peterson (2017). This six-item measure was assessed on a seven-point Likert-type scale (1 = “Strongly disagree” to 7 = “Strongly agree”), and had strong internal reliability (α = .95). An example item is “This proposal is relevant to my life.” Perceived promoter self-interest. To measure perceived promoter self-interest, participants completed a measure comprised of items pulled and adapted from Gerbasi and Prentice’s (2013) self-interest subscale and Tseng and Fan’s (2011) Self-Interest Scale. This five-item measure was captured on a five-point, Likert-type scale (1 = “Strongly disagree” to 5 = “Strongly agree”), and showed strong internal consistency (α = .93). An example item is “The author of this proposal, by and large, is pursuing his/her own interest.” Manipulation checks. Participants completed a set of manipulation checks to ensure that they were able to correctly identify the gender and race of the author of the proposal, as well as the topic of the proposal (exact items available in Appendix B). Participants who could not correctly identify the author’s gender, the author’s race, and the topic of the author’s proposal were not included within further analyses. Specifically, 36 individuals did not correctly identify the race of the promoter, 32 did not correctly identify the promoter’s gender, and 26 did not correctly indicate the topic of the proposal; consequently, their responses were removed. Demographics. Participants concluded the study by indicating their own gender, race/ethnicity, and age. Further, participants provided information with relation to their 23 employment, such as number of years employed, and previous experience with similar workplace diversity initiatives. Additional Measures While the primary measures were used to test the discussed proposed model, the following additional measures were also collected for a variety of purposes. For example, the Attitudes Towards Workplace Diversity scale (Montei, Adams & Eggers, 1996) and the charitable donation measure were included to assess whether the logic of the proposed model extended beyond the context of the organizational initiative. Further, a persuasiveness measure and the McCroskey and Teven’s (1999) credibility scale were included as variables that could potentially be affected by our manipulations and explain any potential outcome differences. Finally, the Social Dominance Orientation scale (Pratto, Sidanius, Stallworth & Malle, 1994) was included to be used as a covariate for analyses, to address participant bias as an alternative explanation for any identified effects. Perceived persuasiveness. To measure participant perceptions of proposal persuasiveness, participants completed a six-item bipolar scale created for the purposes of this study. Participants were asked to indicate the extent to which a pair of opposing adjectives reflected their opinion towards the proposal, by marking on a five-point scale which adjective more accurately reflected their attitudes. An example item is “Not Persuasive/Persuasive,” with the entire scale available in Appendix B. Results of pilot testing the measure in a sample of undergraduate students at a large Midwestern university suggested strong internal consistency (α = .91), with sufficient variance and range to proceed with using the measure as developed (SD = 0.69). The measure demonstrated further evidence of internal consistency within the present study (α = .93). 24 Perceived promoter credibility. Participant perceptions of the promoter’s credibility were assessed using two subscales within McCroskey and Teven’s (1999) Ethos/Credibility scale. The two subscales (competence and trustworthiness) were each comprised of six, bipolar items. An example item for the competence subscale is “informed/uninformed,” while and example for the trustworthiness subscale is “untrustworthy/trustworthy.” Because of high correlations between the two subscales, the subscales were combined and treated as a single credibility scale. This credibility scale demonstrated solid internal consistency (α = .90). Attitudes Towards Workplace Diversity. To assess attitudes towards organizational diversity in general, participants completed Montei et al.’s (1996) Attitudes Toward Diversity Scale (α = .90 in validation study). This 30-item instrument asked participants to indicate the extent to which they agree with each of the following statements from 1 (Strongly Disagree) to 5 (Strongly Agree). Example items include “I know some workers who would be fired if they were not minorities” and “Under most circumstance, I would prefer a male supervisor.” This measure displayed high internal reliability in the current study (α = .92). Social Dominance Orientation. To measure individual differences in support for inequity between social groups, participants completed the 16-item Social Dominance Orientation scale (Pratto et al., 1994; α = .90 in validation study). Measured on a seven-point Likert-type scale (1 = “Extremely negative” to 7 = “Extremely positive”), an example item for this measure is “Inferior groups should stay in their place.” This measure demonstrated solid internal consistency (α = .95), and was used as a covariate in all analyses involving the promoter demographic manipulations. Charity Donation. To assess more general behaviors in support of workplace diversity, participants completed a measure assessing diversity-related behaviors more distal to the 25 proposed initiative. Specifically, participants were asked to respond to a scenario in which they had the option to support a diversity-focused organization over other options. After completing the survey, participants were informed that along with their compensation, the researchers associated with this survey pledged to donate some additional funds to one of three charities for each survey completed. Participants were tasked with choosing which of the three charities the donation associated with their survey would support. Of the three charities, one was related to workplace diversity (American Association for University Women), while the other two were unrelated to diversity causes (Animal Welfare Institute and Coral Reef Alliance). Each of these charities received four-star overall ratings on Charity Navigator, and donations were in fact made to each of these causes upon completion of the study. 26 CHAPTER 4 RESULTS Means, standard deviations, and bivariate correlations between all variables are available in Table 1. Social Dominance Orientation (SDO) scores were used as a covariate for all analyses involving promoter gender and promoter race manipulations to control for underlying levels of bias, given the consistent significant correlations between SDO and the outcomes of interest. Given the high correlation between personal relevance and initiative attitudes (see Table 1), a series of confirmatory factor analyses were conducted to inform whether the measures are distinct from one another. The results of such analyses suggest that the measures are indeed distinctive, as a two-factor structure with personal relevance items loading onto one factor and initiative attitudes items loading on another had better model fit (RMSEA = .10, CFI = .95, TLI = .94, SRMR = .04) than did a one-factor model with all personal relevance and initiative attitude items loading onto a single factor (RMSEA = .19, CFI = .83, TLI = .80, SRMR = .06). Therefore, the measures were treated as distinct within subsequent analyses. As informed by the hypothesis separating participants into those of majority-group and minority-group status, all analyses involving participant demographics treated the variable as dichotomous (White male participants vs. Female and racial minority participants). Exploratory analyses looking at the race and gender of participants separately often revealed interactions supporting the division of groups in this manner, as the main effect of either demographic variable was often qualified by interactions comparing these two groups. Hypothesis Testing: Hypothesis 1. Hypothesis 1 predicted a relationship between promoter demographics and personal relevance ratings, moderated by participant demographics. Accordingly, a three-way 27 ANCOVA with promoter gender, promoter race, and