Regulatory Fit from Stereotype Threat: Enhancing Women’s Leadership Aspirations

TABLE OF CONTENTS

INTRODUCTION

1 THEORETICAL PART
1.1 Stereotype Threat
1.1.1 Stereotype Threat and Performance
1.1.1.1 Paradigms
1.1.1.2 Underlying Processes
1.1.1.3 Moderators
1.1.1.4 Coping Mechanisms and Interventions
1.1.2 Stereotype Threat and Other Outcome Variables
1.1.2.1 Disengagement and Disidentification
1.1.2.2 Domain Aspirations and Motivation or Avoidance?
1.2 Regulatory Fit
1.2.1 Regulatory Focus Theory
1.2.1.1 Paradigms
1.2.1.2 Consequences
1.2.1.3 Moderators
1.2.2 From Regulatory Focus to Regulatory Fit
1.2.2.1 Consequences
1.2.2.2 How and When Regulatory Fit Effects Emerge
1.3 Regulatory Fit from Stereotype Threat
1.3.1 Stereotype Threat and Regulatory Focus
1.3.2 Stereotype Threat and Regulatory Fit
1.3.2.1 Motivational Intensity and Performance
1.3.3 Summary of the Main Hypothesis and Experimental Overview

2 EMPIRICAL PART
2.1 Study 1
2.1.1 Method
2.1.1.1 Design and Participants
2.1.1.2 Procedure
2.1.2 Results and Discussion
2.1.2.1 Manipulation Check
2.1.2.2 Leader Role Motivation
2.1.2.3 Leadership Motivation – BIP
2.1.2.4 Test Performance and Effort
2.2 Study 2
2.2.1 Method
2.2.1.1 Design and Participants
2.2.1.2 Procedure
2.2.2 Results and Discussion
2.2.2.1 Manipulation Check
2.2.2.2 Test Performance and Effort
2.2.2.3 Leader Role Motivation
2.2.2.4 Leadership Motivation – BIP
2.3 Discussion Study 1 and Study 2
2.4 Study 3
2.4.1 Method
2.4.1.1 Design and Participants
2.4.1.2 Procedure
2.4.2 Results and Discussion
2.4.2.1 Manipulation Check
2.4.2.2 Leadership Motivation
2.4.2.3 Impression Related Concerns and Pressure Related Feelings
2.4.2.4 Test Performance and Effort

3 GENERAL DISCUSSION
3.1 Implications
3.2 Limitations
3.3 Future Directions
3.4 Conclusions

4 REFERENCES

5 APPENDIX A: INDEPENDENT VARIABLES

6 APPENDIX B: DEPENDENT VARIABLES AND MEDIATOR

7 APPENDIX C: COMPLETE QUESTIONNAIRES

INTRODUCTION

“We women are not made for governing and if we are good women, we must dislike these masculine occupations; but there are times which force one to take interest in them mal gré bon gré, and I do, of course, intensely.”

- Victoria, British monarch (Queen of Great Britain and Ireland, 1837-1901),

Letter to Leopold I, King of the Belgians, Feb. 3, 1852

“No woman in my time will be Prime Minister or Chancellor or Foreign Secretary - not the top jobs. Anyway I wouldn’t want to be Prime Minister. You have to give yourself 100%.”

- Margaret Thatcher, British Conservative politician (British Prime Minister, 1979-1990),

Interview in Sunday Telegraph, London, Oct. 26, 1969

Two women, primary political leaders of their country, more than a century apart and yet their statements have something in common. Both note that a leadership post is a masculine occupation not appealing to women of their time, nonetheless both are genuinely and intensely or 100% committed to their leadership position. Of course the examples given above are grand exceptions with regard to women’s career paths. However, even though stereotypically masculine occupations are male dominated this does not always mean that they are solely aspired by men (e.g., European Commission, 2007). But what makes women withdraw from stereotypically masculine occupations and why do some women despite the masculine stereotype still aspire those careers?

