Sage Journals: Discover world-class research

Abstract

Anonymization of job applicant resumes is a recommended strategy to increase diversity in organizations, but large-scale tests have shown mixed results. We consider decision-makers’ social dominance orientation (SDO), a measure of anti-egalitarianism/endorsement of group-based hierarchy, to illustrate the limits of anonymization. Across four pre-registered studies (N = 3,150), we show that (a) lower SDO individuals are less likely to hire individuals from underrepresented groups when job materials are anonymized and (b) they are more likely to opt into using anonymization. Taken together, these results suggest that opt-in anonymization policies may sometimes reduce the diversity of who is selected. Furthermore, people appear to have inaccurate lay beliefs about the consequences of anonymization. Our results suggest policy evaluations of diversity interventions should consider the interaction of heterogeneous treatment effects and selection effects, which may inadvertently lead to outcomes that are contrary to the stated policy goals.

Keywords

anonymization blinding diversity social dominance orientation

Anonymization of resumes—or the removal of information (e.g., names, photos) that could potentially identify an applicant’s social category or demographic information—is one of the most commonly recommended strategies by both academics and practitioners to increase diversity in hiring for organizations (Bortz, 2018; Fath et al., 2021; Feintzeig, 2016; Goldin & Rouse, 2000; Kottasova, 2015). Previous research demonstrates that the availability of social category information influences subsequent judgments and decisions about an individual (e.g., Biernat et al., 1991; Darley & Gross, 1983), which suggests that removing demographic information could de-bias evaluation processes by removing access to demographic attributes. And yet, some large-scale efforts to implement anonymization of resumes in the real world have had mixed results (Åslund & Skans, 2012; Bøg & Kranendonk, 2011; Krause et al., 2012) or backfired (Behaghel et al., 2015). Why is this the case? In the present research, we explore the consequences of anonymization, lay people’s understanding of the practice, and how an individual’s social dominance orientation (SDO; Ho et al., 2015; Sidanius & Pratto, 1999)—an ideological variable that captures support for group-based hierarchies—influences anonymization outcomes and decisions to use anonymization at all. By considering how SDO interacts with anonymization, we hope to shed light on why anonymization may not be a panacea for increasing diversity in organizations.

Anonymization/Deidentification/“Blinding”

Decades of correspondence or audit studies have revealed discrimination in employment contexts on the basis of social category or demographic information such as race, ethnicity, and gender (Baert, 2018; Bertrand & Duflo, 2017; Bertrand & Mullainathan, 2004; Quillian et al., 2017). In these field experiments, fictional job materials are generated and sent out to real job openings, differing only in the particular characteristic (e.g., race) being studied. Unequal callback rates from employers constitute discrimination since the qualifications of candidates are identical, except for the social identity characteristic being varied. Other work on hiring discrimination reveals that individuals will justify decisions biased by social category information by re-weighting the relative value of certain qualifications to mirror those of their preferred candidate (Norton et al., 2004). If employers believe that they should not be selecting job candidates on the basis of social category or demographic information, one solution is to implement anonymization (also referred to as de-identification or “blinding”) to remove the biasing information so that discrimination becomes impossible (Goldin & Rouse, 2000; Onyeador et al., 2021). Anonymization may be seen as an instantiation of a colorblind approach to diversity, in contrast to a multicultural approach to diversity (Correll et al., 2008; Kung et al., 2023; Plaut et al., 2018; Rattan & Ambady, 2013).

In a landmark study, Goldin and Rouse (2000) examined the impact of an anonymized evaluation procedure, where orchestra musicians auditioned behind a curtain as opposed to out in the open. They found that this anonymized evaluation process led to increased representation of women in the five different orchestras, from 6% in 1970 to 21% in 1993. Due to the results of this study and the straightforward logic of the practice, anonymization has been recommended widely, with firms such as HSBC and Deloitte implementing anonymized evaluation procedures (Feintzeig, 2016; Kottasova, 2015).

Although companies have started implementing anonymized procedures, a number of field experiments in European countries examining the impact of anonymization on hiring have found mixed results as to whether anonymization improves job outcomes for women and ethnic/racial minorities (Åslund & Skans, 2012; Behaghel et al., 2015; Bøg & Kranendonk, 2011; Krause et al., 2012). In an experimental program in France, firms could opt into receiving anonymized resumes or not from a public employment service (Behaghel et al., 2015). Anonymization decreased the diversity of selected candidates: It increased the gap in interviewing rates between majority and minority candidates (e.g., residents of deprived neighborhoods, immigrants), and as a result, the hiring gap also widened. These mixed results suggest that we need a more nuanced understanding of anonymization before implementing it in organizations or suggesting it as a broad policy prescription.

Assumptions Underlying Anonymization as a Diversification Strategy

In some ways, the effectiveness of anonymization is self-evident: by removing cues of social identity, anonymization prevents their use in selection or evaluation processes. However, reducing discrimination on the basis of identity and increasing demographic diversity are not the same. In order for anonymization to increase demographic diversity, several key assumptions must be true. First, if we believe that people have unequal access to education and opportunities based on their social identities (e.g., because systemic and institutionalized racism exists; Rucker & Richeson, 2021), then anonymization will reify existing societal biases, given that the people from groups who are disadvantaged or marginalized in society will not—on average—have the same qualifications as those from advantaged or privileged groups. By removing social identity from evaluation processes, we remove our ability to account for systemic and institutionalized advantages and disadvantages. Furthermore, given two equivalently qualified candidates—one from a historically advantaged background and one from a historically disadvantaged background—it can be rational for a decision maker to preferentially select the candidate from the historically disadvantaged background, given that it is likely that that person had to overcome more to achieve the same qualifications as the candidate from the historically advantaged background (Bohren et al., 2019). Anonymization would prevent this rational behavior from occurring, as these candidates would be equally likely to be selected.

Second, anonymization assumes that people and organizations do not have valid reasons for actively pursuing candidates from historically marginalized backgrounds to increase team or organizational diversity. For some roles, candidates with historically marginalized identities might be perceived as possessing greater expertise than their less historically marginalized counterparts, which in turn increases hireability (Torrez et al., 2024). Other work has documented many reasons why diversity may provide benefits to teams or organizations (Herring, 2009; Hoogendoorn et al., 2013; McLeod et al., 1996; Proudfoot et al., 2024; Richard et al., 2007; Sinaceur et al., 2010; Sommers, 2006; Wright et al., 1995). Even in the absence of systemic or institutionalized disadvantage, anonymization presumes that organizations do not have any active reason to pursue diversity because it prevents selection based on social identity.