participant demographics as the fixed factors and perceived personal relevance as the outcome was conducted to examine this possibility. No support for this hypothesis was found, as a three-way interaction between the independent variables was not significant F(1, 332) = .04, p = .834, meaning that the effect of promoter race and promoter gender on personal relevance ratings did not differ depending on participant demographics. Main effects of promoter race [F(1, 332) = .75, p = .388] and promoter gender on perceived personal relevance were also not significant, F(1, 332) = .33, p = .694. However, a main effect of participant demographics did emerge, such that White male participants found the proposal less personally relevant (M = 4.65, SD = 1.62) than female or racial minority participants (M = 5.11, SD = 1.54), F(1, 332) = 4.10, MSE = 11.85, p = .014 η2p = .02. Therefore, Hypothesis 1 was not supported. Hypothesis 2. Hypothesis 2 suggested that perceived personal relevance ratings would positively predict attitudes towards the proposed initiative. Indeed, simple regression with personal relevance as the predictor and attitudes towards the diversity initiative as the outcome showed that personal relevance scores significantly predicted attitudes towards the initiative, b = .71, SE = .03, p < .001, R2 = .63. Therefore, Hypothesis 2 was supported. Hypothesis 3. Hypothesis 3 predicted that perceived personal relevance would mediate the relationship between promoter demographics and attitudes towards the proposed initiative, as moderated by participant demographics. However, as Hypothesis 1 was not supported, a mediation of the moderation relationship is not possible. Further, the lack of direct relationship between promoter demographics and personal relevance ratings suggests that an un-moderated mediation relationship is not possible. Therefore, Hypothesis 3 was not supported. 28 Hypothesis 4. Hypothesis 4 predicted a direct relationship between the manipulated promoter demographics and attitudes towards the proposed initiative. A two-way ANCOVA with promoter gender and promoter race as the fixed factors on attitude ratings revealed a main effect of promoter race. Specifically, initiatives proposed by White authors elicited more positive attitude ratings (M = 5.51, SD = 1.29) that initiatives proposed by Black authors (M = 5.22, SD = 1.51), F(1, 335) = 6.87, MSE = 10.12, p = .043, η2p = .02 . No significant differences were found on attitude ratings depending on promoter gender, F(1, 335) = .08, p = .784. However, the effect of manipulated race on attitude ratings (even when controlling for participant SDO scores) suggests support for Hypothesis 4. A marginal interaction between promoter race and promoter gender also emerged, as the effect of promoter gender appeared to depend on promoter race, F(1, 335) = 2.92, MSE = 4.29, p = .089. Specifically, as displayed in Figure 2, Black male and White male promoters elicited relatively equal attitude ratings toward the proposed diversity initiative; however, White female promoters appeared to elicit more positive attitude ratings as compared to Black female promoters. In fact, White female promoters appeared to elicit the highest attitude ratings overall, while Black female promoters elicited the lowest attitude ratings of the demographic four conditions. While this interaction did not reach statistical significance, it is interesting to note as a potential trend. Hypothesis 5. Hypothesis 5 predicted that promoter race and promoter gender would be related to participant perceptions of promoter self-interest. A two-way ANCOVA with promoter gender and promoter race as the fixed factors on ratings of perceived promoter self-interest was conducted, and revealed a main effect of promoter race, F(1, 333) = 4.21, MSE = 3.83, p = .041, η2p = .01. Specifically, Black initiative promoters were rated as more self-interested (M = 3.10, 29 SD = 1.05) than were White initiative promoters (M = 2.91, SD = 0.95). No main effect of promoter gender [F(1, 333) = 0.49, p = .484] or interaction between promoter gender and promoter race were identified, F(1, 344) = 0.17, p = .680. However, the effect of manipulated race on perceived promoter self-interest ratings provides support for Hypothesis 5. Hypothesis 6. Hypothesis 6 stated that ratings of perceived promoter self-interest would negatively predict attitudes towards the proposed diversity initiative. Using simple regression with self-interest ratings as the predictor and initiative attitude ratings as the outcome, I found support for this hypothesis, as self-interest ratings had a negative predictive relationship with initiative attitude ratings, b = -.61, SE = .07, p < .001, R2 = .19. Further, an examination of both perceived personal relevance and perceived promoter self-interest jointly predicting initiative attitudes showed that both predictors remain significant even when accounting for the other predictor, as seen in Table 2. Together, perceived personal relevance and perceived promoter self-interest accounted for almost two thirds of the variance in initiative attitudes (R2 = .65). Thus, Hypothesis 6 was supported. Hypothesis 7. Given the effect of promoter race on perceived promoter self-interest ratings, and the significant relationship between promoter self-interest ratings and participant attitudes towards the proposed initiative, I used Hayes’ PROCESS Macro to test for mediation, as proposed by Hypothesis 7. Using PROCESS Model 4 with promoter race as the predictor, self-interest ratings and the mediator, and initiative attitudes as the outcome, I found support for Hypothesis 7 as the model was significant, F(3, 329) = 58.75, MSE = 1.32, p < .001, and accounted for a large proportion of variance, R2 = .35. Specifically, using 5000 bootstrapped samples, results revealed a significant indirect effect of promoter race on initiative attitudes 30 through perceived promoter self-interest ratings, b = -.10, SE = .05, p = .043, 95% CI = [-.20, - .01]. Therefore, Hypothesis 7 was supported. Hypothesis 8. Hypothesis 8 predicted that attitudes towards the proposed initiative would predict behavioral actions in support of the initiative. Two variables were examined in relation to behavioral support for the initiative: voting behavior (yes/no in support of the initiative) and money allotment. To examine if initiative attitudes significantly predicted voting behavior, a logistic regression was conducted with initiative attitudes as the predictor and voting behavior as the outcome, finding that indeed initiative attitudes significantly predicted voting behavior, b = - 1.74, SE = .18, p < .001, Cox and Snell R2 = .45, Nagelkerke R2 = .63. Specifically, those who voted in support of the initiative reported roughly 65% more positive attitudes towards the proposed initiative than those who voted against the proposal. Hypothesis 8 was further examined by testing whether initiative attitudes predicted money allotment. Using simple regression with initiative attitudes as the predictor and number of dollars allotted to the initiative as the outcome, I found support for Hypothesis 8 such that positive initiative attitudes predicted the amount of money allotted to the diversity initiative, b = 581.65, SE = 42.66, p < .001, R2 = .35. Specifically, results show that those one standard deviation above the mean on initiative attitudes allotted $2682.88 (SD = 1258.51) on average to the proposed diversity initiative, as compared to participants one standard deviation below the mean, who allotted on average $457.29 (SD = 836.69). Therefore, Hypothesis 8 was supported across both behavioral measures. Expanded Path-Model: Given the support for Hypotheses 4 through 8, a path-model linking the race manipulation to behavioral outcomes (as depicted in Figure 3) was examined in Mplus. As in 31 previous analyses, SDO was modeled as a covariate for these analyses. Path estimates, standard errors, and p-values for the models examining both voting behavior and charity donation behavior are listed in Tables 3 and 4, respectively. Looking to these path estimates, the positive and significant relationships between promoter race and perceived promoter self-interest show that White promoters were