Introduction

Attracting a gender diverse workforce is a major challenge in today’s corporate world. Key drivers of this challenge are the high level of international competition for labor (e.g., Blau and Kahn 2005; Krisor et al. 2013; Pries 2010), the business case for female representation in top management (e.g., Bear et al. 2010; Francoeur et al. 2008; Torchia et al. 2011), and in many countries, the political pressure to increase gender diversity in organizations (e.g., European Directorate-General for Internal Policies 2013). Despite numerous actions aimed at achieving gender equality, women are still often underrepresented in upper management in many countries around the globe (e.g., Grant Thornton 2014). In this study, we explore a potential reason for this circumstance that we refer to as “the paradox of diversity management strategies”—the discrepancy between that which individuals acknowledge as effective for organizations to achieve gender-balanced upper echelons and that which individuals perceive as attractive organizational attributes. Existing evidence indicates that identity conscious initiatives that are either based on opportunity enhancement for, or preferential treatment of, female employees are in fact often more effective for increasing the proportion of women in top management positions than more general and less binding identity blind initiatives such as training or diversity evaluations (Kalev et al. 2006). Despite their effectiveness, however, elements of fairness theory (Folger and Cropanzano 1998, 2001) and previous empirical research suggest the existence of a paradox such that identity conscious initiatives may not always be well received and lead to ‘backlash’ from non-beneficiaries, such as white men (Leslie et al. 2014). This proposed diversity paradox is similar to the diversity-validity dilemma described by Pyburn et al. (2008), whereby organizations’ attempts to increase gender diversity are hindered because some selection procedures that exhibit high validity result in women (as a group) scoring lower than men (as group).

Today, “more than half of the countries in the world have implemented some type of political quota” regarding female representation in public decision-making bodies (Pande and Ford 2011, p. 8). For example, the first empirical long-term investigation of the gender quota in Norway found a positive effect on the number of female board chairs as well as “spill-over effects on top leadership positions” (Wang and Kelan 2013, p. 463). The present study was conducted in Germany, where the parliament recently passed a government-mandated gender quota that will be established by 2016 (German Ministry of Family Affairs 2014b). The German quota is similar to recent action taken by many other countries within (Austria, Belgium, Denmark, Finland, France, Ireland, Italy, Spain, and the Netherlands) and outside the European Union (Iceland, Israel, Switzerland, and Norway). In brief, the German quota regulations recently set forth require both genders to be represented with at least 30 % on the supervisory boardsFootnote 1 of major organizations. In case of non-compliance, board seats reserved for the underrepresented gender have to remain empty. The organizations further have to publicize goals for gender diversity within their executive boards and also in general upper management positions. Given the potential impact of this legislation for today’s workplace, it is important to know whether and how potential employees react toward quota regulations.

With this work, we therefore seek to contribute to a better understanding of the following key practical questions: What happens when organizations employ strategies aimed at increasing the proportion of women in top management teams? Are people attracted to organizations that implement diversity strategies they perceive as effective, or do they wish to avoid such organizations as potential employers? What is the role of governmental legislation in this context? To the best of our knowledge, the current study is the first to combine various types of organizational gender diversity initiatives with the effect of a forthcoming government-mandated gender quota from an employer attractiveness perspective. If, as we argue, a government-mandated gender quota helps to increase the perceived attractiveness of organizations using identity conscious diversity initiatives, then such governmental legislation could become an important driver in resolving the proposed paradox of diversity management strategies. In this respect, one of the most unique features of the current work is the simultaneous consideration of three perspectives on the paradox of diversity management: If decisions at the level of the organization (i.e., implementation of certain initiatives) are perceived differently by individuals depending on the regulations set forth at the higher country-level, our study enables conclusions for individual employees as well as diversity practitioners in organizations and beyond.

The German Context

The German context is of high relevance with regard to our research questions for various reasons. First, the recent legislative changes described previously have led to substantial public scrutiny of gender diversity—job seekers from Germany are thus likely to be attentive to organizational gender diversity attributes. Second, (also) due to governmental pressure, many employers publicly promote their interest and efforts with regard to gender diversity through various outlets such as their corporate websites (Singh and Point 2006). Third, the German labor market shares central gender diversity attributes with other Western societies: For example, despite the fact that the female labor force participation rate is approximately 54 % (United States 56 %, United Kingdom 56 %; Worldbank 2013), “the executive floors in Germany continue to be predominantly a male monoculture”(German Ministry of Family Affairs 2014a, p. 61), comparable to Fortune 500 and FTSE 100 top management teams (Bernardi et al. 2006; Catalyst 2013a; Vinnicombe et al. 2014). Similarly, less than one-third (31 %) of German senior managers are female (United States 20 %, United Kingdom 19 %; Catalyst 2013b), and even less than one-fifth of all board seats (14 %) are occupied by women (United States 17 %, United Kingdom 21 %; Catalyst 2014). Key drivers of these gender differences in Germany are issues related to the division of labor within the household, or, in other words, family issues (European Commission 2012). Although parental support in Germany is traditionally more elaborate than in other countries such as the United States (e.g., Dustmann and Schönberg 2012), the employment rate of women with children is approximately 20 % points lower than the employment rate of childless women (European Commission 2012).

