Antecedents of facades of conformity: when can employees “ be themselves ” ? Antecedents of facades of conformity

Purpose – An individual engages in a façade of conformity by attempting to appear to embrace their organization ’ s values when, in truth, they do not. While numerous studies investigate the negative outcomes associated with facades of conformity, fewer studies consider its antecedents. Therefore, this study aims to investigate the association between diversity-related influences – including individuals ’ beliefs, other unit members ’ beliefs, unit gender diversity and unit racial diversity – and individuals ’ propensities to engage in a façade of conformity. Design/methodology/approach – This paper administered an online survey to 2,122 employees nested within151unitslocatedatahospitalsitelocatedinthesoutheasternUnitedStates.Hierarchicallinearmodelingandrelativeweightsanalyseswereusedtotestthestudyhypotheseswhichaimedtodeterminehowobjectivediversityandperceptionsassociatedwithdiversityincreaseordiminishfacadesofconformity. Findings – In this paper individuals ’ and other unit members ’ beliefs that their organization values diversity were negatively associated with facades of conformity; however,there was a positive association between unit gender diversity and facades of conformity. There were no statistically significant associations involving unit racialdiversityorinteractiveeffects.Overall,theresultsindicatethatitislesslikelythatemployeeswillengageinfaçadesofconformitywhendiversityisvaluedwithinorganizations. Originality/value – By further expanding understanding of the concept of façades of conformity within the humanities and social sciences literature, this study highlight the importance of allowing and encouraging employees to “ be themselves. ”


Introduction
Quality of work life is a multifaceted concept that focuses on the favorable interactions among individual workers, the nature of work itself and the environment in which the work occurs (Bagtasos, 2011;Martel and Dupuis, 2006). Although this notion is now decades old (e.g. Elizur and Shye, 1990;Kirby and Harter, 2001;Lawler, 1982;Nadler and Lawler, 1983;Walton, 1980), there has been a renewed focus in recent years considering how diversity within organizational workforcesand the management of that diversityaffects individuals' work experiences (Badawy et al., 2018;Kara et al., 2018;Mart ınez-Buelvas et al., 2021;Tarigan et al., 2021). This renewed focus coincides with continuously changing workforce demographics and growing societal attention to various diversity-related issues (e.g. Chawla et al., 2021). Indeed, the topic of workforce diversity is receiving "consistent and increasing attention by organizations, the business media, and the popular press" (Roberson et al., 2017, p. 483).
Although diversity is both important and generally regarded as a positive attribute within organizations, there remain limited investigations of façades of conformity, or "false representations created by employees to appear as if they embrace organizational values" (Hewlin, 2003, p. 633). This is an important omission since the presence of diversityor lack thereofinfluences individuals' behaviors at work and their workplace experiences. Notably, some studies confirm that there are numerous downsides to facades of conformity and, viewed from the opposite angle, numerous benefits to authentic displays of behavior and selfexpression (Cable and Kay, 2012;Hewlin et al., 2017). Yet consideration of the underlying causes and antecedents of facades of conformity remain under-investigated.
Using social information processing theory as a lens for our investigation aimed at bridging the gap in this unique area of diversity research, we propose and empirically test a multilevel model with two components. First, we consider the association between individuals' perceptions of how diversity is valued in their organization and facades of conformity. Second, we consider how three unit-level influencesgender diversity within a unit, racial diversity within a unit and other unit members' diversity perceptionsare associated with facades of conformity (both as predictors and cross-level moderators). Overall, we anticipate that individual and other unit members' positive diversity perceptions will be associated with lower levels of facades of conformity, while greater gender diversity and racial diversity within units will be associated with higher levels. In testing our hypotheses, we use hierarchical linear modeling (HLM) and relative weights analysis with data collected from a large hospital system located in the southeastern United States. By further integrating the concept of façades of conformity into the humanities and social sciences literature, we not only highlight the importance of allowing and encouraging employees to "be themselves" but also aim to generate a better understanding of when individuals feel more (or less) compelled to engage in a façade of conformity.