consistently rated as less self-interested than Black promoters. The negative relationships between both self-interest perceptions and promoter race with initiative attitudes suggests a partial-mediation relationship, such that promoter race impacted initiative attitudes both directly, as well as through the self-interest mediator. Finally, the relationships between initiative attitudes and behavioral outcomes were significant across both models in the expected directions. Looking to fit statistics, both models fit the data reasonably well. The fit statistics for the first model examining voting behavior as the behavioral outcome were all in line with traditional acceptability standards (RMSEA = .07, CFI = .99, TLI =.96). Similarly, the fit statistics of the second model including charity donation as the behavioral outcome were also in line with traditional cutoffs (RMSEA = .07, CFI = .97, TLI = .92). Therefore, these values suggest support for both path models. Exploratory Analyses: Additional Variables. A number of additional outcomes were included in the data set to examine whether our manipulated promoter demographics had any impact on variables outside of our hypothesized model. To explore this possibility, a MANCOVA with promoter gender and promoter race as the fixed factors, and promoter persuasiveness, promoter credibility, the Attitudes Towards Diversity Scale (ATDS), and a charitable donation measure as the outcomes, was examined. Multivariate tests revealed an overall main effect of promoter race, F(4, 301) = 32 4.07, p = .003, η2p = .05. The univariate tests revealed that this difference was due to the charitable donation measure. Specifically, participants in the White promoter condition were more likely to donate to a diversity-related charity over other charities (M = 0.40, SD = 0.49) as compared to participants in the Black promoter condition (M = 0.27, SD = 0.45), F(1, 305) = 6.16, MSE = 1.29, p = .014, η2p = .02. No other significant differences were found across these outcomes, including any interactions between promoter race and promoter gender. That is, promoter race and gender did not affect how persuasive or credible the promoter was rated by participants. Further, promoter race and gender had no effect on ATDS ratings. Participant Demographics. Exploratory analyses revealed significant differences on a number of outcomes dependent on participant demographics. Specifically, a MANCOVA with participant demographics (i.e. White male vs. Female or racial minority) as the fixed factor, and initiative attitudes, promoter self-interest perceptions, proposal voting behavior, money allotment, and charitable donation behavior as the outcomes, was conducted. Multivariate tests revealed an overall main effect of participant demographics, F(5, 310) = 2.89, p = .015, η2p = .04. An examination of the univariate tests suggest that these differences were due to the initiative attitudes, promoter self-interest perceptions, and voting behavior measures. Specifically, White male participants overall had less positive attitudes toward the diversity initiative (M = 5.09, SD = 1.64) as compared to participants with a minority identity (M = 5.57, SD = 1.19), F(1, 314) = 8.85, MSE = 12.96, p = .027, η2p = .03. Further, White male participants rated the promoter as more self-interested (M = 3.20, SD = 1.01) than other participants (M = 2.86, SD = 0.97), regardless of the promoter’s demographics, F(1, 314) = 7.19, MSE = 6.60, p = .022, η2p = .02 . Finally, White male participants were more likely to vote against the proposed initiative (M = 1.40, SD = 0.49) than other participants (M = 1.25, SD = 0.43), F(1, 314) = 7.83, MSE = 1.45, p 33 = .005, η2p = .02. No significant differences were found on either the money allotment or charitable donation measure. 34 CHAPTER 5 DISCUSSION The purpose of this study was to examine whether the demographics of an individual proposing a diversity initiative could impact subsequent attitudes and behavior in support of the initiative. Using logic grounded in the Elaboration Likelihood Model (Petty & Cacioppo, 1986), and the Attributional Analysis of Persuasion (Eagly & Chaiken, 1975), I predicted that majority- group individuals proposing diversity initiatives would be more likely to elicit positive initiative attitudes and behavioral support as compared to female or racial minority individuals proposing the same initiative. Results suggest that promoter race, but not promoter gender, may be influential in fostering positive attitudes and behavior in support of the initiative. A specific discussion of the results of each hypothesis follows. Hypotheses 1 through 3 reflect the logic rooted in the Elaboration Likelihood Model. That Hypothesis 1 (the prediction that promoter demographics will affect perceived personal relevance for majority-group participants) was not supported suggests that promoter demographics may not be particularly impactful in altering personal relevance perceptions. However, given the finding that personal relevance perceptions did vary by participant demographics, such that White male participants found the proposal less personally relevant than female and racial minority participants, personal relevance may remain an important variable to consider when seeking to maximize support for a diversity initiative. Further, as perceptions of personal relevance were indeed predictive of initiative attitudes as was posited in Hypothesis 2, alternative levers of increasing personal relevance should be explored. For example, perhaps individuals could learn how diversity initiatives are relevant to them through diversity training or other educational efforts. Although the promoter demographic manipulations in this study may 35 not have related to personal relevance perceptions as predicted, the findings from this research do suggest the significance of personal relevance when considering enacting proposed organizational initiatives. Alternatively, the lack of support for Hypothesis 1 may suggest not that our demographic manipulations were ineffective, but rather that participants were not using the central route of processing when considering the proposed initiative. The support of Hypothesis 4 (the prediction that promoter demographics will directly relate to attitudes towards the initiative) may suggest that participants were instead using the peripheral route; this would involve participants not deeply considering the proposal at hand, but instead forming attitudes based on simple surface cues (Petty & Cacioppo, 1986). Indeed in organizations, it is possible that wide variability in deep consideration exists when considering organizational initiatives such as those of focus here. Specifically, one could expect that the majority of employees who hear about an organizational initiative are not likely to deeply process, as compared to those who are intimately involved in developing and enacting the initiative, who may alternatively process the situation more in depth. With this in mind, the support of Hypothesis 4 suggests that when quickly considering a proposed diversity initiative, individuals are indeed influenced by the promoter’s race. Interestingly, this direct relationship between promoter race and initiative attitudes (as well as the indirect relationship between race and attitudes through self-interest perceptions) was found while controlling for participant SDO scores. Given the strong correlations between SDO and all outcome measures as seen in Table 1, the results suggest that the effect of promoter race on our dependent variables is present irrespective of participant preferences for inequality among social groups. As SDO has often been used as a proxy for general bias given the score’s consistent relationship with racial-ethnic prejudice and sexism (Pratto et al., 1994; Sidanius & 36 Pratto, 1993), this suggests that participant bias is not the only driver of our observed outcome differences, but that other mechanisms, such as our hypothesized self-interest mediator, are also important factors. Further, one could wonder whether these direct effects of promoter demographics were the result of White and male privilege (McIntosh, 1989). Generally, discussions of demographic privilege