Finally, the practice of anonymization is based on the assumption that all identifying details can be completely removed from application materials. Truly anonymizing job materials is much harder than just removing a name at the top of the resume. Often, extracurricular or leadership activities (e.g., Women’s Rugby Football Club President) can “leak” social identity cues (Rivera & Tilcsik, 2016). While these cues could be removed, doing so may remove valuable information about a candidate’s values, qualifications, or interests. Furthermore, reading a candidate’s resume through the lens of their identity can provide important context for their experiences and achievements. Anonymization, even if well-intentioned, may sometimes hinder an evaluator’s ability to gain a nuanced understanding of the candidate’s background, challenges they may have overcome, and the full scope of their potential contributions to an organization.

Given that there are reasons to expect these assumptions to not be true, anonymization may not always be an effective intervention to increase diversity in organizations. Thus, although anonymization may eliminate the direct influence of social identity on selection decisions, which should in theory reduce identity-based discrimination, this may sometimes reduce the likelihood of people selecting historically marginalized candidates who would increase the demographic diversity of a team or organization when these assumptions are violated (or when decision makers believe these assumptions to not be true).

(Anti-)Egalitarianism and Social Dominance Orientation

One potentially important factor in determining the consequences of anonymization may be evaluators’ SDO. SDO is an ideological variable that measures the extent to which people prefer and endorse group-based inequality (Ho et al., 2015; Sidanius & Pratto, 1999). Individuals higher in SDO (i.e., anti-egalitarians¹) aim to maintain hierarchy between groups in society, while individuals lower in SDO (i.e., egalitarians) aim to attenuate hierarchy between groups in favor of intergroup equality. Previous work reveals that SDO is essential in understanding the composition and culture of organizations: Organizations and potential employees both tend to select one another to have compatible orientations for hierarchy enhancement or attenuation (Danbold & Unzueta, 2020; Dupree & Torrez, 2021; Haley & Sidanius, 2005; Pratto & Espinoza, 2001; Pratto et al., 1994; Sidanius et al., 1996). Ironically, recent work has also found that when hierarchy-enhancing myths such as gender stereotypes and essentialism support organizational diversity, egalitarians increasingly endorsed these beliefs (de Leon & Kay, 2021).

We propose that SDO can help us understand the impact of anonymization on selecting historically marginalized candidates. People lower in SDO may be more likely to believe systemic or institutionalized discrimination exists (Kteily et al., 2017), which may lead them to take steps to counteract this inequity. Furthermore, they may also be more likely to believe that diversity has benefits for organizations (Cokley et al., 2010). As such, when lower SDO decision-makers have access to job candidates’ social identity cues, they may exhibit a preference for selecting equivalently qualified candidates from historically marginalized groups. On the other hand, if anonymization is implemented, this preference cannot be acted upon, thereby decreasing the selection of historically marginalized candidates by lower SDO decision-makers.

In contrast, individuals with higher SDO may be more likely to discriminate against candidates with historically marginalized identities when such identity information is available (Kteily et al., 2011). Higher SDO individuals make decisions that are perceived to uphold the hierarchical status quo (Sidanius et al., 1992). When candidate information is not anonymized, higher SDO decision-makers may use social identity cues to make selections that maintain the status quo, choosing to select individuals from dominant groups. However, when anonymization is implemented, higher SDO individuals lose the ability to directly act on these cues, potentially resulting in more frequent selection of historically marginalized candidates.

Taken together, SDO may help us understand when anonymization is likely to lead to increases or decreases in the selection of historically marginalized candidates, due to its varied impact on higher and lower SDO decision-makers. Specifically, we predict that the effect of anonymization on rates of selecting marginalized candidates will be moderated by SDO such that anonymization (compared with no anonymization) will decrease rates of hiring the marginalized candidate for lower SDO individuals and increase rates of hiring the marginalized candidate for higher SDO individuals (Hypothesis 1).

Lay Perceptions of Anonymization and Who Is Likely to Opt in

Beyond helping us understand the consequences of anonymization, SDO may also help us understand preferences for anonymization as an organizational strategy. In general, we predict that people do not understand the nuances of anonymization and when it is likely to increase selected diversity or decrease it. Because anonymization is at face value a fair process and because it is often cited as a method to decrease the influence of bias and stereotypes in hiring processes (Gawronski et al., 2020; Onyeador et al., 2021), people may have lay beliefs that anonymization will generally increase the diversity of who is selected. However, as we have argued, this may not always be true. As such, we predict that people will have inaccurate lay beliefs about the consequences of anonymization for the selection rates of people from historically underrepresented groups (Hypothesis 2).

Given these inaccurate lay beliefs, we propose that anonymization should be particularly appealing to people who care about fairness and equity, which are people lower in SDO. Anonymization may seem like a natural way to attenuate hierarchy in organizations, given that people from historically marginalized identities are often targets of stereotypes and discrimination. Indeed, past research has shown that SDO can help explain people’s reactions to organizational policies like affirmative action, with people lower in SDO supporting these policies when they are understood to disrupt the status quo, whereas higher SDO individuals generally do not support affirmative action policies unless they are framed to uphold the hierarchical status quo (e.g., because it would entail only the recruitment of underrepresented groups that would remain at the bottom of organizational hierarchies; Aberson, 2016; Chow et al., 2013; Federico & Sidanius, 2002; Greig et al., 2023; Gutiérrez & Unzueta, 2013; Harrison et al., 2006; G. C. Ho & Unzueta, 2015; Kravitz et al., 2008; Kravitz & Klineberg, 2000; Unzueta et al., 2012; Waldfogel et al., 2021). Despite having opposite instantiations (as affirmative action actively takes into account identity, whereas anonymization attempts to prevent this from happening), because both anonymization and affirmative action have been touted as ways to promote more diverse outcomes (Crosby et al., 2006), people’s reactions to these policies may be similar. In other words, lower SDO individuals may be the ones most likely to adopt anonymization procedures in their hiring processes (Hypothesis 3) because they perceive it as a practice that will promote their egalitarian ideals.

Combined with our previous prediction, this suggests that the people who might most want to adopt anonymized selection processes (i.e., those lower in SDO) are also those for whom anonymization will lead to decreases in selected diversity, which may help explain the lackluster results of field implementations of anonymization in the real world. Furthermore, we suggest that this effect is driven by people’s misperceptions of the consequences of anonymization. We predict that if we frame anonymization as a practice that does not promote egalitarian ideals (i.e., because it can decrease organizational diversity), then lower SDO people will be less likely to opt into anonymized hiring practices than higher SDO people (Hypothesis 4).

The Present Research

We first consider how SDO helps predict how anonymization consequently affects hiring decisions. Next, we examine whether people have inaccurate lay beliefs about the outcome of anonymized versus identifiable hiring processes (and the direction of the inaccuracy). Finally, we examine how SDO influences whether people opt into using anonymized hiring practices and how differences in the framing of projected outcomes influences these choices.