Theoretical background
Facades of conformity Hewlin (2003) coined the term façade of conformity, defining it as an instance where an individual appears to embrace or exhibit support for organizational values when, in fact, they do not. By creating façades of conformity, individuals engage in behaviorsincluding emotional displays, verbal statements and gesturesto signal their alignment with the organization's espoused values (Phillips et al., 2016). This can include compliant behaviors that adhere to formal procedures or the suppression of dissenting views (Stormer and Devine, 2008). When employees are compelled to engage in a façade of conformity, it drastically reduces their quality of work life. Specifically, facades of conformity are linked to increased psychological strain, work-family conflict, job stress and decreased employee voice, job satisfaction and work engagement (Chou et al., 2019;Hewlin et al., 2017;Liang, 2020).
Previous work distinguishes facades of conformity from similar concepts like surface acting and impression management (Hewlin, 2003). However, it is also important to distinguish facades of conformity from psychological safety, another commonly studied construct that has gained much research attention and shares similar importance in promoting quality of work life (e.g. Edmondson and Lei, 2014;Frazier et al., 2017;Newman et al., 2017). Psychological safety is defined as a "shared belief held by members of a team that the team is safe for interpersonal risk taking" (Edmondson, 1999, p. 350). Despite some conceptual overlap, a few differences are evident. For example, a facade of conformity is a behavior whereas psychological safety is more perceptual in nature. Psychological safetyas a shared belief among team membersis often operationalized as a workgroup level construct whereas facades of conformity reside as an individual-level variable. The critical distinction comes in terms of the conceptual underpinnings: psychological safety primarily focuses on learning and risk taking in regard to making, speaking up about and correcting mistakes (Tucker and Edmondson, 2003) whereas facades of conformity primarily focuses on (the expression of) values, specifically when the organization's values conflict with one's personal values that are "linked to one's identity and rooted in one's upbringing, socioeconomic status, and cultural background" (Hewlin, 2003, p. 634). While facades of conformity may involve behaviors associated with disclosing mistakes, it is broader in nature in that it also includes behaviors involving disclosure about personal information such as espousing one's political views, cultural values or religious beliefs (Doblhofer et al., 2019).
While a full review of the facades of conformity literature is beyond the scope of our work, it is important to understand the causes of facades of conformity given the negative consequences associated with them. While there is much research focusing on the negative downstream effects, a continued focus on downstream effects risks the classic issue of focusing on the symptoms of a problem rather than addressing its root cause (Tucker and Spear, 2006). From a practical standpoint, organizations should aim to understand and fix the underling factors driving the use of façades of conformity among their employees (e.g. the ways in which diversity is portrayed, discussed, viewed and valued among organizational members). In the next section, we develop a set of hypotheses on the basis of social information processing theory to investigate why employees engage in facades of conformity.
Social information processing of diversity cues from the organization Social information processing theory suggests that individuals take in cues and make subsequent attributions to better understand their work environment (Salancik and Pfeffer, 1978;Zalesny and Ford, 1990). These cues provide insight into the "social reality" in which individuals are embedded and, once processed, inform action regarding acceptable and unacceptable behaviors. One way this occurs is through the way in which the organization treats employees fairly and equally, regardless of their various individual and unique attributes (McKay and Avery, 2015). These cues may be subtle and can come from anywhere in the organization, including through the presence of diversity, equity and inclusion initiatives or through brochures and pamphlets that pictorially depict gender and racial diversity among coworkers or employee-customer interactions (Avery et al., 2004). Despite their subtlety, these cues can be quite powerful. For example, job applicants tend to have stronger memory and better information recall of recruitment websites that include racial diversity cues (Walker et al., 2012).
On the one hand, cues that indicate non-participative work environmentsdefined as those "lacking tolerance for members expressing diverse ideas, opinions, and values"should encourage facades of conformity (Hewlin, 2009, p. 729;Phillips et al., 2016). On the other hand, pro-diversity cueswhich create numerous downstream benefits including improved job satisfaction, commitment, engagement and performance (Holmes et al., 2021) are likely to minimize facades of conformity such that individuals feel free to act in a way that is consistent with their own values (even if those values are not in complete alignment with the organization).
H1. When individuals believe that their organization values diversity, they will be less likely to use of facades of conformity.
Sources of social information within the unit While social information cues are available from numerous sources, the primary source of social information for individuals comes from experiences within their unit or workgroup (Perrigino et al., 2021a). A readily available source of social information comes from surfacelevel diversity, or easily observable features about unit members including their race and gender (Harrison et al., 1998). Research commonly finds that the presence of surface-level diversitymore formally defined as "the extent to which a unit (e.g. a workgroup or organization) is heterogeneous with respect to demographic attributes"has a negative impact on both individual and unit-level outcomes (Pelled et al., 1999, p. 1; see also Harrison and Klein, 2007). Faultline theory helps explain why this occurs: faultlines are "hypothetical dividing lines that may split a group into subgroups based on one or more attributes" including gender and race (Lau and Murnighan, 1998, p. 328). Although faultlines in and of themselves are neither inherently good nor inherently bad, a large body of research suggests that the presence of surface-level diversity may create negative effects (e.g. Lau and Murnighan, 2005; see also van Knippenberg and Mell, 2016). For example, the presence of gender diversity within units negatively affects creativity (Pearsall et al., 2008) while both gender diversity and racial diversity negatively affect the degree to which members are "on the same page" with each other (Fisher et al., 2012).