outline how individuals of certain groups are afforded a level of power and opportunity not available to those outside of those groups. One would expect that if privilege were the underlying mechanism of our findings, that the White male promoter would elicit the most positive reactions, as theoretically he would be the most privileged of the demographic conditions. However, as the White male promoter was not viewed the most positively (in fact, marginal interactive effects suggest the White female promoter as the most effective), it seems that participants were not reacting simply to power and privilege. Further, given the evidence found in support of the self-interest mediator, it seems likely that other mechanisms are at play to explain the findings identified in this study. Hypotheses 5 through 7 outline predictions based on logic grounded in the Attributional Analysis of Persuasion. Hypothesis 5 specifically posited that promoter demographics relate to perceptions of self-interest, such that majority-group promoters will be perceived as less self- interested than minority-group promoters. This hypothesis was supported, as promoter race was indeed related to self-interest perceptions. Further, the support of Hypothesis 6 displays the important link between these self-interest perceptions and subsequent initiative attitudes, as self- interest ratings were negatively predictive of attitudes towards the proposed diversity initiative. In fact, self-interest perceptions were found to mediate the relationship between promoter race and initiative attitudes as predicted in Hypothesis 7, and these attitudes were linked to behavioral 37 support for the initiative as detailed in Hypothesis 8. The path model linking the race manipulations to initiative voting behavior was examined (Figure 3), and fit statistics suggest support for the model. It is also worth noting that promoter race may affect variables outside of the proposed model. While exploratory analyses revealed no differences in promoter credibility, promoter persuasiveness, and ATDS ratings, charitable donation behavior was found to differ depending on the race of the initiative promoter. Specifically, findings suggest that White promoters were more likely to elicit donations to a diversity-focused charity than were Black promoters. Consequently, it appears the effects of promoter demographics may extend beyond the context of the proposed initiative to more general diversity-related behavior. Further, the lack of differences in promoter credibility and promoter persuasiveness may underscore the importance of the mediators identified in the primary model, as these exploratory variables which may have acted as alternative explanations for the reported findings were found to be unaffected by the manipulations. Collectively, these exploratory analyses suggest that although demographic manipulations may not be strong enough to affect stable attitudes such as ATDS ratings, they may in fact transcend the situation of the proposed initiative to impact more general diversity- related behavior. An interesting point is that direct and indirect effects of promoter demographics on outcomes of interest were found only for the race manipulation, and not the gender manipulation. One potential explanation could be that given racial minorities’ lack of representation in upper level roles compared to both White men and women, perhaps racial minorities are perceived as having the most to gain from diversity initiatives. For example, a recent report from the U.S. Equal Employment Opportunity Commission (EEOC) suggests that while women now comprise 38 close to 40% of senior and mid-level managerial roles, less than 6% of those roles are filled by racial minorities (EEOC, 2015). Therefore, given the larger gap in representation experienced by minorities compared to women, it is possible that Black promoters were perceived as more self- interested in proposing a diversity initiative, as racial minorities may appear to have more to gain from such programs. It is also possible that gender stereotypes may be operating so as to counteract any potential differences in self-interest perceptions between male and female promoters. Specifically, given the long-documented stereotypes associating women with communal characteristics such as kindness, compassion, and warmth (Rudman & Glick, 2001; Fiske, Cuddy, Glick & Xu, 2002), perhaps women proposing diversity initiatives were perceived as selfless and altruistic rather than acting in a way to help themselves. This could potentially explain the marginal interaction found between promoter gender and promoter race on initiative attitudes, as the White female promoter was found to elicit the most positive attitude ratings towards the initiative than any other promoter condition. An interesting additional set of findings suggests not only do promoter demographics relate to attitudes towards a proposed diversity initiative, but so too do the demographics of individuals evaluating the initiative. Specifically, the data suggest that White male participants rated initiative promoters as more self-interested, held less positive attitudes towards the diversity initiative, and were less likely to vote in favor of the initiative as compared to female and racial minority participants. Such findings are in line with previous studies noting the differences between White males’ and other workers’ opinions towards workplace diversity (Kossek & Zonia, 1993; Mor Barak et al., 1998), and reiterate a need to focus on this sub-group when seeking support for diversity-related causes. Although the effect of personal relevance was 39 not found to be unique to White male participants as hypothesized in this study, future research should continue to explore ways in which to target individuals of this sub-group to garner as much support as possible for organizational diversity initiatives. Theoretical and Practical Implications With respect to theory, results of the current study suggest support for the Attributional Analysis of Persuasion, as the hypotheses grounded in logic inspired by the AAP were ultimately supported. The present findings also extend AAP by examining a slightly different mechanism than that which is traditionally studied. As studies applying AAP tends to define message expectancy as the mediating variable, the current study goes a step further by exploring promoter self-interest as the hypothesized variable contributing to differential message expectancy. Future research should incorporate both self-interest and message expectancy into a single model, and examine whether self-interest is indeed the explanatory mechanism for differences in message expectancies. Practically, the importance of self-interest perceptions as displayed by the current data has implications for how proposals of all types may best be presented. While the present findings suggest that certain individuals are viewed as more self-interested than others when proposing a diversity initiative, perhaps these proposals may be crafted in such a way so as to counter self- interest perceptions. By potentially acknowledging one’s state of benefit, or alternatively removing oneself as a beneficiary from a proposed policy-change, perceptions of self-interest may be mitigated. However, future research is needed to conclusively identify the best strategies one could use to reduce self-interest perceptions. The current findings also have important implications for the budding literature on workplace allyship (Sabat, Martinez & Wessel, 2013; Sabat et al., 2014). While the term “ally” 40 has traditionally come from the lesbian, gay, bisexual and transgender (LGBT) community, referring to heterosexual and cis-gendered individuals striving to advocate on behalf of non- heterosexual minorities (Washington & Evans, 1991), the term has recently been used to refer to any member of a non-stigmatized group who engages in supportive behaviors on behalf of those with stigmatized identities (Sabat et al., 2013). The findings here not only underscore the utility of out-group allies, but also reveal a mechanism through which allyship may be effective. Specifically, the support found for the perceived self-interest mediator