In Study 1, we examine the impact of anonymization in a hypothetical hiring decision between a (White) woman and (White) man with similar qualifications and test whether SDO moderates the outcome. We predict that the effect of anonymization on rates of selecting the female candidate will be moderated by SDO such that anonymization (compared with no anonymization) will decrease rates of hiring the female candidate for lower SDO individuals and increase rates of hiring the female candidate for higher SDO individuals (Hypothesis 1). In Study 2, we ask people to predict the outcomes of Study 1 to understand whether lay beliefs about anonymized hiring processes are accurate. We predict that people have inaccurate lay beliefs about the impact of anonymized hiring processes such that they will predict that a woman will be hired at a lower rate than the actual rate we found in Study 1 (Hypothesis 2). In Study 3, we examine whether an individual’s SDO influences the likelihood of adopting an anonymized hiring process. We propose that when hiring managers or decision makers are lower in SDO, they will be more likely than those higher in SDO to adopt anonymization as a hiring practice because they believe that anonymization promotes fairness, which resonates with their egalitarian values (Hypothesis 3). In Study 4, we examine how the stated consequence of anonymization influences the likelihood of adoption and whether this varies based on an individual’s SDO. We believe that we will replicate our findings from Study 3 when anonymization is explicitly framed as a process that will increase organizational diversity, but in contrast, when anonymization is framed as a process that will decrease organizational diversity, we predict that social egalitarians will be less likely than anti-egalitarians to adopt anonymization as a hiring practice (Hypothesis 4). These four studies examine the consequences of anonymization, whether people have accurate lay beliefs about the effects of an anonymized hiring process, and how the perceived outcome of anonymization (i.e., increasing or decreasing organizational diversity) and a decision-maker’s SDO interact to determine the likelihood of adopting anonymization as part of a hiring process. To be clear, we are not claiming to tread new ground about our understanding of SDO as it relates to organizational policies that support or challenge the status quo. Instead, we see our work as combining straightforward predictions from the SDO literature in a nonobvious way to contribute to our understanding of a commonly recommended and implemented diversity, equity, and inclusion (DEI) strategy. We propose that anonymization may not work as expected in the real world because the organizations that are likely to opt into anonymization are those where it may lead to decreases in selected diversity.

Study 1: How Does Evaluator SDO Influence Anonymized and Non-Anonymized Selection Decisions Between a White Woman and White Man?

In Study 1, we examine whether SDO moderates hiring outcomes when using anonymization in hypothetical hiring decisions between a (White) woman and a (White) man with similar qualifications.

Transparency and Openness

For all studies, we report all measures, manipulations, and exclusions. Materials, data, and analyses for all studies, as well as Supplemental Appendix containing our analysis strategy, robustness checks, study tables, and supplemental exploratory analyses, can be found here: https://osf.io/syfqu/.

Methods

Participants

We recruited 1,200 U.S.-based participants² across CloudResearch and Prolific (CloudResearch = 175,³ Prolific = 1,025; 527 women, 667 men, six identify with a gender not listed; M_age = 37.58 years, SD = 13.22; 855 self-identified as mono-racial White; see pre-registrations for more details about our sampling decisions: https://aspredicted.org/gr3v-6xp3.pdf, https://aspredicted.org/2rd9-3dyj.pdf, https://aspredicted.org/pryw-m35g.pdf). In our final sample, 48.2% identified as liberal (ranging from very to slightly), 45.5% identified as conservative (ranging from very to slightly), and 6.3% identified as neither liberal nor conservative. 63.5% of participants reported having some experience making hiring decisions.

We conducted a Monte Carlo power analysis with 1,000 replications to determine the power for α = .05 with our sample size. Results indicated that power was sufficient for a small interaction effect (−0.052).

Social Dominance Orientation

We used the shortened eight-item SDO₇ scale (Ho et al., 2015; for example, “An ideal society requires some groups to be on top and others to be on the bottom”) to measure social dominance on a 1 (strongly oppose) to 7 (strongly favor) scale. We averaged responses to create a single index, α = .90, M = 2.66, SD = 1.37, where higher numbers indicated a higher preference and support for group-based inequality.

Procedure

We asked participants to imagine that they were hiring managers for a technology company and that they were tasked with hiring a software engineer. Participants were randomly assigned to one of two conditions: gender information or anonymized.

In both conditions, participants evaluated the resumes and interviewer notes of the same two candidates. Resumes and interviewer notes were randomly assigned to candidates (i.e., either Laurie Andersen and Hunter McGrath in the gender information condition or “Candidate 1” and “Candidate A”⁴ in the anonymized condition). By randomizing job materials to names, we ensured that the two candidates had equal qualifications on average. In the gender information condition, the names Laurie Andersen and Hunter McGrath appeared on the job materials of the two candidates, so that participants could infer candidate gender based on their names. We chose the names Laurie Andersen and Hunter McGrath to match names on perceptions of Whiteness and associated SES⁵ (Gaddis, 2017). In the anonymized condition, participants evaluated the same job materials of two candidates, but we removed the names and replaced them with the uninformative labels of “Candidate 1” and “Candidate A,” which meant that gender could not be inferred in this condition. After reading resumes and interviewer notes of both candidates, participants were asked which candidate they would prefer to hire and made a hiring decision. Finally, all participants filled out the shortened SDO₇ scale and answered questions about their demographic information.

Results and Discussion

Following our pre-registered analysis plan, we fit linear probability models with robust standard errors to predict the probability of hiring the female candidate (or Candidate A; coded as 1) versus the male candidate (or Candidate 1; coded as 0) from condition (anonymized, gender information), participant SDO, and their interaction.⁶ See Figure 1 for a depiction of results.

Figure 1.

Study 1 Regression-Estimated Likelihood of Choosing the Female Candidate Versus Male Candidate as Predicted by Condition and Participant SDO

Consistent with Hypothesis 1, we found a significant interaction between condition and participant SDO, b_Interaction = −0.043, 95% confidence interval (CI) = [−0.083, −0.003], t(1196) = −2.10, p = .036, f² = 0.004, meaning that participant SDO moderated the effects of anonymization on the selection of the female candidate. Compared with the anonymized condition, participants were more likely to hire the female candidate in the gender information condition, b_GenderInformation = 0.226, 95% CI = [0.106, 0.346], t(1196) = 3.74, p < .001, f² = 0.01. As expected, in the anonymized condition, we did not find significant effects of participant SDO on the probability of hiring the “female” candidate (i.e., Candidate A; this remains true when Candidate 1 is coded as the “female” candidate instead), b_SDO = −0.005, 95% CI = [−0.034, 0.023], t(1196) = −0.37, p = .714. In other words, anonymization did work as “intended” given profiles were designed to be equivalent in qualifications. On the other hand, in the gender information condition, participants with higher SDO were less likely to hire the female candidate, b_SDO = −0.048, 95% CI = [−0.077, −0.020], t(1196) = −3.37, p < .001.