Perceived Effectiveness of Gender-Related Diversity Initiatives

Past research has identified a variety of human resource management practices that can be used to improve organizational diversity (e.g., Kalev et al. 2006; Konrad and Linnehan 1995). Konrad and Linnehan (1995) distinguished identity blind (e.g., management training available to all qualified employees) from identity conscious (e.g., leadership development exclusively for women) initiatives. Although identity blind initiatives are designed to ensure that human resource practices ignore demographic group attributes, stereotypes and biases tend to continue to taint employment decision-making (e.g., King et al. 2006); thus, identity blind initiatives often do not sufficiently improve organizational diversity. By contrast, identity conscious initiatives are designed to ensure that human resource practices consider demographic attributes. Identity conscious initiatives, in turn, can be further classified as (a) opportunity enhancement initiatives whereby target group members are provided with extra resources but their demographic group memberships are not considered in ultimate employment decisions, and (b) preferential treatment initiatives whereby targets are actually given preference in employment decisions (Kravitz 1995; Leslie et al. 2014).

Although field research indicates that identity conscious initiatives are indeed more effective than identity blind initiatives for increasing the proportion of women in management (Kalev et al. 2006), individuals’ perceptions are not always consistent with reality, especially when the issue of gender is involved (McCauley et al. 1988; Swim et al. 1995) and during the early stages of recruitment (for a review, see Connelly et al. 2011). Thus, they are often unable to observe themselves the actual effectiveness of gender diversity initiatives. Given that our proposed paradox is based on individuals’ perceptions regarding the effectiveness of diversity initiatives, and as such perceptions may influence their decisions to pursue employment with an organization, it is important to investigate the extent to which people’s perceptions reflect reality with regard to the effectiveness of gender diversity. Given that identity blind initiatives work only indirectly, individuals should perceive identity conscious initiatives as more effective for increasing the proportion of women in management. Furthermore, preferential treatment initiatives should be perceived as more effective than opportunity enhancement initiatives because the former actually consider group memberships in employment decisions, whereas the latter do not. In other words, we expect that the prescriptiveness (i.e., the extent to which decision makers are constrained by diversity-related policies and procedures when making employment decisions) of diversity initiatives will be related to their perceived effectiveness as follows:

Hypothesis 1a

Individuals perceive identity conscious initiatives as more effective than identity blind initiatives for increasing the proportion of women in upper management.

Hypothesis 1b

Individuals perceive preferential treatment initiatives as more effective than opportunity enhancement initiatives for increasing the proportion of women in upper management.

Impact of Type of Diversity Initiative on Pursuit Intentions

Employers communicate information about their diversity management initiatives on their websites and in other organizational publications such as recruitment brochures and corporate social performance reports (e.g., Singh and Point 2006). Individuals engaged in a job search often examine such sources of information for signals regarding the values of, and working conditions within, organizations in an attempt to maximize their fit with potential employers (e.g., Turban and Greening 1997). The dissemination of diversity-related information through various channels for the purpose of employee recruitment has been conceptualized as a form of organizational impression management intended to signal potential employees that the organization values diversity (Avery and McKay 2006; Goldberg and Allen 2008; Greening and Turban 2000; Turban and Greening 1997). Indeed, research has demonstrated that the inclusion of diversity statements in recruitment materials is positively related to individuals’ perceptions that the organization values diversity (Kim and Gelfand 2003; Rau and Hyland 2003). Notably, however, its impact on the perceived favorability of organizations using them is less clear given that diversity statements have been found to have both negative (e.g., Kim and Gelfand 2003; Martins and Parsons 2007; Richard and Kirby 1998; Williamson et al. 2008) and positive (e.g., Avery 2003; Kim and Gelfand 2003; Martins and Parsons 2007) effects on organizations using them. This inconsistency may be reconciled by differentiating between the types of diversity initiatives.

In the ensuing sections, we use the rationale of counterfactual thinking, derived from fairness theory (Folger and Cropanzano 1998, 2001), as an organizing framework to guide our hypothesis development (see Fig. 1 for an overview).

Fig. 1
figure 1

Overview of research model and hypotheses. The thickness of the lines indicates the strength of the proposed effects, with thicker lines representing stronger effects

Full size image

According to the overarching principle of counterfactuals, identity conscious initiatives may be viewed less favorably than identity blind initiatives because the use of demographic criteria for hiring and promotion (as in identity conscious initiatives) violates both procedural and distributive fairness principles (Bobocel and Farrell 1996; Konrad and Linnehan 1995; Ozawa et al. 1996; Summers 1995). This reasoning is consistent with fairness theory (Folger and Cropanzano 1998, 2001), which argues that when individuals view a decision negatively, they engage in counterfactual thinking about whether another decision would have been better (the would counterfactual), could have been made (the could counterfactual), and should have been made to comply with other ethical or morally appropriate standards (the should counterfactual).

We propose that different types of initiatives lead to different interpretations of counterfactuals, which, in turn, impact people’s pursuit intentions. Since many individuals believe that using identity blind rather than identity conscious initiatives would result in better employment decisions (because they are more directly related to the established and endorsed principle of merit) and that these initiatives should be used because they do not violate ethical or moral standards, they are perceived as more fair than identity conscious initiatives (Ozawa et al. 1996; Shaw et al. 2003; Williamson et al. 2008). Thus, organizations using identity conscious initiatives should be less likely than those using identity blind initiatives to be perceived as attractive (Harrison et al. 2006; Summers 1995). Furthermore, this negative effect is likely to be more pronounced for preferential treatment than opportunity enhancement initiatives due to the higher prescriptiveness of preferential treatment (Harrison et al. 2006). Hence, we hypothesized the following:

Hypothesis 2a

Individuals’ intentions to pursue employment with organizations using identity conscious initiatives are lower than their intentions to pursue employment with those using identity blind initiatives.

Hypothesis 2b

Individuals’ intentions to pursue employment with organizations using preferential treatment initiatives are lower than their intentions to pursue employment with those using opportunity enhancement initiatives.