helps to explain how and why allies are important in advocating on behalf of those who are stigmatized. However, the findings here should not be interpreted as suggesting only those of majority-group, non-stigmatized identities should propose diversity-related initiatives. Indeed, such a proposition would take away agency from those in minority-groups, which would in fact be the opposite outcome of the diversity initiatives being proposed. Of course, women, racial minorities, and minorities of all identities should be foremost involved in their own advocacy. Rather than suggesting otherwise, the current results imply that those of majority-group identities may add value when joining such causes, although additional research is needed to speak definitively on the effects of multiple promoters of varying demographics. The hope is that the present findings encourage those of non-stigmatized identities to add their voices in support of diversity-related causes, while making sure not to overpower or silence the voices of those advocating on their own behalf. Limitations and Future Directions A number of limitations associated with this study should be noted. For instance, the experimental paradigm may be limiting, such that central processing of the proposed message may be unlikely for participants. Given the experiment’s level of artificiality, participants may be 41 instead reacting via surface-level processing, as the lack of lived experiences and any true stakes in the scenario may prevent deep-level processing from occurring. This may also explain the small size of the observed effects; perhaps in a real setting, one may find even stronger effect sizes, given a deeper level of processing than that which likely occurred in this experimental paradigm. Further, the context of this experiment was such that participants only considered one organizational initiative for the sake of simplicity. A more externally valid procedure would involve participants evaluating multiple initiative proposals, and would allow researchers to examine how the diversity-initiative proposal fared against other proposed initiatives depending on the demographics of those promoting each cause. It is possible that in situations of comparison, promoter demographics may play an even more important role for diversity initiative outcomes; as raters have the opportunity to select alternative proposals, research on attributional ambiguity would suggest they may be more likely to act on any negativity they may harbor towards organizational diversity, as the situation of choice would keep their true motives concealed (Snyder, Kleck, Strenta & Mentzer, 1979). Future research should investigate this line of examination in order to more fully understand how the relationships and mechanisms explored in this study may translate to real-world situations. It should also be noted that ratings of all measures were collected at a single time point, suggesting that common method variance may be a concern. However, given the experimental nature of the study, the links between our manipulations and subsequent outcomes can be interpreted as causal rather than simply interrelated. To address continued concerns regarding common method variance, future research could separate the experimental manipulation and 42 collection of ratings into separate time points, and see whether the pattern of findings identified here continues to hold under these circumstances. Further, future research should attempt to explore whether the type of diversity initiative plays a role in the processes examined here. While the current study involved the proposal of a task force seeking to enact a number of diversity initiatives within an organization, it would be worthwhile to explore whether certain diversity initiatives are viewed differently from one another. For instance, proposals for diverse selection policies may elicit a wider range of reactions than proposals for organizational diversity training. Consequently, research is needed to explore how generalizable the present findings are across a variety of potential organizational diversity initiatives (i.e. diversity training, diversity hiring policies, employee resource groups, etc.), given the potential that the type of initiative may elicit different employee reactions. Finally, while the identities examined here were limited to race and gender, future research should explore whether the present identified trends may extend to situations involving other identities, such as religious affiliation, sexual orientation, and immigration status. Such studies would help inform whether the explored mechanisms generalize to all situations of stigmatized and non-stigmatized group members, or alternatively whether the current findings are specific to promoter race. Future research should continue to examine promoter race as well; as the current study compared only White and Black promoters, examinations involving other ethnic groups could speak to further generalizability of the present findings. Conclusion As organizations continue to pursue implementing diversity initiatives, one question lies in how organizations can propose and present such efforts so as to most likely maximize support. The current thesis sought to address this question by exploring whether the demographics of the 43 individual proposing the initiative may impact subsequent attitudinal and behavioral initiative support. Using logic grounded in the Elaboration Likelihood Model and the Attributional Analysis of Persuasion, I hypothesized that individuals of majority-group identities may elicit more supportive initiative attitudes and behaviors through perceptions of participant personal relevance and promoter self-interest. Results suggest that indeed, White promoters were perceived as less self-interested than Black promoters promoting a diversity initiative, which in turn predicted more positive initiative attitudes and behavioral support for the initiative. Consequently, the findings discussed here have important implications for both research and practice, as not only do the results suggest support for a mechanism explaining differential outcomes based on promoter demographics, but also present a lever through which organizations may increase organizational support for proposed diversity initiatives. 44 APPENDICES 45 Table 1. Means, standard deviations, and bivariate correlations between manipulations and measured outcomes. APPENDIX A Tables and Figures 1 Promoter Race 2 Promoter Gender 3 Initiative Attitudes Personal Relevance 4 5 Self-Interest 6 Credibility 7 Persuasion 8 Voting Behavior 9 Money Allotment 10 SDO 11 ATDS 12 Charity Donation M 0.50 0.50 5.33 4.89 3.06 5.51 3.89 0.70 SD 0.50 0.50 1.40 1.56 0.99 1.13 0.93 0.46 1973.79 1424.33 2.51 2.38 0.36 1.32 0.64 0.48 NOTE: *p < .05; †p < .01. Scale reliabilities in parentheses on the diagonal. Race coded as 0 = White, 1 = Black. Gender coded as 0 = Female, 1 = Male. Voting behavior coded as 0 = Against, 1 = In favor. Charity donation coded as 0 = Non-diversity charity, 1 = Diversity-related charity. 1 -- .01 -.11* 2 -- 0 3 (.96) 4 5 6 7 8 9 10 11 -.07 .09 .06 -.07 -.09 -.09 -.02 -.04 -.13† .02 .03 -.04 -.03 .04 .07 -.02 -.03 .01 .79† -.38† .67† .87† .68† .57† -.46† -.52† .25† (.95) -.33† .55† .71† .67† .57† -.41† -.44† .24† (.93) -.43† -.34† -.30† -.18† .32† .50† -.11* (.90) .67† .29† .26† -.50† -.52† .14† (.93) .36† .32† -.51† -.57† .16† -- .61† -.26† -.30† .28† -- -.21† -.24† .19† (.95) .74† -.12* (.92) -.19† 46 t Sig. .000 .000 Table 2. Coefficients, standard errors, and significance values of personal relevance and promoter self- interest predicting initiative attitudes. Personal Relevance Promoter Self-Interest S.E. b .50 .03 -0.14 -4.02 0.74 20.67 b 0.66 -0.20 47 Table 3. Path estimates, standard errors, and p-values of path-model linking promoter race to voting behavior. Path Estimate SE P-Value 0.043 0.21 0.26 0.10 0.04 Self Interest as Predicted By: Promoter Race Social Dominance Orientation Initiative Attitudes as Predicted By: Self-Interest Perceptions Promoter Race Social Dominance Orientation Voting Behavior as Predicted By: Initiative Attitudes NOTE: Voting behavior was coded as 1 = In favor of proposed initiative, 2 = Against proposed initiative. -0.49 -0.24 -0.44 0.07 0.14 0.05 0 0 0 0 -0.67 0.04 0.075 48 Table 4. Path estimates, standard errors, and p-values for path-model linking promoter race to charity donation behavior. Path Estimate SE P-Value 0.043 0.21 0.26 0.10 0.04 Self Interest as Predicted By: Promoter Race Social Dominance Orientation Initiative Attitudes as Predicted By: Self-Interest Perceptions Promoter Race Social Dominance Orientation Charity Donation Behavior as Predicted By: Initiative Attitudes NOTE: Charity donation behavior was coded as 0 = Toward a non-diversity-related charity, 1 = Toward a diversity-related charity. -0.42 -0.33 -0.48 0.07 0.13 0.05 0 0 0 0 0.011 0.30 0.06 49 Figure 1. Proposed model linking promoter demographics to attitudes and behavioral support for proposed diversity initiative. 