In an exploratory floodlight analysis, we unpacked the interaction from our linear probability model using the Johnson–Neyman technique to identify SDO values for which the relationship between condition and choice of the female candidate was significant (at alpha = .05).⁷ Consistent with our predictions, we found that the interaction was significant between the SDO values of 1 and 3.48,⁸ demonstrating that lower SDO participants were significantly more likely to select the female candidate when they had access to gender information compared with anonymized materials. To our surprise, for participants with SDO higher than 3.48, this analysis suggests that there was no impact of condition on likelihood of selecting the female candidate. In other words, lower SDO participants were more likely to hire the (White) female candidate over the (White) male candidate when job materials were not anonymized, but we did not find that higher SDO participants were more likely to hire the (White) female candidate when job materials were anonymized. However, even considering only the effect among lower SDO participants, this study illustrates how implementing anonymized hiring practices can decrease the selection of historically marginalized candidates. In the Supplemental Appendix, we also report results of a study that conceptually replicates these effects when considering candidate race instead of gender in another hiring context (see Supplemental Study 1).

Study 2: Are Lay Beliefs About the Effect of Anonymized Hiring Processes Accurate?

In Study 2, we ask people to predict the outcome of Study 1 to understand whether lay beliefs about the outcomes of anonymized hiring processes versus typical hiring processes are accurate. We expect participants to predict that a woman will be hired at a lower rate in an identifiable hiring process than the actual rate we found in Study 1. We also predict that participants will be directionally inaccurate in their predictions, expecting anonymization to increase rates of hiring of the woman as opposed to decreasing the rate (as we found in Study 1). If people have inaccurate lay beliefs about the effects of anonymization, this could help explain why people opt in to anonymization even if it runs counter to their diversity preferences.

Methods

Participants

We recruited 150⁹ U.S.-based participants from Amazon Mechanical Turk (63 women, 84 men, 3 non-binary; M_age = 39.30 years, SD = 10.64; 110 self-identified as mono-racial White). This study was pre-registered on AsPredicted.org: https://aspredicted.org/ntf3-gcvr.pdf.

Procedure

We told participants that they would see materials from another study and that they would make a prediction about the results of this other study. We presented participants with instructions identical to Study 1 and explained that previous participants were assigned to one of two scenarios: identifiable candidates (i.e., gender information condition) or anonymized candidates (i.e., anonymized condition). We referred to conditions as scenarios to decrease technical jargon for increased participant understanding. However, for increased clarity here, we will use Study 1 condition names. First, we showed participants the job materials (i.e., resumes and interviewer notes) of Laurie Andersen and Hunter McGrath in the gender information condition. Next, we told them that participants in the anonymized condition evaluated the exact same materials, but the names were replaced with either “Candidate A” or “Candidate 1,” such that there was no identifying information in the job materials. Participants were told that in the anonymized condition, Laurie Andersen was selected by 54% of the participants. We then asked them to predict the percent of participants in the gender information condition that chose to hire Laurie Andersen. We incentivized their predictions by offering them an opportunity to earn a $0.10 bonus based on the accuracy of their prediction.¹⁰ Finally, all participants answered questions about their demographic information.

Results and Discussion

Following our pre-registered analysis plan, we ran a one-sample t-test to compare participants’ predictions about the percent of participants that chose to hire a woman in the gender information condition compared with the actual percentage of participants in Study 1 that chose to hire a woman in that condition (i.e., 65.2%). Consistent with Hypothesis 2, we found that participants significantly underestimated the rate at which participants chose to hire a woman (M = 46.4%, SD = 14.9%) in the gender information condition from Study 1, t(149) = −15.55, p < .001.

To explore the directional accuracy of predictions compared with the provided percentage of participants that chose to hire a woman in the anonymized condition (i.e., 54%), we coded participant responses as directionally accurate (i.e., they predicted that more than 54% of participants chose a woman in the gender information condition) or inaccurate (i.e., they predicted that 54% or fewer participants chose a woman in the gender information condition), as specified in our pre-registration. We ran a one-sample t-test to compare this binary outcome (26% of participants predicted directionally accurate results) to 50% and found that it did not significantly exceed 50%, t(149) = −6.68, p < .001, suggesting significant directional inaccuracy.

Taken together, incentivized participants were unable to predict the results of Study 1 and made directionally inaccurate predictions about how anonymization influences the selection of historically marginalized candidates. Specifically, people predicted that a woman was less likely to be hired when job materials contain identifying information (e. g., names) compared with when job materials are anonymized, which provides some evidence of inaccurate lay beliefs about the outcome differences between identifiable and anonymized hiring processes. Importantly, in combination with Study 1, we show that there is a disconnect between how people think anonymization operates and its actual consequences. This misperception of the consequences of anonymization may lead people who ultimately favor increased diversity in organizations to self-select into anonymized hiring processes, which may run counter to their goals.

Study 3: Who Opts Into Using Anonymized Hiring Processes?

Given that people generally seem to think that anonymization will increase the hiring of historically marginalized candidates, in Study 3, we test whether SDO predicts who is most likely to opt into anonymized hiring processes. As people lower in SDO are more likely to want to increase the hiring of historically marginalized candidates, we expect participants lower in SDO to be more likely to opt into anonymized hiring processes.

Methods

Participants

We recruited 600 U.S.-based participants¹¹ from Prolific who self-reported having hiring experience via the Prolific prescreen (291 women, 299 men, 10 identified with a gender not listed; M_age = 39.88 years, SD = 12.55; 485 self-identified as mono-racial White). This study was pre-registered on AsPredicted.org: https://aspredicted.org/vgqm-7wng.pdf.

Social Dominance Orientation

We again used the shortened eight-item SDO₇ scale (Ho et al., 2015) which was reliable, α = .91, M = 2.25, SD = 1.28.

Procedure

We told participants that they would learn about two different hiring processes that a company was considering and needed to decide which to use. The anonymized and non-anonymized hiring processes were randomly assigned a label (either Process A or Process B) and presented in counterbalanced order accordingly. The processes were described identically, with one crucial difference: in the anonymized hiring process, applicant resumes were de-identified prior to evaluation. Participants were also explicitly told the difference between the two processes: “There is only one difference between Process A and Process B: In Process A, names and other identifying information will be removed from resumes prior to evaluation.”

After reading about the two processes, all participants chose one of the two hiring processes, filled out the shortened SDO₇ scale, and answered questions about their demographic information.

Study 3 Results and Discussion

Following our pre-registered analysis plan, we fit a linear probability model with robust standard errors to predict the probability of choosing the anonymized hiring process versus the typical process without anonymization from participant SDO.¹² Consistent with Hypothesis 3, we found a main effect of SDO such that participants with higher SDO were less likely to choose anonymization, b_SDO = −0.096, 95% CI = [−0.125, −0.068], t(598) = −6.56, p < .001, f² = 0.07. In other words, lower SDO participants were more likely to choose the anonymized hiring process compared with higher SDO participants. See Figure 2 for a depiction of results.

Figure 2.

Study 3 Regression-Estimated Likelihood of Choosing the Anonymized Hiring Process Versus Non-Anonymized Hiring Process as Predicted by Participant SDO

Study 4: Does the Stated Consequence of Anonymization Moderate Who Opts Into Using a Hiring Process That Includes Anonymization?