Connecting back to social information processing theory, surface-level diversity not only allows for the creation of demographic faultlines within units but also activates stereotypedriven schemas (Priyashantha et al., 2021;Stanciu, 2017). On the one hand, a lack of demographic diversity is likely to decrease the use of facades of conformity since individuals tend to be more comfortable around "similar others" (Montoya and Horton, 2013). In other words, the need to engage in a façade of conformity may be lower when around similar others because the members of the unit may have more commonalities and, by extension, can feel freer to express their viewpoints and be their true selves. On the other hand, the presence of demographic diversity is likely to increase the use of facades of conformity where individuals feel more inhibited in being their true selves. van Dijk et al. (2017) explain that this reluctance stems from a fear of demonstrating behavior that defies stereotypical norms. In this situation, it might be more difficult to process a complexity of social information regarding what types of viewpoints are more (versus less) acceptable. Until additional social information in garnered (e.g. interactions with unit members to allow for a deeper understanding of the social situation; discussed next), we anticipate that the use of facades of conformity will be more common in units characterized by greater demographic diversity. Specifically, we argue that this will be the case for gender diversity and racial diversity since these are two of the most salient and readily observable sources of social information.
H2. (a) Unit gender diversity and (b) unit racial diversity are positively associated with an individual's use of facades of conformity.
In contrast to surface-level diversity, deep-level diversity "includes differences among members' attitudes, beliefs, and values" and manifests through verbal and nonverbal communication with others (Harrison et al., 1998, p. 98). From a social information processing standpoint, these interactions provide additional cues as to what types of behavior are acceptable or unacceptable (Luria, 2008). Whereas the presence of surface-level diversity might indicate that deep-level diversity is present (e.g. gender and racial diversity as indicative of different attitudes, beliefs and values based on demographic faultlines within units), the connection between the two is far more complex (Nkomo et al., 2019). For example, the presence of surface-level diversity could mask a lack of deep-level diversity (i.e. a demographically dissimilar unit where members share similar beliefs and values) while the lack of surface-level diversity could mask the presence of deep-level diversity (i.e. a demographically homogeneous unit with divergent beliefs and values). Thus, we can consider the two separately. A potentially powerful cue regarding facades of conformity concerns the extent to which unit members believe that their organization values and supports the presence of demographic diversity within the workplace (Dwertmann et al., 2016). When other members in an individual's unit consider the organization as supportive of diversity, individuals should be less likely to use facades of conformity since the social information cues emanating from these interactions indicate that it is acceptable to be oneselfregardless of whether surface-level diversity is present. For example, a unit might have a wide degree of racial diversity and a 50/50 balance of men and women. Yet if interactions include the espousing of diversity-related values and promote the acceptance of differing viewpoints, we anticipate that this will diminish the need for individuals to engage in facades of conformity. Indeed, these types of positive interactions should signal to the individual the acceptability of differences in beliefs, opinions or values within the unit, hence the reduced need to display a façade of conformity.
H3. Other unit members' perceptions that their organization values diversity are negatively associated with an individual's use of facades of conformity.
Finally, we note that unit-level characteristics can have both a direct and moderating effect on individual-level outcomes (see Kuenzi and Schminke, 2009 for a review). In other words, we anticipate that unit gender diversity and unit racial diversity not only are likely to (1) increase the use of facades of conformity (per Hypotheses 2a and 2b), but also are likely to (2) attenuate the proposed relationship in Hypothesis 1. Similarly, we anticipate that other unit members' perceptions that their organization values diversity not only are likely to reduce the use of facades of conformity (per Hypothesis 3), but also are likely to enhance the proposed relationship in Hypothesis 1. Figure 1 depicts our complete hypothesized model.
H4a. Unit gender diversity moderates the relationship between individuals' beliefs that their organization values diversity and using facades of conformity, such that this relationship is weakened when there is more diversity within units.
H4b. Unit racial diversity moderates the relationship between individuals' beliefs that their organization values diversity and using facades of conformity, such that this relationship is weakened when there is more diversity within units.
H4c. Other unit members' perceptions that their organization values diversity moderates the relationship between the individual's beliefs that their organization values diversity and using facades of conformity, such that this relationship is strengthened when there are more positive diversity perceptions within units.