50 Figure 2. Marginal interaction between promoter gender and promoter race on initiative attitudes. s e d u t i t t A e v i t a i t i n I 6 5.5 5 4.5 Male Female Promoter Demographics White Black 51 Figure 3. Path-model tested in Mplus. 52 APPENDIX B Measures 1. What was the race of the author whose proposal you just read? Manipulation Checks 2. What was the gender of the author whose proposal you just read? a. White b. Black c. Hispanic d. East Asian a. Male b. Female 3. What was the topic of the proposed initiative you just read? a. Health and wellness incentive plan b. Diversity task force c. Upgrade of company desktops 53 Attitudes Towards the Proposed Initiative Scale Instructions: Please rate the extent to which the following adjectives most closely reflect your attitudes towards the proposed initiative. 1. Not valuable _ _ _ _ _ _ _ Valuable 2. Not needed _ _ _ _ _ _ _ Needed 3. Irrelevant _ _ _ _ _ _ _ Relevant 4. Implausible _ _ _ _ _ _ _ Plausible 5. Impractical _ _ _ _ _ _ _ Practical 6. Unlikely to be implement _ _ _ _ _ _ _ Likely to be implemented 7. Not worthwhile _ _ _ _ _ _ _ Worthwhile 8. Low quality _ _ _ _ _ _ _ High quality 9. Not beneficial _ _ _ _ _ _ _ Beneficial 54 Personal Relevance Measure [Adapted from Zhao & Peterson (2017) and Wells et al. (1971)] Using the provided scale, please indicate the extent to which you agree with each of the following statements. Scale: 1 = “Strongly Disagree,” 2 = “Disagree,” 3 = “Disagree Somewhat,” 4 = “Neither Agree nor Disagree,” 5 = “Agree Somewhat,” 6 = “Agree,” 7 = “Strongly Agree” 1. The proposal is relevant to my life. 2. The proposal grasped my attention. 3. The proposal said something important to me. 4. The proposal is meaningful to me. 5. The proposal is worth remembering. 6. The proposal is valuable to me. 55 Perceived Self-Interest Measure [Adapted from Gerbasi & Prentice (2013) and Tseng & Fan (2011)] Scale: 1 = “Strongly disagree,” 2 = “Disagree,” 3 = “Neither agree nor disagree,” 4 = “Agree,” 5 = “Strongly agree” 1. The author of this proposal is looking for opportunities to achieve higher social status. 2. The author of this proposal is looking for ways to get ahead. 3. The author of this proposal is keeping an eye out for his/her interests. 4. The author of this proposal, by and large, is pursuing his/her own interest. 5. The author of this proposal is protecting his/her own interests above other considerations. 6. The author of this proposal is acting self-interestedly. 56 Perceived Persuasiveness Scale Instructions: Please rate the extent to which the following adjectives most closely reflect your attitude toward the proposal you just read. 1. Not persuasive _ _ _ _ _ Persuasive 2. Unconvincing _ _ _ _ _ Convincing 3. Poorly argued _ _ _ _ _ Well argued 4. Poorly presented _ _ _ _ _ Well presented 5. Did not make a good case _ _ _ _ _ Made a good case 6. Illogical _ _ _ _ _ Logical 57 Ethos/Credibility Scale (McCroskey & Teven, 1999) Please rate your impression of the author of the proposal by considering the following adjectives. The closer you rate the author to an adjective within the pairing, the more certain you are of your evaluation. 1. Intelligent _ _ _ _ _ _ _Unintelligent 2. Untrained _ _ _ _ _ _ _ Trained 3. Inexpert _ _ _ _ _ _ _ Expert 4. Informed _ _ _ _ _ _ _ Uninformed 5. Incompetent _ _ _ _ _ _ _ Competent 6. Bright _ _ _ _ _ _ _ Stupid 7. Honest _ _ _ _ _ _ _ Dishonest 8. Untrustworthy _ _ _ _ _ _ _ Trustworthy 9. Honorable _ _ _ _ _ _ _ Dishonorable 10. Moral _ _ _ _ _ _ _ _ _ _ Immoral 11. Unethical _ _ _ _ _ _ _ Ethical 12. Phoney _ _ _ _ _ _ _ Genuine NOTE: Competence subscale is comprised of items 1-6; Trustworthiness subscale is comprised of items 7-12. 58 Attitudes Toward Diversity Scale (Montei, Adams & Eggers, 1996) Scale: 1 = “Strongly Disagree,” 2 = “Disagree” 3 = “Neither Agree nor Disagree,” 4 = “Agree,” 5 = “ Strongly Agree” 1. All in all, I would say that minority workers are just as productive as other workers.* 2. I often pick up the slack for some of my female coworkers who are less productive. 3. Sometimes I have to compensate for the lack of productivity of minority workers. 4. The most qualified workers in my job seem to be male. 5. I find that minority workers seem to be less productive on average. 6. The minorities in this organization have a greater degree of difficulty getting along with others. of my disapproval.* 7. If a member of my work group were prejudiced, he or she would be less likely to fit in.* 8. If one of my coworkers were racist, I would confront that person and let him or her know 9. Workers who are prejudiced have no place in my organization.* 10. I do not feel comfortable with coworkers who are racist.* 11. I feel that women have a more difficult time handling positions of authority relative to 12. I would feel just as comfortable with a Black or Hispanic supervisor as I do with a White men. supervisor.* supervisors. supervisor. supervisors. 13. It seems that minorities in supervisory positions are ineffective relative to other 14. Most of the women in management positions do an outstanding job.* 15. I feel that diversity is good for my organization even if it means I will have a supervisor who is a minority. 16. Relative to male supervisors, female supervisors seem to be less effective. 17. Under most circumstances, I would prefer a male supervisor. 18. I would feel less comfortable with a female supervisor than I would with a male 19. Most of the minority supervisors in my organization possess the same leadership qualities as do those supervisors who are White.* 20. It seems as if some of the women I work with need to be more assertive to be effective 21. I know some workers who would be fired if they were not minorities. 22. It does not bother me that some preferential hiring goes on because we need more of a mix in my organization.* test scores to even things out. * 23. Because some tests are known to be biased towards minorities, I feel it is alright to adjust 24. I am against hiring by quotas even when done out of necessity. 25. I know many more qualified White males who should have been hired instead of some of the minorities that have been hired lately. 26. We would have a more creative work environment if more women and minorities were hired.* 27. I feel it is wrong for an organization to have two sets of test scores for minorities and nonminorities, even when the test is somewhat biased. 59 28. Some of the members of my organization were hired just because they are women. 29. I feel that increasing the hiring of women and minorities can only help my organization.* 30. Some of the workers in my organization were only hired because they are minorities. Note: * indicates reverse-coded item. 60 Social Dominance Orientation (Pratto, Sidanius, Stallworth & Malle, 1994) Which of the following statements do you have a positive or negative feeling towards? For each statement, indicate a number from 1 to 7 which represents the degree of your positive or negative feeling. Scale: 1 = “Extremely Negative,” 2 = “Somewhat Negative,” 3 = Slightly Negative,” 4 = “Neutral,” 5 = “Slightly Positive,” 6 = “Somewhat Positive,” 7 = “Extremely Positive” 1. We should strive to make incomes as equal as possible.* 2. Group equality should be our ideal.* 3. It’s okay if some groups have more of a chance in life than others. 4. To get ahead in life, it is sometimes necessary to step on other groups. 