Moderators of the Relationship between Type of Diversity Initiative and Pursuit Intentions

We expect several factors to moderate the relationship between the type of diversity initiative and pursuit intentions. Specifically, we hypothesize that information which influences individuals’ perceptions regarding organizational accountability for the diversity initiative as well as individual difference characteristics moderate the influence of type of diversity initiative on individuals’ intentions to pursue employment with an organization. Consistent with this view, previous research investigating the influence of an organization’s diversity initiatives on individuals’ intentions to pursue the organization as an employer has considered moderating factors such as justifications for the initiative (Richard and Kirby 1997; 1998; Williamson et al. 2008) and gender (e.g., Greening and Turban 2000).

Explanation for the Diversity Initiative: Quota

Justifications and excuses are two types of explanations that can be used by decision makers to reduce negative reactions to potentially unfavorable decisions, such as the use of identity conscious initiatives for making employment decisions (Bobocel and Farrell 1996; Shaw et al. 2003). While individuals using excuses shift responsibility (and thus blame others) for the decision away from themselves, those using justifications accept responsibility while attempting to legitimize the unfavorable decision (Bies et al. 1988; Shaw et al. 2003). While research indicates that providing justifications for the use of identity conscious initiatives can be integral to successfully managing diversity (Kravitz et al. 2008; Richard and Kirby 1998), altering perceptions of these initiatives by providing justifications is a complex process in that it depends on the type of justification given (Harrison et al. 2006), and a particular justification may increase organizational attraction for some individuals while decreasing it for others (Williamson et al. 2008).

In their study, Bobocel and Farrell (1996, p. 26) found that an explicit excuse, i.e., “I was not responsible for the decision…” was negatively related to perceptions of interactional fairness. This finding may reflect an attempt at defensive impression management (Gardner and Martinko 1988) that backfired. It is possible, however, that an excuse emanating from a source other than the decision maker him-/herself could improve reactions to identity conscious initiatives. We hence propose that a government-mandated quota for the percentage of women in top management positions may serve as an influential excuse for an organization’s use of identity conscious initiatives. In accordance with attribution theory (Weiner 1986), as long as individuals evaluating the initiative are aware of external causes for it, they may be less likely to hold the organization accountable for potential negative outcomes—which in turn may weaken the negative impact of the initiative on the attractiveness of the organization (see above). This argument is similar to that put forth by Avery and McKay (2006) who noted that external attributions for targeted recruitment programs could lead to the impression that the programs are motivated by legal concerns rather than organizational values.

In accordance with our overarching principle of counterfactuals as derived from fairness theory (Folger and Cropanzano 1998, 2001), given that organizations are generally assumed to be responsible for their own human resource initiatives (Avery and McKay 2006), individuals will believe that organizations using identity conscious initiatives generally could have decided to do otherwise. A government-mandated gender quota, however, may weaken this belief because it implies that organizations have to engage in the more prescriptive diversity initiatives in order to achieve the quota—thus mitigating issues of procedural and distributive unfairness (Bobocel and Farrell 1996). This reasoning leads to the following prediction:

Hypothesis 3

Individuals’ intentions to pursue employment with (a) organizations using opportunity enhancement initiatives and (b) organizations using preferential treatment initiatives are higher when there is a government-mandated gender quota regarding the proportion of women in top management teams than when there is no such quota.

Gender

A meta-analysis of attitudes toward diversity initiatives (Harrison et al. 2006) indicates that the gender of the person evaluating the initiative tends not to matter for identity blind initiatives but often does matter for identity conscious initiatives (e.g., Barber and Roehling 1993; Greening and Turban 2000; Konrad and Linnehan 1995; Kravitz and Platania 1993; McNab and Johnston 2002; Summers 1995; Williams and Bauer 1994). More specifically, men tend to have more negative reactions than women to identity conscious initiatives—despite the finding that some women are also concerned about negative personal outcomes such as lower job satisfaction, greater stress, or being perceived as incompetent (e.g., Harrison et al. 2006; Heilman et al. 1992; Gilbert and Stead 1999).

The gender effect found in response to identity conscious initiatives has been explained using various concepts such as self-interest, prejudice, and perceived fairness (e.g., Harrison et al. 2006; Konrad and Linnehan 1995). In accordance with our principle of counterfactuals derived from fairness theory (Folger and Cropanzano 1998, 2001), while both men and women may believe that less prescriptive initiatives would lead to more favorable organizational outcomes, men should be more likely than women to believe that less prescriptive initiatives would lead to more favorable personal outcomes for themselves—for example, due to more limited opportunities for promotions. In the same vein, men may be more inclined to think that identity conscious initiatives should not be used because they may expect to profit less from such initiatives than women. Taken together, because men can plausibly imagine more counterfactuals than women, men should be more likely than women to perceive identity conscious initiatives as unfair, and thus less likely to consider the organization attractive.

Hypothesis 4

Women’s intentions to pursue employment with organizations using identity conscious initiatives are higher than men’s intentions to pursue employment with organizations using such initiatives.

Attitudes Toward Diversity

De Meuse and Hostager (2001) argued that organizational diversity can lead to negative, neutral, and positive thoughts among different people, which reflect “the extent to which individuals believe there is value in diversity”(see van Dick et al. 2008, p. 1464). The concept of motivated social cognition (Jost et al. 2003) suggests that people with negative attitudes toward diversity are more likely than those with positive attitudes to believe that organizational outcomes will suffer when organizations use more prescriptive diversity initiatives, i.e., they are more likely to believe the negative effects associated with the diversity-validity dilemma outweigh the positive effects associated with the business case for diversity.