The basic root of the research proposed in this thesis is the differential distribution of men and women into occupations also referred to as the sex segregation of occupations. Different explanations for the sex segregation of occupations have been postulated. Overall, explanations relating to external versus internal causes are distinguished. Both are further classified into direct and indirect influences (Stangor & Sechrist, 1998). External influences are those, which come from the outside of the individual such as discrimination, sexism, prejudice, and self- fulfilling prophecies. In the present thesis I will focus on internal influences, which are those within the person such as ability, performance expectancies, and preferences. Specifically, the research proposed here concerns the influence of gender stereotypes on women’s domain preferences and aspirations. In particular, the effects of stereotype threat on women’s leadership aspirations will be examined.

Stereotype threat theory (Steele, 1997) poses that, when a negative stereotype about a group in a certain performance domain exists, members of this group will be threatened by this stereotype, which will in turn lead to a negative impact on their performance and aspirations (i.e., performance and aspirations will decrease; e.g., Davies, Spencer, Quinn, & Gerhardstein, 2002). To date a huge body of research has addressed and confirmed the assumptions of stereotype threat theory concerning performance as an outcome variable (for reviews see Maas & Cadinu, 2003; Smith, 2004; Steele, Spencer, & Aronson, 2002). However, research addressing the effects of stereotype threat on aspirations or motivation is still very rare. In addition, those studies that have shown effects of stereotype threat on motivation do not present a consistent picture whether stereotype threat will decrease or, in contrast to effects on performance, increase motivation (e.g., Davies, Spencer, & Steele, 2005; Nussbaum & Steele, 2007).

The present research aims to address the issue when stereotype threat will have a decreasing or an increasing effect on women’s leadership aspirations or motivation. Furthermore, the question whether stereotype threat effects on motivation and performance will differ is explored. It is proposed that stereotype threat can enhance or decrease motivation depending on a person’s motivational orientation. For this purpose it is referred to regulatory fit theory (Higgins, 2000).

Regulatory fit theory (Higgins, 2000) states that a regulatory fit occurs, when there’s a fit between a person’s motivational orientation, i.e., a person’s regulatory orientation, and his or her goal. Further, regulatory focus theory (Higgins, 1997; 1998) proposes that a person’s regulatory orientation can either be predominantly promotion- or prevention-focused. According to regulatory fit theory (Higgins, 2000) a regulatory fit will occur, if a person’s motivational orientation and his or her goal share the same regulatory orientation. A regulatory fit is assumed to lead to an increase in a person’s motivation and/or a feeling of rightness. The feeling of rightness is proposed as a subjective kind of experience that occurs through a regulatory fit and will lead to an increase in a person’s reactions and evaluations of stimuli during that experience, whatever these reactions or evaluations happen to be (e.g., Higgins, Idson, Freitas, Spiegel, and Molden, 2003; Schwarz, 2006).

The current research proposes that a regulatory fit can occur from stereotype threat. In particular, it is put forward that a prevention focus is associated with stereotype threat conditions whereas a promotion focus is associated with no stereotype threat conditions (i.e., when the negative stereotype is removed and relatively positive expectations are present). According to regulatory focus theory (Higgins, 1997; 1998) a person with a prevention focus is concerned with losses and failures and particularly sensitive to the presence or absence of negative outcomes. At the same time the possibility of a failure and negative outcomes activate prevention goals (e.g., the approach of nonlosses and the avoidance of losses). A person with a promotion focus, on the other hand, is concerned with gains and successes and particularly sensitive to the presence or absence of positive outcomes. Respectively, the possibility of a success and positive outcomes activate promotion goals (e.g., the approach of gains and the avoidance of nongains). In stereotype threat conditions the absence or presence of the possibility of a failure and negative outcomes is made salient by the presence of a negative stereotype about one’s group’s abilities. Therefore, stereotype threat conditions should instigate prevention goals. In no stereotype threat conditions, on the other hand, the negative stereotype about one’s group is removed and relatively positive expectancies are present in the situation, making the absence or presence of the possibility of a success and positive outcomes salient. Hence, no stereotype threat conditions should induce promotion goals. These assumptions have been supported by previous research showing that stereotype threat conditions are associated with a prevention focus and no stereotype threat conditions are associated with a promotion focus (Seibt & Förster, 2004)^[1].