5. We should do what we can to equalize conditions for different groups.* 6. It’s probably a good thing that certain groups are at the top and others are at the bottom. 7. Inferior groups should stay in their place. 8. We would have fewer problems if groups were treated more equally.* 9. It would be good if groups could be equal.* 10. In getting what you want, it is sometimes necessary to use force against other groups. 11. All groups should be given an equal chance in life.* 12. If certain groups stayed in their place, we would have fewer problems. 13. We should strive for increased social equality.* 14. Sometimes other groups must be kept in their place. 15. Some groups of people are simply inferior to other groups. 16. No one group should dominate in society. * Note: * indicates reverse-scored items 61 Author Description Stimuli. APPENDIX C Experimental Stimuli 62 63 Proposal Stimulus. 64 REFERENCES 65 REFERENCES Ajzen, I. (1991). The theory of planned behavior. Organizational Behavior and Human Decision Processes, 50(2), 179-211. doi:https://doi.org/10.1016/0749-5978(91)90020-T Ajzen, I., & Fishbein, M. (1977). Attitude-behavior relations: A theoretical analysis and review of empirical research. Psychological Bulletin, 84(5), 888-918. doi:http://dx.doi.org/10.1037/0033-2909.84.5.888 Avery, D. R. (2003). Reactions to diversity in recruitment advertising- are differences black and white? Journal of Applied Psychology, 88(4), 672-679. doi:http://dx.doi.org/10.1037/0021- 9010.88.4.672 Avery, D. R., Hernandez, M., & Hebl, M. R. (2004). Who’s watching the race? Racial salience in recruitment advertising. Journal of Applied Social Psychology, 34(1), 146-161. doi:10.1111/j.1559-1816.2004.tb02541.x Avery, D. R., & McKay, P. F. (2006). Target practice: An organizational impression management approach to attracting minority and female job applicants. Personnel Psychology, 59(1), 157-187. doi:10.111/j.1744-6570.2006.00807.x Bendick Jr, M., Egan, M. L., & Lofhjelm, S. M. (2001). Workforce diversity training: From anti- discrimination compliance to organizational development. Human Resource Planning, 24(2), 10-25. Bezrukova, K., Spell, C. S., Perry, J. L., & Jehn, K. A. (2016). A meta-analytical integration of over 40 years of research on diversity training evaluation. Psychological Bulletin, 142(11), 1227-1274. Burnkrant, R. E. & Unnava, H. R. (1989). Self-referencing: A strategy for increasing processing of message content. Personality and Social Psychology Bulletin, 15(4), 628-638. doi:10.1177/0146167289154015 Cacioppo, J. T., Petty, R. E., & Sidera, J. A. (1982). The effects of a salient self-schema on the evaluation of proattitudinal editorials: Top-down versus bottom-up message processing. Journal of Experimental Social Psychology, 18(4), 324-338. doi: http://doi.org/10.1016/0022-1031(82)90057-9 Chaiken, S. (1979). Communicator physical attractiveness and persuasion. Journal of Personality and Social Psychology, 37(8), 1387-1397. doi:http://dx.doi.org/10.1037/0022- 3514.37.8.1387 66 Combs, G. M., & Luthans, F. (2007). Diversity training: Analysis of the impact of self‐ efficacy. Human Resource Development Quarterly, 18(1), 91-120. doi:10.1002/hrdq.1193 Czopp, A. M., & Monteith, M. J. (2003). Confronting prejudice (literally): Reactions to confrontations of racial and gender bias. Personality and Social Psychology Bulletin, 29(4), 532–544. doi:http://doi.org/10.1177/0146167202250923 Dobbin, F., & Kalev, A. (2016, July-August). Why diversity programs fail. Harvard Business Review. Retrieved from: https://hbr.org/2016/07/why-diversity-programs-fail Dovidio, J. F., Kawakami, K., & Gaertner, S. L. (2002). Implicit and explicit prejudice and interracial interaction. Journal of Personality and Social Psychology, 82(1), 62-68. Eagly, A., & Chaiken, S. (1975). An attribution analysis of the effect of communicator characteristics on opinion change: The case of communicator attractiveness. Journal of Personality and Social Psychology, 32, 136-144. Eagly, A. H., Wood, W., & Chaiken, S. (1978). Causal inferences about communicators and their effect on opinion change. Journal of Personality and Social Psychology, 36(4), 424-435. doi:http://dx.doi.org/10.1037/0022-3514.36.4.424 Equal Employment Opportunity Commission (2015). 2015 job patterns for minorities and women in private industry. Retrieved from https://www1.eeoc.gov/eeoc/statistics/employment/jobpat-eeo1/2015/index.cfm#select_label on January 25, 2018. Fazio, R. H., Chen, J. M., McDonel, E. C., & Sherman, S. J. (1982). Attitude accessibility, attitude-behavior consistency, and the strength of the object-evaluation association. Journal of Experimental Social Psychology, 18(4), 339-357. doi:https://doi.org/10.1016/0022- 1031(82)90058-0 Fiske, S. T., Cuddy, A. J., Glick, P., & Xu, J. (2002). A model of (often mixed) stereotype content: Competence and warmth respectively follow from perceived status and competition. Journal of Personality and Social Psychology, 82(6), 878-902. Gerbasi, M. E., & Prentice, D. A. (2013). The self- and other-interest inventory. Journal of Personality and Social Psychology, 105(3), 495-514. Gilbert, J. A., Stead, B. A., & Ivancevich, J. M. (1999). Diversity management: A new organizational paradigm. Journal of Business Ethics, 21(1), 61-76. doi:10.1023/A:1005907602028 Goldberg, P. (1968). Are women prejudiced against women? Trans-Action, 5(5), 28–30. doi:http://doi.org/10.1007/BF03180445 67 Gruber, K. J., & Gaebelein, J. (1979). Sex differences in listening comprehension. Sex Roles, 5(3), 299–310. doi:http://doi.org/10.1007/BF00287397 Gulker, J. E., Mark, A. Y. & Monteith, M. J. (2013). Confronting prejudice: The who, what, and why of confrontation effectiveness. Social Influence 8(4), 280-293. doi:http://dx.doi.org/10.1080/15534510.2012.736879 Guynn, J. (2015, May 6). Exclusive: Google raising stakes on diversity. USA Today. Retrieved from: https://www.usatoday.com/story/tech/2015/05/05/google-raises-stakes-diversity- spending/26868359/ Hayes, A. F. (2013). Introduction to mediation, moderation, and conditional process analysis. New York, NY: Guilford Press. Hebl, M. R., Foster, J. B., Mannix, L. M., & Dovidio, J. F. (2002). Formal and interpersonal discrimination: A field study of bias toward homosexual applicants. Personality and Social Psychology Bulletin, 28(6), 815-825. Hekman, D. R., Johnson, S. K., Foo, M. D. & Yang W. (2016). Does diversity-valuing behavior results in diminished performance ratings for nonwhite and female leaders? Academy of Management Journal. doi:10.5465/amj.2014.0538 Herdman, A. O., & McMillan-Capehart, A. (2010). Establishing a diversity program is not enough: Exploring the determinants of diversity climate. Journal of Business and Psychology, 25(1), 39-53. doi:10.1007/s10869-009-9133-1 Highhouse, S., Stierwalt, S. L., Bachiochi, P., Elder, A. E., & Fisher, G. (1999). Effects of advertised human resource management practices on attraction of African American applicants. Personnel Psychology, 52(2), 425-442. doi:10.1111/j.1744-6570.1999.tb00167.x Holladay, C. L., Knight, J. L., Paige, D. L., & Quiñones, M. A. (2003). The influence of framing on attitudes toward diversity training. Human Resource Development Quarterly, 14(3), 245- 263. doi:10.1002/hrdq.1065 Holladay, C. L., & Quiñones, M. A. (2008). The influence of training focus and trainer characteristics on diversity training effectiveness. Academy of Management Learning & Education, 7(3), 343-354. doi:10.5465/AMLE.2008.34251672 Hom, P. W., Roberson, L., & Ellis, A. D. (2008). Challenging conventional wisdom about who quits: revelations from corporate America. Journal of Applied Psychology, 93(1), 1-34. doi:http://dx.doi.org/10.1037/0021-9010.93.1.1 Kalinoski, Z. T., Steele-Johnson, D., Peyton, E. J., Leas, K. A., Steinke, J., & Bowling, N. A. (2013). A meta-analytic evaluation of diversity training outcomes. Journal of Organizational Behavior, 34(8), 1076-1104. doi:10.1002/job.1839 68 Kahle, L. R. & Homer, P. M. (1985). Physical attractiveness of the celebrity endorser: A social adaptation perspective. Journal of Consumer Research, 11(4), 954-961. doi:https://doi.org/10.1086/209029 Kaplan, D. M., Wiley, J. W., & Maertz, C. P. (2011). The role of calculative attachment in the relationship between