In accordance with our counterfactual principle from fairness theory (Folger and Cropanzano 1998, 2001), this means that people with negative attitudes toward diversity are more likely to believe that using less prescriptive initiatives would lead to better organizational outcomes. Furthermore, we expect that individuals with negative attitudes toward diversity are more likely to believe that organizations should use less prescriptive initiatives because they are more consistent with societal expectations of fairness and morality than more prescriptive initiatives (Shaw et al. 2003). Therefore, beliefs that organizations would be better off using less prescriptive initiatives and that they should do so is likely to have a negative impact on intentions to pursue employment with organizations using more prescriptive initiatives. Thus, we hypothesize the following:

Hypothesis 5

Individuals’ intentions to pursue employment with organizations using identity conscious initiatives are higher, the more positive their attitudes toward diversity are.

Method

Participants

Approximately 9000 subscribers of PsyWeb, a German online panel for psychological research, received a link to a survey. Seven hundred and thirty-two completed surveys were returned, for a response rate of 8 %, which seems satisfactory given that participation in the panel is completely voluntary and not rewarded, as was participation in this study, and also compared to the response rate of other studies conducted via the same panel (e.g., Kanning et al. 2014). Thirty-nine participants were excluded due to factors such as missing data. The final sample size was N = 693. The majority of participants were German nationals (97 %) of which only 8 % indicated they have an immigration background (mainly from Poland). The combination of this information with data on the distribution of different ethnicities among immigrants in Germany (German Federal Statistical Office 2014) leads to the conclusion that the large majority of participants in this study were Caucasian, i.e., the racial majority in Germany.

Measures

We used three 7-point Likert type items (1 = strongly disagree to 7 = strongly agree) to assess the diversity statements’ perceived effectiveness with a higher score indicating higher perceived effectiveness (see Table 1 for sample items of all measures). Similarly, pursuit intentions were assessed using the five intentions toward the company items from Highhouse et al. (2003) organizational attraction scale, with higher scores indicating higher pursuit intentions. We measured attitudes toward diversity with an adapted version of De Meuse and Hostager’s (2001) Reaction-to-Diversity Inventory, an established and validated measure (Hostager and De Meuse 2002). Finally, we used two manipulation checks. First, the manipulation check concerning the quota manipulation (described in detail below) was a scale on which participants indicated how much they agreed with the following statement: “[…] you were given the information that the country in which you imagined conducting a job search will be implementing a government-mandated gender quota for all companies.” Second, to ensure that our manipulations resulted in the intended perception of the initiatives as either identity blind or identity conscious, we assessed the diversity statements’ perceived gender consciousness and concreteness (Table 1). All manipulation check items were answered on 7-point Likert type scales with higher scores indicating higher agreement with the relevant statements.

Table 1 Overview of the applied measures
Full size table

Material: Diversity Statements

The independent variable of statement type was manipulated. To test our hypotheses and to ensure external validity, we developed one identity blind and two identity conscious (opportunity enhancement and preferential treatment) diversity statements that were similar to existing statements included on German employers’ websites. In accordance with previous research (e.g., Avery et al. 2004; Casper et al. 2013; Williamson et al. 2008), we reviewed all recruiting websites of the 30 major companies traded at the German stock exchange to maximize the realism of these statements. The length and writing style of all three statements were consistent.

The identity blind statement focused on equality and the importance of organizational diversity in general, hence women should not be perceived as benefitting more than men from the initiative:

Providing equal opportunities for all our employees is central to our corporate values and leadership principles. That means the skills and personalities of each employee are valued and supported. All our employees are given the opportunity to participate in training on understanding and benefitting from the added value of diversity. In addition, we support families and encourage the balance between work and family life for all our employees.

Both identity conscious statements focused on practices designed to increase the proportion of women in upper management positions. The opportunity enhancement statement was based on the most frequently mentioned gender initiatives on the employer websites of the German companies we reviewed aimed at enhancing opportunities for women:

As a company, we support women through targeted initiatives. To increase the proportion of women in upper management positions, we provide a variety of initiatives for our female employees. Our program, “a network for women,” provides female managers exhibiting high development potential with a platform for exchange. Women also receive coaching in workshops to build essential skills for their career development and prepare them for leadership positions.

The preferential treatment statement also contained specific opportunity enhancement initiatives. However, here the focus was on a voluntarily established (i.e., by the organization itself) gender quota for top management positions:

As a company, we are committed for increasing the proportion of women in management positions. To achieve this goal, we have established a range of initiatives designed to support women (e.g., networking programs, workshops, and training). In addition to these initiatives, we have implemented a gender quota. 30 % of all upper management positions will be filled by women. We have established a company-wide directive requiring special consideration of women, who possess equivalent qualifications to men, in hiring and promotion decisions.

Procedure

The data were collected online. To test our hypotheses, we used a scenario design as established approach in diversity research (e.g., Kidder et al. 2004; Williamson et al. 2008). All participants initially read a brief general instruction in which they were asked to imagine that they were looking for a job abroad (i.e., outside Germany). Participants were then told that they had visited the websites of potential employers, which included information on their diversity initiatives. Participants were also informed that all employers in which they were interested currently had a low proportion of women in upper management (approximately 5 to 10 %), and were then given all three diversity statements simultaneously. The statements were described as being derived from the websites of three different potential employers. Participants were asked to read all three of them. This kind of within-subject approach increases external validity as it comes close to the experience of encountering various potential employers during job searches (e.g., Martins and Parsons 2007).