In Study 4, we wanted to further unpack the impact of SDO on the adoption of anonymized hiring processes. We investigate whether framing anonymization as a procedure that increases or decreases organizational diversity moderates the effect of an individual’s SDO on who is likely to opt into a hiring process that includes anonymization.

Methods

Participants

We recruited 1,200 U.S.-based participants¹³ from Prolific who self-reported having hiring experience via the Prolific prescreen (527 women, 665 men, eight identified with a gender not listed; M_age = 45.94 years, SD = 13.12; 791 self-identified as mono-racial White). This study was pre-registered on AsPredicted.org: https://aspredicted.org/nkwd-xqgm.pdf.

Social Dominance Orientation

We again used the shortened eight-item SDO₇ scale (Ho et al., 2015) which was reliable, α = .92, M = 2.56, SD = 1.46.

Procedure

We asked participants to imagine that they were a human resources manager and that they needed to make decisions about the details of the company’s hiring process. Participants were randomly assigned to one of two conditions: decrease diversity or increase diversity.

In both conditions, they were told that a colleague had suggested they read an article. In the decrease diversity condition, they were given an article titled: “Harvard researchers: Removing names and identifying information from resumes can decrease diversity in organizations.” In the increase diversity condition, they were given an article titled: “Harvard researchers: Removing names and identifying information from resumes can increase diversity in organizations.” The articles described what anonymization entails and the rationale behind the consequence of decreasing/increasing organizational diversity. In each condition, participants answered three comprehension check questions¹⁴ about the article they read and were not able to proceed with the experiment until they answered all of them correctly.

After answering the comprehension check questions, participants were asked which hiring process they would choose to implement as the human resources manager of a new company: a hiring process with anonymization or a hiring process without anonymization. Next, they filled out the shortened SDO₇ scale and answered questions about their demographic information. Finally, participants were debriefed about the study.

Results and Discussion

Following our pre-registered analysis plan, we fit three linear probability models¹⁵ with robust standard errors to predict the probability of choosing the hiring process with anonymization versus the hiring process without anonymization. Results from the first two models were consistent with the third model, where we predicted the probability of choosing the hiring process with anonymization from condition (increase diversity, decrease diversity), participant SDO, and their interaction.¹⁶ See Figure 3 for a depiction of results.

Figure 3.

Study 4 Regression-Estimated Likelihood of Choosing the Anonymized Hiring Process Versus Non-Anonymized Hiring Process as Predicted by Condition and Participant SDO

Consistent with Hypothesis 4, we found a significant interaction between condition and participant SDO, b_Interaction = −0.147, 95% CI = [−0.180, −0.114], t(1196) = −8.21, p < .001, f² = 0.06, meaning that condition significantly moderated the effect of participant SDO on the selection of a hiring process that included anonymization over the selection of a hiring process without anonymization. Compared with the decrease diversity condition, participants were more likely to choose the hiring process with anonymization in the increase diversity condition, b_IncreaseDiversity = 0.806, 95% CI = [0.709, 0.904], t(1196) = 16.89, p < .001, f² = 0.26. When anonymization was described as decreasing organizational diversity, participant SDO had a positive relationship with the likelihood of choosing an anonymized hiring process, b_SDO = 0.069, 95% CI = [0.044, 0.095], t(1196) = 5.29, p < .001. In other words, lower SDO people were more likely to reject anonymization when we told them that it could decrease organizational diversity. On the other hand, when anonymization was described as increasing organizational diversity, participant SDO had a negative relationship with the likelihood of choosing an anonymized hiring process, b_SDO = −0.078, 95% CI = [−0.102, −0.054], t(1196) = −6.37, p < .001. In other words, similar to our results in Study 3, when participants believed that anonymization would increase organizational diversity, lower SDO people were more likely to adopt it as a practice.

In an exploratory floodlight analysis, we unpacked the interaction from our linear probability model using the Johnson–Neyman technique to identify SDO values for which the relationship between condition and choice of the hiring process with anonymization was significant (at α = .05). We found that the interaction was significant between the SDO values of 1 to 4.84 and 6.52 to 7, demonstrating that lower SDO participants were significantly more likely to select the anonymized hiring process when it was framed as increasing organizational diversity compared with when it was framed as decreasing organizational diversity. In contrast, participants highest in SDO were significantly more likely to select the anonymized hiring process when it was framed as decreasing organizational diversity compared with when it was framed as increasing organizational diversity.

These results are consistent with the broader SDO literature (Federico & Sidanius, 2002; Gutiérrez & Unzueta, 2013; G. C. Ho & Unzueta, 2015). When anonymization is framed as attenuating hierarchy (i.e., increasing organizational diversity), lower SDO individuals are more supportive than higher SDO individuals; when anonymization is framed as maintaining hierarchy (i.e., decreasing organizational diversity), higher SDO individuals are more supportive than lower SDO individuals.

General Discussion

Across four studies, we identified the impact of anonymization on hiring outcomes, whether decision-makers have accurate lay beliefs about these hiring outcomes, and how SDO and the stated consequence of anonymization moderates the likelihood of adopting an anonymized hiring process. First, we showed that SDO moderates the effect of anonymization on hiring individuals from underrepresented groups: people lower in SDO are more likely to hire a woman when job materials contain identifying information (e.g., names) compared with anonymized job materials. Next, we found that people have inaccurate lay beliefs about the results of anonymized hiring practices: People predict that a woman is less likely to be hired when job materials contain identifying information compared with when job materials are anonymized. Finally, we showed that lower SDO individuals are more likely to opt into using an anonymized hiring process compared with higher SDO individuals, but that this effect reverses when the stated consequence of anonymization is decreased organizational diversity. Taken together, we find that people have a faulty understanding of anonymization and its consequences and that people who appear to like anonymization the most (i.e., people lower in SDO) may also be those for whom anonymization may be counterproductive to their goals (because anonymization may not help increase diversity).

Given that people appear to have incorrect lay beliefs about the consequences of anonymization, organizations interested in increasing diversity may adopt anonymization thinking that it will be a silver bullet in helping this cause. However, organizations interested in diversifying should consider whether using anonymization is likely to increase or decrease the diversity of who is selected, and an analysis of whether they see evidence of discrimination against people from historically marginalized groups at baseline could help. For example, previous work evaluating a voluntary anonymization program in France revealed differences between firms that opted into the program and those that declined (Behaghel et al., 2015): Firms that declined to participate in anonymization generally interviewed and hired significantly fewer minorities compared with firms that chose to participate. Given this potential baseline discrimination, it may have been fruitful for these firms to adopt anonymization practices. But for firms without evidence of baseline discrimination, anonymization may not be the best solution. Instead, these firms could focus attention on initiatives that increase the diversity of the applicant pool prior to any evaluation.