Study design
Data collection occurred at a large healthcare system with multiple locations in the southeastern United States. Approximately 12,000 employees were employed at the time we conducted this research. The healthcare system contracted a workforce analytics company to launch an organization-wide cultural assessment survey. As part of the survey, we included Antecedents of facades of conformity specific survey measures that pertained to the hypotheses developed above. Although the survey was sent to all employees, we focused our investigation on the system's primary location that included approximately 4,000 employees. We obtained Institutional Review Board approval for this study. The author team completed all required fields explaining the purpose of the study and included descriptions of planned procedures, a draft of the e-mail announcement that would be sent to employees, a list of items that would be included on the survey and signed nondisclosure agreements among all collaborating parties where appropriate. As per the announcement email, participants were ensured that their responses would be both confidential and anonymous such that: "No one at [organization] will ever see or have access to individual participant responses." Completed survey responses were received directly by the workforce analytics company. The company cleaned and anonymized the data, providing a reduced dataset on which the research team could conduct analyses for the current study. This created a tradeoff: we ensured confidentiality and anonymity to enhance participation and the response rate but lost the ability to check certain aspects of our data regarding, for example, unit-level response rates. We return to this point in the Discussion section.

Sample
A number of 2,122 employees (response rate 53%) nested within 151 units completed the survey. Regarding the sample, a wide range of occupational categorizations were represented, with registered nurses (37%), pharmacy and nurse technicians (25%), facility staff (e.g. food and cleaning; 12%) and therapists (8%) representing the majority of respondents. On average, participants were 41.19 years old (standard deviation 5 12.92 years) and had worked for the organization for 8.82 years (standard deviation 5 8.87 years). 82% of the respondents identified as female. 57% of the respondents identified as Caucasian, 29% of the respondents identified as Black/African American, 5% of the respondents identified as Asian, 4% of the respondents identified as Hispanic/Latino, while the remaining 5% were categorized into an "other" category (including American Indian/Alaskan Native; Native American/Pacific Islander and individuals who chose not to identify with any categorization). Because relevant descriptive unit characteristicsincluding gender diversity, racial diversity and sizewere included as variables in our analyses, we discuss these in the section below.