Given that a regulatory fit occurs when a person’s motivational orientation and his or her goal share the same regulatory orientation, a regulatory fit or nonfit^[2] should occur for predominantly prevention- or promotion-focused individuals under stereotype threat or no stereotype threat conditions. Specifically, a regulatory fit is assumed to occur under stereotype threat conditions for prevention-focused individuals and under no stereotype threat conditions for promotion-focused individuals. Contrarily, a regulatory nonfit is assumed to occur under no stereotype threat conditions for prevention-focused individuals and under stereotype threat conditions for promotion-focused individuals.

The Regulatory Fit from Stereotype Threat Assumption will be outlined in further detail in the current work and experimentally tested in three studies. I will start outlining stereotype threat effects on performance to explain the concept in further detail and present proposed processes and moderators underlying these effects. Then I will introduce other outcome variables that have been shown to be affected by stereotype threat and finally elaborate on research showing stereotype threat effects on motivation. Here I will point out the inconsistencies of stereotype threat effects on motivation in the research to date and propose the match of a person’s motivational orientation and his or her goal as a possible explanation. Following, I will introduce regulatory fit and regulatory focus theory and give an extensive overview of the literature on regulatory fit as it relates to the current studies. Finally, I will point out how stereotype threat can lead to a regulatory fit and test this assumption experimentally.

Three experiments will test the Regulatory Fit from Stereotype Threat Assumption with women’s leadership aspirations as the outcome variable. In Studies 1 and 2 the effect of regulatory fit from stereotype threat on motivation will be tested. Study 3 examines whether regulatory fit from stereotype threat will lead to an increase in stimuli persuasiveness due to an effect of feeling right. Furthermore, in all three studies the possible differential effects on motivation and performance and their interaction will be explored.

In sum, the current research question is based on the sex segregation of occupations. Specifically, the questions addressed here concern the link between gender stereotypes and women’s domain aspirations or motivation. In particular, the present research investigates when stereotype threat effects will lead to an increase or decrease in leadership motivation. To answer this question it is drawn from regulatory fit theory. It is put forward that stereotype threat conditions can lead to a regulatory fit and thus enhance motivation and/or lead to an increase stimuli persuasiveness.

1 THEORETICAL PART

I want to begin by outlining the interrelations between the sex segregation of occupations, gender stereotypes, and occupational preferences. The sex segregation of occupations exists on a horizontal level (between occupations) and a vertical level (within occupations). The horizontal sex segregation is shown by women being the majority among service workers, clerical and sales workers and in the care, nursing and education professions whereas occupations such as technical workers, production workers and transport workers are heavily male-dominated (e.g., European Industrial Relations Observatory [EIRO], 2000). The vertical sex segregation is marked by the fact that upper level positions are mostly occupied by men and women are less likely found in managerial positions (EIRO, 2000). In fact, the percentage of women in top- management positions has been found to be not more than 7% in Germany (Hoppenstedt, 2005).

It has been argued that the occupational sex segregation corresponds with and results in gender stereotypes (Eagly, 1987; Eagly, Wood, & Diekman, 2000). Empirical support for this argument can be found in various studies. Ceijka and Eagly (1999), for example, found that the gender ratio in several occupations correlated significantly with the gender-stereotypic images of those occupations. Further, Schein, Mueller, Lituchy, and Liu (1996) found in several studies that successful middle managers were perceived to be more similar to men in general than to women in general. The correspondence between the occupational sex segregation and gender stereotypes was already shown among young boys and girls, who believed that certain jobs should be performed by a woman whereas others should be performed by a man.

These beliefs matched the pattern of occupational sex segregation (Miller & Budd, 1999; Smithers & Zientek, 1992).