More broadly, we contribute to our understanding of DEI initiative effectiveness. Much of this literature focuses on why DEI initiatives fail due to issues like manager or individual backlash (e.g., Dobbin et al., 2015; Dover et al., 2016; Leslie, 2019; Plaut et al., 2011). We propose that organizations need to diagnose their specific organizational barriers and consider whether the solutions they implement are appropriate and effective (Onyeador et al., 2021; Stephens et al., 2020). Our paper helps us better understand how to use choice architecture—or changes to decision processes or structures—to improve diversity in organizations (Chang & Cikara, 2018; Chang et al., 2020; He et al., 2021).

Choice architecture can be an appealing strategy to increase diversity because it typically does not require changing decision-makers’ biases or beliefs, which research has shown can be challenging to change over the long term (Chang et al., 2019; Lai et al., 2016). However, our results illustrate the need to consider (a) how choice architecture interventions are predicted to work, as understanding the mechanisms can illuminate circumstances where these interventions will backfire; (b) who selects into these interventions, as this can help illuminate how these interventions will play out in the real world; and (c) the interaction of these factors. Without carefully considering self-selection and how it interacts with the mechanism of the initiative, organizations may unwittingly be adopting initiatives that will end in lowering diversity as opposed to increasing it (as has empirically been true in many of the field tests of anonymization). We consider the combination of misperceptions, lay beliefs, and selection into a policy to further understanding barriers to diversification in organizations, which places our work in conversation with past work on misperceptions related to DEI (Kraus et al., 2022; Onyeador et al., 2021; Torrez et al., 2024) and the ironic effects of organizational diversity initiatives (Dover et al., 2020; Germano et al., 2021; Kaiser et al., 2013).

A limitation of our work is that we rely on hypothetical scenarios that focus on specific industries (e.g., software), and our studies were conducted with online participants from the United States. In addition, social desirability in our studies could be a concern, leading to elevated rates of choosing the female candidate (Study 1) and Black male candidate (Supplemental Study 1). Future work should examine other intersections of social identities (e.g., Black women), including, but not limited to, gender, race, and sexual orientation and whether the relationships we document here also play out in the field in other settings and industries. Future work could also examine the possibility of countervailing effects of lower and higher SDO decision-makers, especially in contexts where both are involved in joint decision-making. Finally, while we focus on SDO in this paper, other ideological beliefs such as belief in meritocracy (McCoy & Major, 2007), belief in a just world (Lerner & Miller, 1978), or political conservatism (Jost et al., 2003) could also be important in uncovering differences in the adoption of anonymization in a hiring process and the subsequent selection outcomes.

The appeal of anonymization is associated with the allure of pure meritocracy—that the only thing that should matter are the qualifications of a candidate. However, its function as an intervention to increase diversity rests on assumptions that discrimination in society does not exist and that organizations do not have reasons to pursue diversity. As such, introducing anonymization procedures can sometimes be counterproductive for increasing diversity. And yet, people appear to have the opposite lay beliefs, driven by perceptions that anonymization will increase diversity. In fact, if the goal is to increase diversity in the workplace, anonymization can ultimately undermine this outcome for the very people who are most motivated to address the underrepresentation of women and racial minorities in organizations.

Supplemental Material

sj-docx-1-psp-10.1177_01461672241304593 – Supplemental material for On the Limits of Anonymization for Promoting Diversity in Organizations

Supplemental material, sj-docx-1-psp-10.1177_01461672241304593 for On the Limits of Anonymization for Promoting Diversity in Organizations by Linda W. Chang and Edward H. Chang in Personality and Social Psychology Bulletin

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iDs

Linda W. Chang

Edward H. Chang

Supplemental Material

Supplemental material is available online with this article.

Notes

References

Aberson

C. L.

(2016). Policy type and justification influences on support for affirmative action policies. Journal of Personnel Psychology, 15(2), 90–93.

Åslund

Skans

O. N.

(2012). Do anonymous job application procedures level the playing field? ILR Review, 65(1), 82–107. https://doi.org/10.1177/001979391206500105

Baert

(2018). Hiring discrimination: An overview of (almost) all correspondence experiments since 2005. Springer.

Behaghel

Crépon

Le Barbanchon

(2015). Unintended effects of anonymous resumes. American Economic Journal: Applied Economics, 7(3), 1–27. https://doi.org/10.1257/app.20140185

Bertrand

Duflo

(2017). Field experiments on discrimination. In Banerjee

A.V.

Duflo

(Eds.), Handbook of field experiments (Vol. 1, pp. 309–393). North Holland Publishing.

Bertrand

Mullainathan

(2004). Are Emily and Greg more employable than Lakisha and Jamal? A field experiment on labor market discrimination. American Economic Review, 94(4), 991–1013.

Biernat

Manis

Nelson

T. E.

(1991). Stereotypes and standards of judgment. Journal of Personality and Social Psychology, 60(4), 485–499. https://doi.org/10.1037/0022-3514.60.4.485

Bøg

Kranendonk

(2011). Labor market discrimination of minorities? Yes, but not in job offers (MPRA Paper No. 33332). https://mpra.ub.uni-muenchen.de/33332/#:~:text=We%20compare%20decisions%20of%20recruiters,in%20the%20job%20offer%20decision

Bohren

J. A.

Imas

Rosenberg

(2019). The dynamics of discrimination: Theory and evidence. American Economic Review, 109(10), 3395–3436.

10.

Bortz

(2018, March 20). Can blind hiring improve workplace diversity? Society for Human Resource Management. https://www.shrm.org/hr-today/news/hr-magazine/0418/pages/can-blind-hiring-improve-workplace-diversity.aspx

11.

Chang

E. H.

Kirgios

E. L.

Rai

Milkman

K. L.

(2020). The isolated choice effect and its implications for gender diversity in organizations. Management Science, 66(6), 2752–2761. https://doi.org/10.1287/mnsc.2019.3533

12.

Chang

E. H.

Milkman

K. L.

Gromet

D. M.

Rebele

R. W.

Massey

Duckworth

A. L.

Grant

A. M.

(2019). The mixed effects of online diversity training. Proceedings of the National Academy of Sciences, 116(16), 7778–7783. https://doi.org/10.1073/pnas.1816076116

13.

Chang

L. W.

Cikara

(2018). Social decoys: Leveraging choice architecture to alter social preferences. Journal of Personality and Social Psychology, 115(2), 206–223.

14.

Chow

R. M.

Lowery

B. S.

Hogan

C. M.

(2013). Appeasement: Whites’ strategic support for affirmative action. Personality and Social Psychology Bulletin, 39(3), 332–345.

15.

Cokley

K. O.

Tran

Hall-Clark

Chapman

Bessa

Finley

Martinez

(2010). Predicting student attitudes about racial diversity and gender equity. Journal of Diversity in Higher Education, 3(3), 187–199.

16.