diversity climate and retention. Human Resource Management, 50(2), 271-287. doi:10.1002/hrm.20413 Kidder, D. L., Lankau, M. J., Chrobot-Mason, D., Mollica, K. A., & Friedman, R. A. (2004). Backlash toward diversity initiatives: Examining the impact of diversity program justification, personal and group outcomes. International Journal of Conflict Management, 15(1), 77-102. doi:http://dx.doi.org/10.1108/eb022908 Kim, S. S., & Gelfand, M. J. (2003). The influence of ethnic identity on perceptions of organizational recruitment. Journal of Vocational Behavior, 63(3), 396-416. doi:http://doi.org/10.1016/S0001-8791(02)00043-X Kossek, E. E., & Zonia, S. C. (1993). Assessing diversity climate: A field study of reactions to employer efforts to promote diversity. Journal of Organizational Behavior, 14(1), 61-81. doi:10.1002/job.4030140107 Kravitz, D. A., & Klineberg, S. L. (2000). Reactions to two version of affirmative action among Whites, Blacks and Hispanics. Journal of Applied Psychology, 85(4), 597-611. doi:10.1037//0021-9010.85.4.597 Kravitz, D. A., Klineberg, S. L., Avery, D. R., Nguyen, A. K., Lund, C., & Fu, E. J. (2000). Attitudes towards affirmative action: Correlations with demographic variables and with beliefs about targets, actions and economic effects. Journal of Applied Social Psychology, 30(6), 1109-1136. doi:10.1111/j.1559-1816.2000.tb02513.x Lee, C. K. C., Fernandez, N., & Martin, B. A. (2002). Using self-referencing to explain the effectiveness of ethnic minority models in advertising. International Journal of Advertising, 21(3), 367-379. doi:10.1080/02650487.2002.11104937 Lev-Ram, M. (2015, August 13). Apple’s diversity numbers reveal plenty of progress to be made. Time Magazine. Retrieved from: http://time.com/3997379/apple-female-employees/ Martin, B. A., Lee, C. K. C. & Yang, F. (2004). The influence of ad model ethnicity and self- referencing on attitudes: Evidence from New Zealand. Journal of Advertising 33(4), 27-37. doi:10.1080/00913367.2004.10639172 McCroskey, J. C., & Teven, J. J. (1999). Goodwill: A reexamination of the construct and its measurement. Communications Monographs, 66(1), 90-103. 69 McGuire, W. J. & Padawer-Singer, A. (1976). Trait salience in the spontaneous self-concept. Journal of Personality and Social Psychology, 33(6), 743-754. doi:10.1037/0022- 3514.33.6.743 McIntosh, P. (1989). White privilege: Unpacking the invisible knapsack. Peace and Freedom, pp. 10-12. McKay, P. F., Avery, D. R., Tonidandel, S., Morris, M. A., Hernandez, M., & Hebl, M. R. (2007). Racial differences in employee retention: Are diversity climate perceptions the key?. Personnel Psychology, 60(1), 35-62. doi:10.1111/j.1744-6570.2007.00064.x Montei, M. S., Adams, G. A., & Eggers, L. M. (1996). Validity of scores on the attitudes towards diversity scale (ATDS). Educational and Psychological Measurement, 56(2), 293-303. Mor Barak, M. E., Cherin, D. A., & Berkman, S. (1998). Organizational and personal dimensions in diversity climate: Ethnic and gender differences in employee perceptions. The Journal of Applied Behavioral Science, 34(1), 82-104. Newman, D. A., & Lyon, J. S. (2009). Recruitment efforts to reduce adverse impact: Targeted recruiting for personality, cognitive ability, and diversity. Journal of Applied Psychology, 94(2), 298-317. doi:http://dx.doi.org/10.1037/a0013472 Pendry, L. F., Driscoll, D. M., & Field, S. C. (2007). Diversity training: Putting theory into practice. Journal of Occupational and Organizational Psychology, 80(1), 27-50. doi:10.1348/096317906X118397 Petty, R. E. & Cacioppo, J. T. (1986). The elaboration likelihood model of persuasion. Advances in Experimental Social Psychology, 19, 123-205. doi: https://doi.org/10.1016/S0065- 2601(08)60214-2 Petty, R. E., Fleming, M. A., Priester, J. R., & Feinstein, A. H. (2001). Individual versus group interest violation: Surprise as a determinant of argument scrutiny and persuasion. Social Cognition, 19(4), 418. doi:10.1521/soco.19.4.418.20758 Pratto, F., Sidanius, J. Stallwoth, L. M., & Malle, B. F. (1994). Social dominance orientation: A personality variable predicting social and political attitudes. Journal of Personality and Social Psychology, 67(4), 741-763. Priester, J. R., & Petty, R. E. (1995). Source attributions and persuasion: Perceived honesty as a determinant of message scrutiny. Personality and Social Psychology Bulletin, 21(6), 637- 654. doi:10.1177/0146167295216010 Regan, D. T. & Fazio, R. (1977). On the consistency between attitudes and behavior: Look to the method of attitude formation. Journal of Experimental Social Psychology, 13(1), 28-45. doi:https://doi.org/10.1016/0022-1031(77)90011-7 70 Rudman, L. A., & Glick, P. (2001). Prescriptive gender stereotypes and backlash toward agentic women. Journal of Social Issues, 57(4), 743-762. Ryan, A. M. & Powers, C. (2012). Workplace diversity. In N. Schmitt (Ed.), The Oxford handbook of personnel assessment and selection. New York: Oxford University Press. p. 814-831. Ryan, A. M., Sacco, J. M., McFarland, L. A., & Kriska, S. D. (2000). Applicant self-selection: Correlates of withdrawal from a multiple hurdle process. Journal of Applied Psychology, 85, 163-179. Rynes, S., & Rosen, B. (1995). A field survey of factors affecting the adoption and perceived success of diversity training. Personnel Psychology, 48(2), 247-270. doi:10.1111/j.1744- 6570.1995.tb01756.x Sabat, I. E., Lindsey, A. P., Membere, A., Anderson, A., Ahmad, A., Kind, E., & Bolunmez, B. (2014). Invisible disabilities: Unique strategies for workplace allies. Industrial and Organizational Psychology, 7(2), 259-265. Sabat, I. E., Martinez, L. R., & Wessel, J. L. (2013). Neo-activism: Engaging allies in modern workplace discrimination reduction. Industrial and Organizational Psychology, 6(4), 480- 485. Schmit, M. J., & Ryan, A. M. (1997). Applicant withdrawal: The role of test-taking attitudes and racial differences. Personnel Psychology, 50, 855-876. Sidanius, J., & Pratto, F. (1993). The inevitability of oppression and the dynamics of social dominance. In P. M. Sniderman, P. E. Tetlock, & E. G. Carmines (Eds.), Prejudice, politics and the American dilemma. Stanford, CA: Stanford University Press. Sierra, J. J., Hyman, M. R., & Torres, I. M. (2009). Using a model’s apparent ethnicity to influence viewer responses to print ads: A social identity theory perspective. Journal of Current Issues and Research in Advertising, 31(2), 41-66. doi:10.1080/10641734.2009.10505265 Snyder, M. L., Kleck, R. E., Strenta, A., & Mentzer, S. J. (1979). Avoidance of the handicapped: An attributional ambiguity analysis. Journal of Personality and Social Psychology, 37(12), 2297-2306. Tseng, F. C., & Fan, Y. J. (2011). Exploring the influence of organizational ethical climate on knowledge management. Journal of Business Ethics, 101(2), 325-342. Washington, J., & Evans, N. J. (1991). Becoming an ally. In N. J. Evans, & V. A. Wall (Eds.) Beyond tolerance: Gays, lesbians and bisexuals on campus (pp. 195-204). Alexandria, VA: American Association for Counseling and Development. 71 Wells, W. D., Leavitt, C. & McConville, M. (1971). A reaction profile for TV commercials. Journal of Advertising Research, 11(6), 11-17. Wingfield, N. (2015, January 6). Intel allocates $300 million for workplace diversity. The New York Times. Retrieved from: https://www.nytimes.com/2015/01/07/technology/intel-budgets- 300-million-for-diversity.html Wood, W., & Eagly, A. H. (1981). Stages in the analysis of persuasive messages: The role of causal attributions and message comprehension. Journal of Personality and Social Psychology, 40(2), 249-259. doi:http://dx.doi.org/10.1037/0022-3514.40.2.246 Xu, Y. J. (2008). Gender disparity in STEM disciplines: A study of faculty attrition and turnover intentions. Research in Higher Education, 49(7), 607-624. doi:10.1007/s1162-008-9097-4 Zhao, X., & Peterson, E. (2017). Effects of temporal framing on response to antismoking messages: The mediating role of perceived relevance. Journal of Health Communication, 22(1), 37-44. doi:http://dx.doi.org/10.1080/10810730.2016.1250844 Zweigenhaft, R.L., & Domhoff, G. W. (2011) The new CEOs: Women, African American, Latino, and Asian American leaders of Fortune 500 companies. Lanham, MD: Rowman & Littlefield. 72