Furthermore, Miller and Hayward (2006) showed that the occupational sex segregation strongly correlated with young boys’ and girls’ sex-role stereotyping as well as their occupational preferences. Additional support for the influence of the occupational sex segregation on occupational preferences via gender stereotypes was shown in studies by Jacobs and Eccles (2000) where boys chose stereotypically masculine jobs whereas girls choose stereotypically feminine jobs. Moreover, data from the Equal Opportunity Commission (EOC, 2001) provides evidence that most individuals prefer jobs that are viewed as gender-congruent (i.e., correspond to the occupational sex segregation and gender stereotypes of one’s own sex).

Taken together, there is clear evidence for a link between the occupational sex segregation, gender stereotypes, and domain aspirations. The link between gender stereotypes and domain aspirations is multi-faceted and has been addressed by many different authors (e.g., Eagly, 1987; Eccles, 1994; Konrad, Ritchie, Lieb, & Corrigal, 2000). For example, in Eagly’s social role theory (Eagly, 1987; Eagly, Wood, & Diekman, 2000) it is argued that gender stereotypes are the result of the segregation of women and men into different social roles such as occupations. Those gender stereotypes in turn elicit sex-differentiated behavior and choices such as domain aspirations. Konrad et al. provided an extensive review with a meta-analysis on gender-differences in job attribute preferences, which showed that men and women generally preferred job attributes that are consistent with gender roles and stereotypes. Further, a comprehensive model of achievement-related choices was offered by Eccles (1994). Eccles proposes that achievement behaviors such as choice of activity (e.g., domain selection), persistence, and actual performance are a result of the interaction of a number of variables that can be grouped into two basic components, a component concerning social factors and a component concerning psychological factors. In this model gender stereotypes are seen as one among many interacting variables that influence achievement related behavior such as domain aspirations.

Although, it is acknowledged here that the influence of gender stereotypes on behavior operates within a wide array of variables, such as socialization, social structures and roles, this thesis will address how gender stereotypes can influence behavior in a specific situation namely through stereotype threat (Steele, 1997). Hence, the focus in this thesis is on how women’s career aspirations can be influenced by stereotype threat.

1.1 Stereotype Threat

Stereotype threat theory has proposed valuable insights and created a vast body of research on how negative stereotypes can impede behavior of a stereotyped group. Stereotype threat has been characterized as a situational threat that occurs when negative stereotypes about one’s group are thought to apply. As a result one might perceive to be judged or treated in terms of the stereotype or might inadvertently confirm it (cf. Steele et al., 2002). Stereotype threat has been studied for more than a decade now, and different authors have offered different definitions of the concept encompassing the one offered above in different ways. Those definitions mostly differ in their statement by whom one might be judged or treated in terms of the stereotype (i.e., the self, out-group others, or in-group others) and who might be judged or treated in terms of the stereotype or might be at risk to confirm it (i.e., the self or one’s group; for a review see Shapiro & Neuberg, 2007).

Since all of the above definitions are applicable to the theoretical model that I will put forward at a later point as well as to the later proposed studies I want to slightly alter a definition of stereotype threat proposed by Wheeler and Petty (2001, p. 804) to take in the full range of proposed definitions: Stereotype threat is “defined as the pressure an individual faces when he or she may be at risk of confirming negative, self-relevant group stereotypes [in others’ eyes, or in one’s own]^[3]”.

Further definitions of stereotype threat include its outcome variables such as stating that stereotype threat is “the apprehension people feel when performing in a domain in which their group is stereotyped to lack ability” (Aronson & Inzlicht, 2004, p. 830) or “a negative stereotype can elicit a disruptive state […] that undermines performance and aspirations […] – a situational predicament termed stereotype threat” (Davies, et al., 2002, p. 1616). There is a massive body of research evidence showing that stereotype threat impairs the performance of members of the stereotyped group in the stereotyped domain (for reviews see Maas & Cadinu, 2003; Smith, 2004; Steele et al., 2002). Far less empirical evidence exists on the impact of stereotype threat on aspirations or interest and motivation for a domain (i.e., Davies, et al., 2002; Davies et al., 2005; Nussbaum & Steele, 2007; Rosenthal & Crisp, 2006; Smith, Sansone, & White, 2007). Therefore, in order to give an overview of the phenomena of stereotype threat including its proposed mechanisms, moderators, and interventions I will first outline the effects of stereotype threat on performance. However, of particular interest for this thesis are the effects of stereotype threat on motivation and related outcome variables. Hence, these will be discussed in greater detail in the section thereafter.