Measures
Unless noted otherwise, scales were based on a 5-point Likert scale with scores ranging from 1 5 strongly disagree to 5 5 strongly agree. We also identify whether each measure refers to an individual-level variable (level 1) or a unit-level variable (level 2).
Individual perceptions that the organization values diversity (Level 1 predictor). We assessed individual perceptions that the organization values diversity using two items based on previous research (McKay et al., 2011): "This organization demonstrates a commitment to workforce diversity" and "This organization values employees from different backgrounds" (α 5 0.87).
Unit members' perceptions that the organization values diversity (Level 2 predictor). We used the same two items to assess unit members' perceptions that the organization values diversity by taking an average of each of the individual unit members' scores based on an additive composition model (Chan, 1998). Importantly, this involved a separate calculation where we removed the focal individual's score from this aggregation to avoid contamination issues that could result in inflated scores (see Bhave et al., 2010). For example, in a hypothetical 3-member unit where individual perception scores were 3.0, 4.0 and 5.0, the unit average for the individual with a score of 3.0 would be 4.5 (average of 4.0 and 5.0); the unit average for the individual with a score of 4.0 would be 4.0 (average of 3.0 and 5.0) and the unit average for the individual with a score of 5.0 would be 3.5 (average of 3.0 and 4.0). Consistent with our conceptualization above, this operationalization accurately captures other unit members' perceptions.
Unit gender diversity (Level 2 predictor). Gender was a dichotomous variable, with all 2,122 participants identifying as either male or female. We used Blau's index of heterogeneity to calculate unit gender diversity (see Campbell and M ınguez-Vera, 2008). With two categories, scores can range from 0.00 (no unit gender diversity; e.g. an all-female unit) to 0.50 (maximum gender diversity; i.e. a unit where there is an equal number of males and females). Within our sample, scores covered the full range (mean 5 0.24, standard deviation 5 0.15).
Unit racial diversity (Level 2 predictor). Participants were asked to identify themselves based on equal employment opportunity (EEO) class. We condensed this into five categories: Caucasian, Black/African American, Asian, Hispanic/Latino and an "other" category. We again used Blau's index of heterogeneity to calculate unit racial diversity. With five categories, the possible range of scores goes from 0.00 (no unit racial diversity) to 0.80 (i.e. all five categories equally represented). Within our sample, unit scores ranged from 0.00 to 0.72 (mean 5 0.44, standard deviation 5 0.19).
Facades of conformity (Level 1 outcome). To assess facades of conformity, we used a fouritem scale derived from Hewlin's (2009) work. All items had the stem, "I hide my true self at work in the following area:" with "physical," "cultural," "spiritual" and "emotional" completing each stem (α 5 0.91). Higher scores indicated a greater propensity for someone to engage in a façade of conformity.
Control variables. Best practice suggests that control variables should be included in analyses when there is theoretical justification to do so (Bernerth and Aguinis, 2016). Previous research suggests that workgroup size can influence perceptions associated with diversity-related constructs (Boehm et al., 2014). Given the wide variability in unit size (mean 5 14.05; standard deviation 5 16.37), we included unit size as a control variable.

Planned analyses
Our data consists of individuals (level 1) nested within units (level 2). As a result, we used HLM to examine our hypotheses. HLM is appropriate in situations where grouping or clustering occurs since error terms are not independent and the use of ordinary least squares (OLS) regression risks generating incorrect error terms and results (Garson, 2013;Raudenbush and Bryk, 2002). Stated in simpler terms, there is an increased chance of erroneously identifying statistically significant results when using OLS with nested data. When using HLM, a minimum level-2 sample size of n 5 30 is recommended to generate enough power to detect statistically significant effects, while the accumulation of research in the organizational behavior and industrial/organizational psychology literature finds that a sample size of n 5 65 is typical (Shen et al., 2011). With 151 units, we were confident that we had enough power to detect significant effects if they existed. Table 1 displays the means, standard deviations, reliabilities and correlations among the study variables. Table 2 displays the results of our hypotheses tests. Model 1 is the null model. The null model does not include any predictors but provides useful information criteriaincluding the log-likelihood statistic, Akaike's Information Criteria (AIC), and Schwarz's Bayesian Criterion (BIC)for comparing subsequent results such that lower numbers indicate a better fit to the data. In Model 2, we entered unit size as a control variable. This did not have any statistically significant effect on the facades of conformity outcome andbased on increases in the log-likelihood, AIC and BIC statisticsindicated a worse-fitting model to the data. In Step 3, we entered the four predictors. Hypothesis 1 predicted a negative association between individuals' beliefs that their organization valued diversity and façades of conformity; this hypothesis received strong support (B 5 À0.323, p < 0.001). Hypothesis 2a also received support: unit gender diversity had a positive association with facades of conformity (B 5 0.504, p < 0.01), suggesting individuals were more likely to engage in a façade of conformity when gender diversity within the unit was higher. Hypothesis 2b received minimal support: although racial diversity had a positive association with facades of conformity (B 5 0.241), the effect was statistically significant only at a p < 0.10 level. Hypothesis 3 received support: there was a negative association between other unit members' perceptions that the organization values diversity and the individual's facades of conformity (B 5 À0.165, p < 0.01). Notably, the loglikelihood, AIC and BIC statistics all decreased, indicating that this model was a better fit to the ΔR 2 0.00 0.08 Note(s): ***p < 0.001; **p < 0.01; *p < 0.05; t p < 0.10; Dependent Variable 5 Facades of Conformity; n 5 2,122 employees nested within 151 units at a single hospital site; AIC 5 Akaike's Information Criteria; BIC 5 Schwarz's Bayesian Criterion Table 1. Means, standard deviations, correlations, and reliabilities Table 2. Hierarchical linear modeling (HLM) results data compared to both Model 1 (null) and Model 2 (control variable only). A chi-square difference testwhich compares the log-likelihood ratio across modelsconfirmed that Model 3 was a statistically significant better fit (p < 0.001).