At this point, the sample was split. One-third of the sample participated in a within-subjects experiment to assess the diversity initiatives of the three potential employers for their perceived effectiveness with regard to increasing the proportion of women in upper management positions. These participants responded to the manipulation check regarding our classification of the statements as either identity blind or identity conscious. The remaining two-thirds participated in a mixed factorial design experiment to assess their intentions to pursue employment with the three organizations whose diversity statements they were provided (within-subjects factor). These participants were exposed to one of two conditions regarding the existence of a government-mandated quota (between-subjects factor). Participants in the first condition (Government Quota condition), read the following information:

Within the next 2 years, the country in which you are conducting your job search will implement a government-mandated gender quota of 30 % for top management team positions. This means all organizations will be required to fill thirty percent of top management team positions with women. Organizations that do not meet this requirement will face extensive penalties and economic sanctions from the government. All organizations in which you are considering employment will be affected by this legislation.

The second condition (No-Government Quota condition) was a control condition in which no information regarding a government-mandated gender quota was provided. Participants in both conditions responded to the government quota manipulation check item. An overview of the study design is depicted in Fig. 2.

Fig. 2
figure 2

Overview of study design

Full size image

We chose to assess the dependent variables of perceived effectiveness and pursuit intentions with different portions of the sample in order to reduce socially desirable responding, which has been found to occur in survey research on issues for which social norms exist (for a review, see Moorman and Podsakoff 1992). Awareness of social expectations influences participants’ responses such that they provide responses that are consistent with (expected) norms to not appear sexist or racist (Swim et al. 1995). Since diversity in the workplace has become a social norm (Walt and Ingley 2003), this research suggests that participants would be reluctant to indicate a preference for working in organizations with diversity initiatives they perceived as ineffective.

Results

Means, standard deviations, and inter-correlations of the main variables employed in this study are presented in Table 2. Our intended categorization of the three statements as either identity blind or identity conscious was successful. A repeated measures one-way analysis of variance (rANOVA) revealed significant differences for gender consciousness [F(1, 236) = 63.10, p < 0.001, \(\eta_{\text{P}}^{2}\) = 0.73] between the identity blind statement with M = 2.62, SD = 1.00, and the two identity conscious statements (opportunity enhancement with M = 5.41, SD = 1.12 and preferential treatment with M = 5.39, SD = 1.23). Similarly, a rANOVA showed that the identity blind statement with M = 3.52, and SD = 1.39 was perceived as significantly less concrete [F(1, 236) = 69.45, p < 0.001, \(\eta_{\text{P}}^{2}\) = 0.23] than the identity conscious statements (opportunity enhancement M = 4.13, SD = 1.58; preferential treatment with, M = 4.94, SD = 1.40).

Table 2 Means, SD, and correlations for main variables
Full size table

Perceived Effectiveness

Table 3 depicts the mean perceived effectiveness as well as the mean pursuit intentions regarding the three diversity initiatives. We conducted a rANOVA to test our hypotheses of effectiveness differences. As predicted, we found an effect of statement type on perceived effectiveness, F(2, 235) = 88.44, p < 0.001, \(\eta_{\text{P}}^{2}\)  = 0.43. As can be seen in the descriptive statistics depicted in Table 3, the identity blind statement (M = 3.84, SD = 1.25) was perceived as less effective than the opportunity enhancement statement (M = 4.10, SD = 1.29) with d = −0.20, and also as less effective than the preferential treatment statement (M = 5.26, SD = 1.25) with d = −1.14. To test our hypotheses, first the two identity conscious statements were combined and then contrasted against the identity blind statement. In support of Hypothesis 1a, the comparison revealed a significantly lower perceived effectiveness of the identity blind statement, F(1, 236) = 66.11, p < 0.001, \(\eta_{\text{P}}^{2}\) = 0.22. As predicted in Hypothesis 1b, we also found that the preferential treatment statement led to higher perceived effectiveness ratings than the opportunity enhancement statement, t(236) = −10.85, p < 0.001 (one-tailed) with d = 0.91.

Table 3 Means and SD of perceived effectiveness ratings and pursuit intentions
Full size table

Pursuit Intentions

We conducted a rANOVA to test our hypotheses of pursuit intention differences for the organization with the identity blind statement as compared to the organizations with the identity conscious statements. Mauchly’s test indicated a violation of the assumption of sphericity, χ 2(2) = 93.72, p < 0.001, and the degrees of freedom were corrected using Greenhouse-Geisser estimate. As predicted, the rANOVA revealed a significant main effect of statement type on pursuit intentions, F(1.69, 766.95) = 195.03, p < 0.001, \(\eta_{\text{P}}^{2}\) = 0.30. In line with expectations, Table 3 shows that pursuit intentions were higher for the organization with the identity blind statement (M = 5.69, SD = 1.13) as compared to the organization with the opportunity enhancement statement (M = 4.19, SD = 1.37) with d = 1.19, and also as compared to the organization with the preferential treatment statement (M = 4.28, SD = 1.51) with d = 1.02. As predicted in Hypothesis 2a, a comparison revealed that the identity blind statement led to significantly higher pursuit intentions as compared to the combined identity conscious statements, F(1,455) = 276.15, p < 0.001, \(\eta_{\text{P}}^{2}\) = 0.38. Contrary to Hypothesis 2b, however, mean pursuit intentions were not significantly different between the organization with the opportunity enhancement statement and the organization with the preferential treatment statement [t(456) = −1.46, p = 0.14, d = −0.06].