1.1.1 Stereotype Threat and Performance

A meanwhile classic example of stereotype threat effects on the academic performance of women is a study by Spencer, Steele, and Quinn (1999). They asked male and female participants to complete a test measuring mathematical ability. Half of the participants were told that the test has shown no gender differences in the past (no stereotype threat condition), presumably rendering the stereotype irrelevant. The other half of the participants who were in the stereotype threat condition were given no information about gender differences on the test. The assumption was that in the stereotype threat condition the negative stereotype about women’s math performance would become relevant simply by presenting a test in the stereotyped domain. The results showed that women and men performed equally well in the no stereotype threat condition but that women in the stereotype threat condition underperformed in comparison to men and in comparison to women in the no stereotype threat condition (see Figure 1).

Figure 1. Performance on a math test as a function of gender and test characterization (adapted from Spencer et al., 1999)

Abbildung in dieser Leseprobe nicht enthalten

Further evidence for the detrimental effects of stereotype threat on performance relates, for example, to African-American’s academic performance (e.g., Steele & Aronson, 1995), low SES (Socio-Economic Status) participants’ verbal performance (Croizet & Claire, 1998), women’s negotiation abilities (Kray, Thompson, & Galinsky, 2001), men’s performance on an affective decision task (Leyens, Désert, Croizet, & Darcis, 2000), decreased memory performance of elderly (Levy, 1996), or Caucasians’ athletic performance when compared to African- Americans’ (Stone, Lynch, Sjomeling, & Darley, 1999).

1.1.1.1 Paradigms

The general paradigm for stereotype threat effects to occur is that the stereotype has to be relevant in the situation. Many manipulations have been used by researchers studying stereotype threat. However, most manipulations can be grouped into three categories.

First, stereotype threat can be manipulated by activating group differences in performance either implicitly or explicitly. The implicit activation of group differences in performance is accomplished by giving participants a test of performance for the stereotyped domain without mentioning any group differences in performance for the stereotype threat condition assuming that the stereotype would be activated merely by the test itself. In the no stereotype threat condition the stereotype is rendered irrelevant by stating that no group differences in performance have occurred on this test in the past (e.g., Spencer et al., 1999). The explicit activation of group differences is either indirect by stating that group differences (e.g., gender differences) in performance have occurred on the test in the past or blatant and direct by stating that the stereotyped group (e.g., women) was found to perform worse than the non-stereotyped group (e.g., men). In the no stereotype threat conditions the stereotype is in both cases again removed by stating that no group differences in performance have occurred on this test in the past (e.g., Spencer et al., 1999).

Second, stereotype threat can be manipulated by varying the test diagnosticity for the stereotyped domain. The stereotype is made relevant by depicting the test as diagnostic of a domain for which the stereotype applies or, when removing the stereotype threat, depicting the test as diagnostic of a domain for which the stereotype does not apply (e.g., in a study with African-Americans by stating that the proposed test is diagnostic of verbal vs. problem solving abilities; Steele & Aronson, 1995).

Third, stereotype threat can be manipulated in a subtle manner usually by activating the category membership of the negatively stereotyped group, e.g. by having participants indicate their gender or race prior to the test, by exposing participants to gender stereotypic TV commercials, or putting them in a solo-status condition among members of a group which is not negatively stereotyped for the performance domain (e.g., Davies et al., 2002; Steele & Aronson, 1995; Sekaquaptewa & Thompson, 2003).

1.1.1.2 Underlying Processes

Why do these detrimental effects of negative stereotypes on people’s behavior occur? Much research has been conducted to determine the processes that drive stereotype threat effects on performance. Up to date there is no unitary answer or as Osborne (2001, p. 296) stated stereotype threat “is an effect without a clearly explicated mechanism”.