Results
We entered the different interaction terms in Model 4. However, none of these were statistically significant and we rejected Hypotheses 4a, 4b and 4c. For parsimony, we do not include these results in Table 2. However, we note that the log-likelihood, AIC and BIC statistics all increased, confirming that Model 4 was a worse fit and providing justification for retaining Model 3 as our final model. In recognition of the potential for interactive effects among multiple unit-level characteristics (Phillips and Loyd, 2006), we also conducted posthoc analyses by entering additional interaction terms (e.g. the effect of unit gender diversity and unit racial diversity on facades of conformity). However, we again failed to find any statistically significant interactions and did not improve the fit beyond Model 3.
To further examine our results, we conducted a relative weights analysis using the tool provided by Tonidandel and LeBreton (2015). We included this as a post-hoc analysis since relative weights analysisalso referred to as dominance analysisis "a useful tool for practitioners and researchers to identify the relative importance of predictors in a linear regression" (Braun et al., 2019, p. 593). As shown in Table 3, individual diversity perceptions accounted for approximately 76% of the total explained variance (R 2 5 0.08). Combined with the low ICC (1) value of 0.05 in the null model (which suggests that 95% of all variance is attributable to individual-level factors), the relatively weak predictive power of the three level-2 variables helps account for why none of the interaction terms were statistically significant. This may also be unsurprising since effect sizes tend to be stronger when the predictor and outcome variable exist at the same level of analysis (Gully et al., 2002).

Discussion
Individuals' beliefs that their organization values diversity and other unit members' beliefs that the organization values diversity were associated with lower levels of facades of conformity (i.e. a positive result). We also found that unit gender diversity was positively associated with facades of conformity (i.e. a negative result). The distinction across findings is best understood through the lens of social information processing theory. Surface-level diversity including diversity associated with gender and race can cause individuals to feel a sense of reluctance in showing their "true self" around different others. Yet when diversity is valued and supported in the organization, it can allow individuals to get to know and understand each other. When this occurs, individuals should feel more comfortable revealing various beliefs and personal details even if they are divergent from those of other unit members.

Limitations and future research directions
The strengths and contributions of our work discussed above should be considered in tandem with its limitations. Our use of others' perceptions should help alleviate concerns  Table 3.

Relative weights analysis
Antecedents of facades of conformity about common method variance (Bhave et al., 2010). However, the most robust relationship involving the association between individuals' beliefs that their organization values diversity and facades of conformity was based on self-report data. This limitation highlights a broader challenge for future research since both variables are based on individuals' own perceptions and experiences, meaning those individuals might be the most accurate reporting source (Rothbard et al., 2005). To overcome both this and our use of anonymized data, future research can apply more qualitative, interview-based methods to shed additional light on the nuances of these experiences.
Although we followed previous research in addressing facades of conformity as a gestalt, it is multi-faceted in nature. Consistent with the identity theory literature that acknowledges that individuals can possess multiple identities (Ramarajan, 2014), individuals may feel the need to hide somebut not allphysical, cultural, spiritual and/or emotional aspects of themselves. Facades of conformity and impression management are unique, as impression management focuses on "career-related outcomes such as gaining approval from superiors and successful role adaptation" whereas facades of conformity focuses more on the suppression of personal values in work-related contexts (Hewlin, 2009, p. 728). Yet future research can follow the impression management literature by identifying different motives and behaviors. For example, a woman's reasoning for attempting to hide her pregnancy in its initial stages (e.g. Little et al., 2015) may be quite different than someone else's reasoning for attempting to hide their spiritual beliefs or marital status in the workplace (e.g. Perrigino et al., 2021b;Sitzmann and Campbell, 2021).
Despite the strength in sample sizeboth at the individual and unit levels of analysisour data was collected in a single organizational setting. Moreover albeit consistent with other studies collecting data from healthcare organizations our sample was mostly female (82%). Although informative for the research questions at hand, issues of gender diversity and racial diversity may be perceived differently and may differ in their level of sensitivity in other industries, organizations and countries. This makes the generalizability of our results a concern. Gender diversity may be less impactful in nations that ascribe to more egalitarian (vs traditional) gender role norms. For example, the United States ranked 53rd out of 153 countries in the World Economic Forum's Global Gender Gap Report (2019). In the most egalitarian nations including Nordic countries (with Iceland, Norway, Finland and Sweden ranked first through fourth on the list) the influence of gender diversity may be less impactful. The impact of racial diversity may differ within different regions of a country and could be a more powerful force in areas where racial unrest incidents occur. Future research should include more geographically diverse samples and consider macro-level variables (e.g. current political climate; traditional vs egalitarian gender role norms).
Finally, our study's design did not allow us to infer causality. In discussing the implications for practice below, we caution that our hypothesized antecedents of facades of conformity are more accurately described as associations. Future research that uses longitudinal designs and time-lagged waves of data can help overcome this shortcoming.