Quota Versus No-Quota

Examination of participants’ responses to the manipulation check item on quota revealed a significant difference between the two quota conditions in line with expectations, t(453) = 14.75, p < 0.001. Notably, however, only 56 % of participants in the No-Quota condition and 78 % of participants in the Quota condition answered the manipulation check correctly. Recall that participants indicated on a 7-point Likert type scale their agreement with the following statement: “[…] you were given the information that the country in which you imagined conducting a job search will be implementing a mandatory gender quota for all companies.” As the analysis of the effect of a government-mandated gender quota on pursuit intentions required a careful reading and correct understanding of the instructions and information provided, including participants that failed the manipulation check would reduce internal validity. In other words, without having recognized the quota, one would not expect effects of a quota to emerge. Thus, we decided to test our hypotheses using only participants who answered the manipulation check correctly. We, however, also report results for the full sample, i.e., including participants who answered the manipulation check incorrectly.

Means and standard deviations can be seen in Table 3. Inspection of Table 3 shows that participants who were presented with the quota information indicated higher pursuit intentions than those in the control condition (i.e., No-Quota) for the organization with the preferential treatment statement, t(305) = 1.95, p = 0.03, one-tailed, d = 0.23. Mean pursuit intentions for the opportunity enhancement statement did not differ significantly between participants in the quota condition and the control condition, t(305) = −0.24, p = 0.82, d = −0.03. There was no difference between participants in the quota condition and participants in the control condition, t(305) = −0.35, p = 0.73, d = −0.04, for the organization with the identity blind statement (which was expected because an identity blind statement generally adheres to fairness principles and thus does not invite counterfactuals).

In the analysis including participants that failed the manipulation check, we found no differences in mean pursuit intentions between participants in the quota and control condition for organizations using (a) the opportunity enhancement statement [t(454) = 1.34, p = 0.18, d = −0.13], (b) the preferential treatment statement [t(454) = −0.29, p = 0.77, d = −0.03], or the identity blind statement [t(454) = 0.14, p = 0.89, d = 0.02].

Gender Effects

A two-way rANOVA with gender and statement type as independent variables and pursuit intentions as dependent variable (Greenhouse-Geisser corrected) indicated a significant interaction, F(1.69, 765.2) = 3.63, p = 0.03, \(\eta_{\text{P}}^{2}\) = 0.01. Table 3 shows the means of pursuit intentions by gender and statement type. As predicted in Hypothesis 4, analyses of simple main effects (one-tailed) revealed significantly higher pursuit intentions for women than for men for the preferential treatment statement, t(454) = −3.68, p < 0.001, d = 0.35. Our analysis also revealed a significant effect in that women, as compared to men, who showed slightly higher pursuit intentions for the opportunity enhancement statement, t(454) = −1.63, p = 0.05, d = 0.16. Thus, Hypothesis 4 was largely supported. As expected, there were no differences in pursuit intentions between men and women for the identity blind diversity statement [t(454) = −0.63, p = 0.53, d = 0.06].

Attitudes Toward Diversity

To test the effect of attitudes toward diversity on pursuit intentions, depending on statement type, we conducted a rANOVA with statement type as independent variable, pursuit intentions as dependent variable, and attitudes toward diversity as covariate. Contrary to our prediction, the interaction effect (Greenhouse-Geisser corrected) did not reach statistical significance, F(1,454) = 0.23, p = 0.63, \(\eta_{\text{P}}^{2}\) = 0.001. Thus, Hypothesis 5 (i.e., individuals with more positive attitudes toward diversity have higher pursuit intentions for organizations using identity conscious statements) was not supported. Following De Meuse and Hostager’s (2001) classification, we ran further exploratory analyses and dichotomized participants into a combined category of realists/pessimists (−35 to +10) and optimists (+11 to +35) which was similar to a median-split (Mdn = 10). Once again, we did not find a significant effect of attitudes toward diversity. To further explore whether individuals’ pursuit intentions for the organization with the identity blind statement would be affected by their attitudes toward diversity, we conducted a regression analysis. Attitudes toward diversity were unrelated to pursuit intentions for the organization with the identity blind statement, t(305) = 0.13, p = 0.90.

Discussion

Existing theory (Folger and Cropanzano 1998, 2001) and empirical evidence indicate a major challenge for organizational decision makers. They are under increasing pressure to achieve gender diversity in top management positions, and hence depend on more prescriptive gender initiatives (e.g., preferential treatment) to attain these goals. At the same time, these initiatives might negatively affect an organization’s attractiveness as an employer (Beechler and Woodward 2009; Ng and Burke 2005). This study was designed to investigate the organizational challenge of a diversity management paradox, whereby organizational initiatives perceived as effective for achieving gender-balanced upper echelons are not perceived as attractive organizational attributes. The key finding of this experimental study, depicted in Fig. 3, supports the existence of the paradox we proposed. Identity conscious initiatives were perceived as more effective than identity blind initiatives; thus, empirically establishing that individuals’ perceptions of the differential effectiveness of the two types of diversity initiatives (identity blind and identity conscious) are consistent with actual differences found in field research (Kalev et al. 2006). Despite being perceived as more effective, however, identity conscious initiatives negatively impacted the perceived favorability of organizations that use them.

Fig. 3
figure 3

Illustration of the paradox of diversity management strategies

Full size image

Notably, however, participants who recognized that a government-mandated gender quota would be forthcoming were in fact found to indicate higher pursuit intentions for organizations using a preferential treatment initiative, as predicted. This finding suggests a unique interplay between different theoretical levels of analysis in that governmental legislation (country-level) impacts pursuit intentions (individual-level) toward companies using different diversity strategies (organizational level). This finding highlights that the study and practice of diversity management is well advised to consider the intertwined and complex reality of today’s workplace. In accordance with our hypothesis, we further found an effect of gender on intentions to pursue employment with organizations using identity conscious initiatives, with women indicating higher pursuit intentions.