According to Steele and Aronson (1995) minority members are afraid to confirm the stereotype which creates a tension due to the preoccupation with the inadvertent confirmation of the stereotype. Thus, two mechanisms are suggested here to go hand in hand: anxiety and disruptive thoughts. Indeed both have been among the most frequently studied processes to account for stereotype threat effects.

Anxiety or arousal. Many researchers have attempted to show that the detrimental effects of stereotype threat on performance are due to an increase in anxiety or arousal. The results however provide a mixed picture. Often no effects at all could be established (e.g., Steele & Aronson, 1995; Aronson, Lustina, Good, Keough, Steele, & Brown, 1999; Stone et al., 1999). Other researchers found marginal or full evidence for partial mediation (e.g., Osborne, 2001; Spencer et al., 1999). In addition to most of the anxiety measures which were assessed on self-report scales Blascovich, Spencer, Quinn, and Steele (2001) found support for anxiety as a plausible mechanism by showing an increase in blood pressure under stereotype threat.

Disruptive thoughts. Further, disruptive thoughts such as the number of negative thoughts, self-doubts, low performance expectancies, or task confidence were tested as a mediational process for stereotype threat effects on performance. It is argued that stereotyped group members’ cognitive resources will be occupied by these thoughts limiting further resources for performance. Again across different studies no clear picture emerges when comparing the results. Sometimes no effects of mediation were found as for performance expectancies (e.g., Keller, 2002; Kray et al., 2001; Sekaquaptewa & Thompson, 2003) and self-efficacy (e.g., Oswald & Harvey, 2001). Others found effects of stereotype threat on the mediator (performance expectancies: e.g., Stangor, Carr, & Kiang, 1998; self-doubt: e.g., Steele & Aronson, 1995) but no effect of the mediator on performance. Then some researchers found evidence of partial mediation (performance expectancies: Cadinu, Maas, Frigerio, Impagliazzo, & Latinotti, 2003; self-doubt: Stone, 2002).

Other mediators that are related to anxiety and/or intrusive thoughts such as increased evaluation apprehension and lowered working memory capacity have also been studied. In particular, no evidence for evaluation apprehension as a mediator could be obtained so far (e.g., Spencer et al., 1999; O’Brien & Crandall, 2003). However, results concerning working memory capacity have been proven to be most promising when compared to other investigated processes. Schmader and Johns (2003) found evidence of stereotype threat effects on working memory and established working memory as a significant mediator of stereotype threat effects on women’s math performance.

Overall, the processes underlying stereotype threat effects cold not be fully established yet. Except for working memory capacity none of the mediators investigated could show a full mediation. Smith (2004) proposes that the fact that none of the processes could be proven so far when studied individually would imply that different processes cumulatively cause the effects. As a result Smith put forward a multiple-mediator model of stereotype threat effects. However, Shapiro and Neuberg (2007) suggest that the mixed picture of results on stereotype threat processes is due to the different definitions and operationalizations of stereotype threat and offer a multi-threat framework proposing different kinds of stereotype threat that differ in their underlying processes. Nevertheless, the multiple-mediator model and the multi-threat framework still remain to be tested empirically.

Even though the empirical evidence concerning the underlying mechanisms of stereotype threat remains somewhat puzzling, some parts are clear and consistent across all studies. A negative stereotype is salient in the situation and the members of the stereotyped group do not want to confirm this stereotype. Whether they fear or are anxious to confirm it, have thoughts and worries about confirming it, put too much effort into disconfirming the stereotype, other processes, all or some of the above lead to performance decrements still remains to be shown.

1.1.1.3 Moderators

The mostly acknowledged moderators in the stereotype threat literature were already identified by Steele (1997) stating that one of the general features of stereotype threat is “an identified-with setting” (p. 617) implying that for stereotype threat to occur it is necessary that an individual identifies a) with the domain and b) with the stereotyped group. Steele (1997) further states that “the work of dispelling stereotype threat through performance probably increases with the difficulty of work in the domain” (p. 618). Accordingly, task difficulty is suggested as another major moderator variable for stereotype threat effects on performance.