Correll

Park

Allegra Smith

(2008). Colorblind and multicultural prejudice reduction strategies in high-conflict situations. Group Processes & Intergroup Relations, 11(4), 471–491.

17.

Crosby

F. J.

Iyer

Sincharoen

(2006). Understanding affirmative action. Annual Review of Psychology, 57(1), 585–611.

18.

Danbold

Unzueta

M. M.

(2020). Drawing the diversity line: Numerical thresholds of diversity vary by group status. Journal of Personality and Social Psychology, 118(2), 283–306. https://doi.org/10.1037/pspi0000182

19.

Darley

J. M.

Gross

P. H.

(1983). A hypothesis-confirming bias in labeling effects. Journal of Personality and Social Psychology, 44(1), 20–33. https://doi.org/10.1037/0022-3514.44.1.20

20.

de Leon

Kay

A. C

. (2021). Ironic Egalitarianism: When hierarchy-attenuating motives increase hierarchy-enhancing beliefs. Organizational Behavior and Human Decision Processes, 164, 86–101. https://doi.org/10.1016/j.obhdp.2021.02.006

21.

Dobbin

Schrage

Kalev

(2015). Rage against the iron cage: The varied effects of bureaucratic personnel reforms on diversity. American Sociological Review, 80(5), 1014–1044.

22.

Dover

T. L.

Kaiser

C. R.

Major

(2020). Mixed signals: The unintended effects of diversity initiatives. Social Issues and Policy Review, 14(1), 152–181.

23.

Dover

T. L.

Major

Kaiser

C. R.

(2016). Members of high-status groups are threatened by pro-diversity organizational messages. Journal of Experimental Social Psychology, 62, 58–67.

24.

Dupree

C. H.

Torrez

(2021). Hierarchy profiling: How and why a job’s perceived impact on inequality affects racial hiring evaluations. Journal of Experimental Social Psychology, 96, 104185.

25.

Fath

Larrick

R. P.

Soll

J. B.

Zhu

(2021). Why putting on blinders can help us see more clearly. MIT Sloan Management Review, 62(4), 38–45.

26.

Federico

C. M.

Sidanius

(2002). Racism, ideology, and affirmative action revisited: The antecedents and consequences of principled objections to affirmative action. Journal of Personality and Social Psychology, 82(4), 488–502. https://doi.org/10.1037//0022-3514.82.4.488

27.

Feintzeig

(2016, January 5). The boss doesn’t want your resume. The Wall Street Journal. www.wsj.com/articles/the-boss-doesnt-want-your-resume-1452025908

28.

Gaddis

S. M.

(2017). How Black are Lakisha and Jamal? Racial perceptions from names used in correspondence audit studies. Sociological Science, 4(19), 469–489. https://doi.org/10.15195/v4.a19

29.

Gawronski

Ledgerwood

Eastwick

P. W.

(2020). Implicit bias and antidiscrimination policy. Policy Insights from the Behavioral and Brain Sciences, 7(2), 99–106. https://doi.org/10.1177/2372732220939128

30.

Germano

A. L.

Ziegler

S. A.

Banham

Cheryan

(2021). Do diversity awards discourage applicants from marginalized groups from pursuing more lucrative opportunities? Psychological Science, 32(9), 1346–1361.

31.

Goldin

Rouse

(2000). Orchestrating impartiality: The impact of “blind” auditions on female musicians. American Economic Review, 90(4), 715–741.

32.

Greig

S. M.

Coundouris

S. P.

Fletcher

G. L.

Henry

J. D.

Hippel

C. V.

(2023). The help that hinders? A meta-analysis of reactions to affirmative action. European Journal of Social Psychology, 53(7), 1475–1523.

33.

Gutiérrez

A. S.

Unzueta

M. M

. (2013). Are admissions decisions based on family ties fairer than those that consider race? Social dominance orientation and attitudes toward legacy vs. affirmative action policies. Journal of Experimental Social Psychology, 49(3), 554–558. https://doi.org/10.1016/j.jesp.2012.10.011

34.

Haley

Sidanius

(2005). Person-organization congruence and the maintenance of group-based social hierarchy: A social dominance perspective. Group Processes & Intergroup Relations, 8(2), 187–203.

35.

Harrison

D. A.

Kravitz

D. A.

Mayer

D. M.

Leslie

L. M.

Lev-Arey

(2006). Understanding attitudes toward affirmative action programs in employment: Addressing reactions to redressing discrimination. Journal of Applied Psychology, 91, 1013–1036.

36.

J. C.

Kang

S. K.

Lacetera

(2021). Opt-out choice framing attenuates gender differences in the decision to compete in the laboratory and in the field. Proceedings of the National Academy of Sciences, 118(42), 1. https://doi.org/10.1073/pnas.2108337118

37.

Herring

(2009). Does diversity pay? Race, gender, and the business case for diversity. American Sociological Review, 74(2), 208–224.

38.

A. K.

Sidanius

Kteily

Sheehy-Skeffington

Pratto

Henkel

K. E.

Stewart

A. L.

(2015). The nature of social dominance orientation: Theorizing and measuring preferences for intergroup inequality using the new SDO-scale. Journal of Personality and Social Psychology, 109(6), 1003–1028.

39.

G. C.

Unzueta

M. M.

(2015). Antiegalitarians for affirmative action? When social dominance orientation is positively related to support for egalitarian social policies. Journal of Applied Social Psychology, 45(8), 451–460. https://doi.org/10.1111/jasp.12311

40.

Hoogendoorn

Oosterbeek

Van Praag

(2013). The impact of gender diversity on the performance of business teams: Evidence from a field experiment. Management Science, 59(7), 1514–1528.

41.

Jost

J. T.

Glaser

Kruglanski

A. W.

Sulloway

F. J.

(2003). Political conservatism as motivated social cognition. Psychological Bulletin, 129(3), 339–375.

42.

Kaiser

C. R.

Major

Jurcevic

Dover

T. L.

Brady

L. M.

Shapiro

J. R.

(2013). Presumed fair: Ironic effects of organizational diversity structures. Journal of Personality and Social Psychology, 104(3), 504–519.

43.

Kottasova

(2015, October 26). Big finance tackles racism by hiring “nameless” graduates. CNN Money. money.cnn.com/2015/10/26/news/blind-hiring-white-sounding-names/

44.

Kraus

M. W.

Torrez

Hollie

(2022). How narratives of racial progress create barriers to diversity, equity, and inclusion in organizations. Current Opinion in Psychology, 43, 108–113.

45.

Krause

Rinne

Zimmermann

K. F.

(2012). Anonymous job applications in Europe. IZA Journal of European Labor Studies, 1(1), 1–20. https://doi.org/10.1186/2193-9012-1-5

46.

Kravitz

D. A.

Bludau

T. M.

Klineberg

S. L.

(2008). The impact of anticipated consequences, respondent group, and strength of affirmative action plan on affirmative action attitudes. Group & Organization Management, 33(4), 361–391.

47.