Practical implications
Connecting our research to practice, our study's findings indicate that effective diversity management should help inhibit employees' uses of facades of conformity. One avenue through which this can occur is via policy implementation. As discussed, organizational policies are important signals and sources of social information that organizational leaders and the organization itself values diversity. However, these policies might prove meaningless if there is no follow-through in terms of implementation (Dwertmann et al., 2016). A large body of literature supports the finding that several types of organizational policies used for "window dressing" purposes often fail to add any significant value (e.g. Bikos, 2020;Kalev et al., 2006;Taylor et al., 2018). In some cases, the existence of these policieswithout any implementationmay even be a source of dissatisfaction for employees that reduces trust in and commitment towards the organization (Kirby and Krone, 2002). To promote diversity within organizationsand, by extension, reduce facades of conformityorganizational leaders and champions of HR policies that promote diversity must follow through with genuine implementation efforts. This is true not only for diversity-focused policies but also for other policies or practices that encourage authentic displays of behaviors (e.g. town hall meetings, open forums and team building; Boss et al., 2018;Hewlin et al., 2017).
A second avenue through which this might occur is authentic leadership, a style of leadership promoting self-expression and awareness of others, the creation of trust among employees and commitment to social values (see Gardner et al., 2011 for a review). Indeed, organizational leaders and individuals in positions of authority within organizations are often seen as role models (Brown and Treviño, 2014). Leaders function as a primary source of social information (Luria, 2008) and a growing body of research finds support for "trickle-down" effects where leaders' behaviors are mimicked or modeled at lower levels of the organization (Byun et al., 2020;Mayer et al., 2009;Ruiz et al., 2011;Stollberger et al., 2019). If organizational leaders wish to encourage more authentic behaviors within their organizations, they can both adopt an authentic leadership style and demonstrate desired, authentic behaviors so that their followers can more freely to do the same. This will be particularly effective when such authentic displays embody support for a diversity of ideas, values, beliefs and opinions.
Finally, policymakers at a more macro level can influence facades of conformity through efforts to destigmatize issues that affect equality. Governments play a key role not only in creating new laws and policies to which both corporations and individuals are expected to adhere but also in shaping normative expectations within society (DiMaggio and Powell, 1983). This top-down influence and coercive pressure provides the ability to legitimize more encompassing practices moving forward which, in turn, can further inhibit individuals' perceived need to engage in facades of conformity. As one example, governmental efforts to expand employees' rights associated with work-life balance can cut down on any potential "backlash" individuals might experience for seeking to use flexible work arrangements or prioritizing life outside of the workplace . At the same time, individuals who are compelled to use the same façade of conformity may band together and seek change, engaging in unified, bottom-up activism efforts with the shared goal of legitimizing and destigmatizing certain perspectives or practices (e.g. Briscoe and Safford, 2008).

Conclusion
As social movements, further raise awareness of diversity-related issues in society, organizational leaders possess a keen awareness that managing diversity is a top priority. The mismanagement of diversity can lead to various pernicious effects, least of all an increase in employees' use of façades of conformity. While more work is required as we reiterate that certain limitations associated with our methodology including the use of cross-sectional data our findings that facades of conformity may be reduced when individuals and other unit members hold positive perceptions of diversity within the organization allow us to close on a more positive note. There is immense potential not only for additional research in this area to uncover more positive findings but also for the creation of positive societal change by fostering more all-inclusive attitudes and beliefs across individuals.