An interesting, yet unexpected, finding of the present study is that both identity conscious statements led to equal pursuit intentions, even though the preferential treatment statement received substantially higher effectiveness ratings than the opportunity enhancement statement. The absence of a difference between opportunity enhancement and preferential selection initiatives with regard to pursuit intentions reveals an interesting practical insight: If organizations that are committed to increasing gender equality in top management choose opportunity enhancing initiatives over preferential selection initiatives only because they believe such initiatives will be perceived better by potential and/or existing employees, they are better off implementing preferential selection because it is more effective for helping them achieve the goal of increased female board participation while not additionally decreasing pursuit intentions among potential employees.

Another key finding of this study is the effect of a government-mandated gender quota for women in top management positions on individuals’ intentions to pursue employment with organizations voluntarily using preferential selection initiatives. After the necessary data adjustments, we found pursuit intentions toward the organization using a preferential treatment statement were higher in the Quota condition compared to the control condition. This is of special interest as we set out to determine whether a government-mandated gender quota could serve as a buffer that protects organizations that are using effective diversity initiatives from the negative outcome of decreased employer attractiveness. The presence or absence of information regarding a government-mandated quota, however, had no impact for the identity conscious opportunity enhancement statement. Recall that our government-mandated quota condition stated that organizations failing to achieve 30 % women in top management teams would face penalties and economic sanctions from the government. Given that opportunity enhancement initiatives are not perceived to be as effective as preferential treatment initiatives for increasing the proportion of women in upper management, it may be that participants believed that an even more prescriptive initiative was required to ensure compliance with the government mandate and thus to avoid economic penalties. This interpretation is consistent with our finding that the perceived effectiveness of opportunity enhancement was much closer to that of the identity blind initiative than to that of the preferential treatment initiative (cf. Table 3). Since we did not expect any effect of a government-mandated quota for the identity blind initiative, and as the opportunity enhancement initiative was only perceived as slightly more effective, it makes sense that we did not find an effect for the opportunity enhancement initiative.

Although gender mattered significantly with respect to the preferential treatment initiative, we only found a weak moderating effect of gender on opportunity enhancement. This finding may be explained in terms of the “would” aspect of fairness theory. It is possible that women did not believe the opportunity enhancement initiative would be enough for helping them to obtain a position/promotion (recall that opportunity enhancement was not perceived as effective as the preferential treatment initiative), thus they did not find it attractive, whereas men may have believed that this initiative would be enough for them to lose the position/promotion to a woman, so they did not find it attractive. These dynamics would be expected to narrow the difference between the genders.

We did not find a significant effect of attitudes toward diversity on pursuit intentions toward organizations using identity conscious diversity statements even after conducting additional analyses with the variable of attitudes toward diversity being dichotomized. A potential explanation may concern the scale we used to assess attitudes toward diversity (Hostager and De Meuse 2002), which is comprised of seventy words. Due to the way in which the scale was designed, it was not possible to ensure that participants read all seventy words thoroughly (e.g., mandatory choice) and it may be that due to the length of the survey in total, participants did not carefully respond to this scale, which would have been rather time-consuming. Therefore, limited accuracy of the scale in this study’s design may be responsible for the absence of an effect. Future research using alternative measures would thus be desirable.

Limitations and Avenues for Future Research

This experiment made an important contribution to the extant knowledge by tapping into the link between diversity initiatives and pursuit intentions—a main dimension of employer attractiveness (Highhouse et al. 2003). Given several limitations of our study, however, further research is necessary. This study was designed to shed light on the challenge of a diversity management paradox faced by organizations. We used counterfactual thinking from fairness theory (Folger and Cropanzano 1998, 2001) as a guiding framework for our hypotheses, but we did not measure perceived fairness of the initiatives themselves or the organizations using them. Future research should investigate fairness perceptions as a potential mediator to show that the positive effects of being perceived as an organization that values diversity are offset, at some point, by the negative effects of being perceived as an organization that violates societal expectations of fairness.

The three gender diversity statements used in this study were developed based on the actual diversity activities of the thirty major German companies. These companies’ typical corporate gender diversity activities like “a network for women," “coaching” or “workshops” for women might, however, be unaffordable for smaller companies with fewer employees and resources. Future research should investigate the effects of identity blind and identity conscious corporate gender diversity efforts that are relevant to a broader scope of organizations (i.e., including small- and medium-sized companies). Another limitation of this study is that participants’ awareness of the success (or lack thereof) of quota programs in countries near Germany may have impacted our results. Thus, future research that examines individuals’ knowledge about related quotas in other countries as well as their general attitude toward quotas could be especially useful.

While the results described above are mainly consistent with assumptions from the theoretical propositions of fairness theory (Folger and Cropanzano 1998, 2001), there may be additional processes contributing to our findings. For example, with regard to the impact of gender, the lower pursuit intentions of men relative to women for organizations using identity conscious initiatives may be influenced by neo-sexism. Men are higher than women in neo-sexism, which has been found to be negatively related to support for affirmative action policies designed to benefit women (Tougas et al. 1995). Such reactions are likely to decrease intentions of men to pursue employment with organizations using identity conscious initiatives. Another potential explanation for men’s lower pursuit intentions is that (as a group) they have a stronger preference for inequality among social groups, i.e., they are more social dominance oriented than women, and thus more likely to believe that initiatives designed to reduce gender inequality in management provide women with unearned advantages (Pratto et al. 1994). Such reactions are also likely to decrease intentions of men to pursue employment with organizations using identity conscious initiatives. The higher levels of neo-sexism and social dominance orientation in men may lead them to perceive lower person-organization fit between themselves and organizations using identity conscious initiatives (O’Reilly et al. 1991), as well as increase their perceptions of unfairness, thus making them less likely to pursue employment with such organizations. Taken together, although our work provides the first evidence of the predicted main effects, future research that looks more closely at the underlying mechanisms is desirable. In addition to attitudes such as neo-sexism and social dominance orientation (or lack thereof), being a member of underrepresented groups on dimensions other than gender (e.g., race or sexual orientation) or of multiple groups could also impact the model and are worthy of examination.