Kravitz

D. A.

Klineberg

S. L.

(2000). Reactions to two versions of affirmative action among Whites, Blacks, and Hispanics. Journal of Applied Psychology, 85(4), 597–611.

48.

Kteily

N. S.

Sheehy-Skeffington

A. K.

(2017). Hierarchy in the eye of the beholder: (Anti-)egalitarianism shapes perceived levels of social inequality. Journal of Personality and Social Psychology, 112(1), 136–159.

49.

Kteily

N. S.

Sidanius

Levin

(2011). Social dominance orientation: Cause or “mere effect?” Evidence for SDO as a causal predictor of prejudice and discrimination against ethnic and racial outgroups. Journal of Experimental Social Psychology, 47(1), 208–214.

50.

Kung

F. Y.

Brienza

J. P.

Chao

M. M.

(2023). Mixed reactions to multicultural (vs. colorblind) diversity approach signals: A lay theories of culture perspective. Journal of Experimental Psychology: Applied, 29(1), 162.

51.

Lai

C. K.

Skinner

A. L.

Cooley

Murrar

Brauer

Devos

Nosek

B. A.

(2016). Reducing implicit racial preferences: II. Intervention effectiveness across time. Journal of Experimental Psychology: General, 145(8), 1001–1016. https://doi.org/10.1037/xge0000179

52.

Lerner

M. J.

Miller

D. T.

(1978). Just world research and the attribution process: Looking back and ahead. Psychological Bulletin, 85(5), 1030–1051.

53.

Leslie

L. M.

(2019). Diversity initiative effectiveness: A typological theory of unintended consequences. Academy of Management Review, 44(3), 538–563.

54.

McCoy

S. K.

Major

(2007). Priming meritocracy and the psychological justification of inequality. Journal of Experimental Social Psychology, 43(3), 341–351.

55.

McLeod

P. L.

Lobel

S. A.

Cox

T. H.

Jr. (1996). Ethnic diversity and creativity in small groups. Small Group Research, 27(2), 248–264.

56.

Norton

M. I.

Vandello

J. A.

Darley

J. M.

(2004). Casuistry and social category bias. Journal of Personality and Social Psychology, 87(6), 817–831.

57.

Onyeador

I. N.

Hudson

S. T. J.

Lewis

N. A.

Jr. (2021). Moving beyond implicit bias training: Policy insights for increasing organizational diversity. Policy Insights From the Behavioral and Brain Sciences, 8(1), 19–26. https://doi.org/10.1177/2372732220983840

58.

Plaut

V. C.

Garnett

F. G.

Buffardi

L. E.

Sanchez-Burks

(2011). “What about me?” Perceptions of exclusion and Whites’ reactions to multiculturalism. Journal of Personality and Social Psychology, 101(2), 337–353.

59.

Plaut

V. C.

Thomas

K. M.

Hurd

Romano

C. A.

(2018). Do color blindness and multiculturalism remedy or foster discrimination and racism? Current Directions in Psychological Science, 27(3), 200–206.

60.

Pratto

Espinoza

(2001). Gender, ethnicity, and power. Journal of Social Issues, 57(4), 763–780.

61.

Pratto

Sidanius

Stallworth

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. https://doi.org/10.1037/0022-3514.67.4.741

62.

Proudfoot

Berry

Chang

E. H.

Kay

M. B.

(2024). The diversity heuristic: How team demographic composition influences judgments of team creativity. Management Science, 70(6), 3879–3901.

63.

Quillian

Pager

Hexel

Midtbøen

A. H.

(2017). Meta-analysis of field experiments shows no change in racial discrimination in hiring over time. Proceedings of the National Academy of Sciences, 114(41), 10870–10875.

64.

Rattan

Ambady

(2013). Diversity ideologies and intergroup relations: An examination of colorblindness and multiculturalism. European Journal of Social Psychology, 43(1), 12–21.

65.

Richard

O. C.

Murthi

B. S.

Ismail

(2007). The impact of racial diversity on intermediate and long-term performance: The moderating role of environmental context. Strategic Management Journal, 28(12), 1213–1233.

66.

Rivera

L. A.

Tilcsik

(2016). Class advantage, commitment penalty: The gendered effect of social class signals in an elite labor market. American Sociological Review, 81(6), 1097–1131.

67.

Rucker

J. M.

Richeson

J. A.

(2021). Toward an understanding of structural racism: Implications for criminal justice. Science, 374(6565), 286–290.

68.

Scherer

A. M.

Windschitl

P. D.

Graham

(2015). An ideological house of mirrors: Political stereotypes as exaggerations of motivated social cognition differences. Social Psychological and Personality Science, 6(2), 201–209. https://doi.org/10.1177/1948550614549385

69.

Sidanius

Devereux

Pratto

(1992). A comparison of symbolic racism theory and social dominance theory as explanations for racial policy attitudes. The Journal of Social Psychology, 132(3), 377–395.

70.

Sidanius

Pratto

(1999). Social dominance: An intergroup theory of social hierarchy and oppression. Cambridge University Press.

71.

Sidanius

Pratto

Sinclair

Van Laar

(1996). Mother Teresa meets Genghis Khan: The dialectics of hierarchy-enhancing and hierarchy-attenuating career choices. Social Justice Research, 9(2), 145–170. https://doi.org/10.1007/BF02198077

72.

Sinaceur

Thomas-Hunt

M. C.

Neale

M. A.

O’Neill

O. A.

Haag

(2010). Accuracy and perceived expert status in group decisions: When minority members make majority members more accurate privately. Personality and Social Psychology Bulletin, 36(3), 423–437.

73.

Sommers

S. R.

(2006). On racial diversity and group decision making: Identifying multiple effects of racial composition on jury deliberations. Journal of Personality and Social Psychology, 90(4), 597–612.

74.

Stephens

N. M.

Rivera

L. A.

Townsend

S. S.

(2020). The cycle of workplace bias and how to interrupt it. Research in Organizational Behavior, 40, 100137.

75.

Torrez

Hollie

Richeson

J. A.

Kraus

M. W.

(2024). The misperception of organizational racial progress toward diversity, equity, and inclusion. American Psychologist, 79(4), 581–592.

76.

Unzueta

M. M.

Knowles

E. D.

G. C.

(2012). Diversity is what you want it to be: How social-dominance motives affect construals of diversity. Psychological Science, 23(3), 303–309. https://doi.org/10.1177/0956797611426727

77.

Waldfogel

H. B.

Sheehy-Skeffington

Hauser

O. P.

A. K.

Kteily

N. S.

(2021). Ideology selectively shapes attention to inequality. Proceedings of the National Academy of Sciences, 118(14), Article e2023985118.

78.

Wright

Ferris

S. P.

Hiller

J. S.

Kroll

(1995). Competitiveness through management of diversity: Effects on stock price valuation. Academy of Management Journal, 38(1), 272–287. https://doi.org/10.2307/256736

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.27 MB