We also encourage researchers to explore the boundaries conditions of our model. For example, does the paradox of diversity initiatives apply to dimensions of diversity other than gender? While we expect that the prescriptiveness of diversity initiatives focused on race, age, etc., will influence their perceived effectiveness, the relationship between prescriptiveness of initiatives and pursuit intentions may be different, however, when the dimension of interest is something other than gender. If participants object to preferential treatment policies only because they believe such policies violate the merit principle, then the diversity dimension of interest should have no impact on pursuit intentions. Alternatively, if participants’ pursuit intentions are influenced by bias against (or in favor of) women, and they do not exhibit such bias toward other groups, the paradox may be mitigated (strengthened).

As described in the results section, the evaluation of our quota manipulation check revealed that a sizable percentage of participants did not answer the single item correctly. By excluding participants who failed the manipulation check, we chose a conservative way to ensure the internal validity of our analyses. Two potential explanations may account for participants’ failure to recognize the quota manipulation. First, some individuals might not have read the instructions thoroughly enough and hence may not have recognized the presence (or absence) of information on the government-mandated gender quota. Second, although there was no mention of a government-mandated quota in the No-government quota condition, those participants read about a gender quota that was voluntarily implemented by the organization in the identity conscious preferential treatment statement. Despite the fact that the manipulation check explicitly specifies a “government-mandated gender quota,” it is likely that this created confusion on the part of some participants.

We acknowledge that the characteristics of our sample in terms of gender, age, and education level are important to address. First, we tested for the effect of gender in our model and controlled for age (see Table 2). Second, with regard to the age and education level of the sample, the larger percentage of well-educated and older individuals may have influenced the perceived effectiveness of the initiatives because such individuals may have had a better understanding of the way in which they actually work, thus resulting in stronger effects for perceptions of perceived effectiveness. Thus, the paradox could be stronger for such individuals.

The present work reveals a major challenge for organizations that will be increasingly under pressure to achieve gender diversity at all organizational levels. It is crucial for researchers to provide a better understanding of how to overcome this challenge, thus we recommend follow-up studies to build on the growing body of knowledge regarding the impact of justifications and excuses (e.g., Bobocel and Farrell 1996; Kidder et al. 2004, Richard and Kirby 1998, Williamson et al. 2008). Furthermore, since legislative regulations regarding gender equality are likely to gain momentum around the globe, we call for research on their effects on the link between organizational diversity efforts and reactions of external stakeholders such as potential employees. It would be especially interesting to examine the effects of different types of penalties and incentives stemming from (non-)compliance with government-mandated gender quotas. For example, more substantial penalties and incentives for (non) compliance might strengthen the effect of quotas on intentions to pursue employment with organizations voluntarily using identity conscious initiatives. On the other hand, such penalties and incentives may have a negative impact for organizations using identity blind initiatives because these organizations may be perceived as more likely to violate government regulations, thus receiving substantial fines and negative publicity. This idea is consistent with research finding that individuals like to feel good about the organization for which they work (e.g., Greening and Turban 2000).

It would also be interesting to delve deeper into the mechanisms underlying men’s and women’s reactions to identity conscious initiatives. The impact of factors like self-interest, belief in discrimination, and sexism has been investigated in the past research (see Harrison et al. 2006 for a meta-analysis). The desire to maintain social harmony (Jepsen and Rodwell 2009) is one factor that has not yet been explored but may contribute to negative reactions to identity conscious initiatives, especially for women. Since women tend to be concerned with maintaining social harmony (Gilligan 1982), and identity conscious initiatives are controversial, this may contribute to women’s negative reactions to such initiatives.

Conclusion

The goal of achieving gender equality in organizations’ upper echelons makes sense from both a moral and a business perspective, which when taken together, explain the interest of many organizations in getting more women to the top of the management hierarchy. Identity conscious initiatives have the potential to achieve this goal (e.g., Kalev et al. 2006). However, this study shows that identity conscious initiatives also have the potential to decrease individuals’ intentions to pursue employment with an organization. Our finding that identity conscious initiatives are perceived as not only most effective but also as least appealing could lead to the conclusion that organizations should be cautious with their implementation of such initiatives. In accordance with Shore et al. (2009), however, it may be wise to take a more proactive and positive approach toward identity conscious initiatives. Using an experimental design, we showed that individuals who are told about a forthcoming government-mandated gender quota exhibited less resistance to organizations using gender diversity initiatives perceived and empirically shown to be effective. We believe that such legislation, in countries where it has been implemented, should be understood only as a vehicle for overcoming the systematic bias that has worked against the ascendance of women into top management positions. It is our hope that instead of being driven to achieve gender equality in the upper echelons of management by external factors such as government quotas, organizations ultimately will be driven by internal factors such as improved corporate governance and business performance. As more women attain top management positions, diversity statements espousing an organizational culture that values gender diversity will no longer be merely window-dressing; this should lead to positive stakeholder perceptions, and hence be most effective for overcoming the paradox of diversity management.