Domain identification. Domain identification is often seen as a prerequisite for stereotype threat effects to occur. As a result, many researchers have limited their samples to highly domain identified individuals (e.g., Marx & Roman, 2002; Schmader, 2002; Schmader & Johns, 2003; Spencer et al., 1999). It is argued that an individual has to care about the domain in order for stereotype threat effects to become self-relevant. A few studies have indeed successfully tested this assumption by comparing high and low domain-identified individuals (Aronson et al., 1999; Cadinu et al., 2003; Stone et al., 1999). However, according to the definition of stereotype threat offered above an individual might not only fear to confirm the stereotype in his or her own eyes but also in the eyes of others without seeing the domain as relevant to his or her self-concept . Indeed (Marx, Stapel, & Muller, 2005) have found that stereotype threat effects were mediated by impression related concerns (sample item: “I am concerned what other people think of me”) for high and low identified individuals. In addition, stereotype threat effects have also, for example, been found for the domain of social sensitivity or emotional intelligence in men (Koenig & Eagly, 2005; Leyens, et al., 2000) a group that generally is not likely to be highly identified with those domains. However, other research in which the perceived relevance of the domain was manipulated suggests that men are more vulnerable to stereotype threat in the domain of emotional intelligence when their domain identification is high (Görzig, Keller, & Bless, 2005). In sum, these results suggest that domain identification is sometimes, but not always, a prerequisite for stereotype threat effects to occur. Future research should address the question under which conditions domain identification is a necessity for stereotype threat effects and in which conditions stereotype threat effects can be found for both high and low identified individuals.

Group identification. Identification with the stereotyped group is also often seen as a presumed necessity for stereotype threat effects, but has not been studied as frequently as domain identification. The reasoning is the same as the one for domain identification. An individual has to identify with his or her group in order for the negative stereotype about the group to become self-relevant. Again, this argument can only hold when the individual fears to confirm his or her group’s stereotype in his or her own eyes. When an individual could be seen as a member of his or her group by others and thus fears to confirm the group stereotype in others’ eyes, group identification fails to be a necessary condition in order for stereotype threat effects to appear. Accordingly, the empirical evidence again presents a mixed picture. For example, Schmader (2002) showed that gender group identification is a moderator for stereotype threat effects on women’s math performance whereas Cadinu et al. (2003) failed to show the very same effect.

Task difficulty. The difficulty of the task is another potential moderator of stereotype threat effects. It is argued that performance-impeding factors triggered by stereotype threat such as anxiety and/or intrusive thoughts are particularly relevant under conditions of high task difficulty. The argument put forward concerning anxiety is in line with research showing that enhanced arousal increases performance on moderately difficult items but decreases performance on very difficult items (e.g., Hill & Wigfield, 1984; Yerkes & Dodson, 1908). As mentioned in the section on mediators, intrusive thoughts occupy a person’s cognitive resources, which should become particularly debilitating under difficult performance tasks. Supporting these arguments, Spencer et al. (1999) found that stereotype threat effects on women’s math performance only occurred under conditions of high task difficulty. Nonetheless, the assumed mechanisms (i.e., anxiety and intrusive thoughts) have not yet proven to be present under all conditions of stereotype threat. Thus, it cannot be assumed that stereotype threat effects are always enhanced under high test difficulty. In fact, it has been demonstrated in a recent study that stereotype threat effects can also occur on relatively easy test items (Wade, 2007).

Although the moderators outlined above have been agreed upon by many scholars, the theoretical background and empirical evidence seem not as clear. Further research should be considered to identify the conditions under which certain variables do or do not moderate stereotype threat effects. Furthermore, it is relevant to differentiate between moderators for different types of domains and groups as well as the nature of the task.

1.1.1.4 Coping Mechanisms and Interventions

Members of stereotyped groups have developed different types of coping behaviors when faced with stereotype threat. They have been shown to engage in self- handicapping by providing external explanations (e.g., lack of sleep) or engaging in behavior which would link a possible failure to other sources than their ability level (e.g., lack of training; e.g., Keller, 2002; Steele & Aronson, 1995; Stone, 2002). Further, they have shown to disengage their self-worth from performance in the domain (Stone et al., 1999). As an attempt not to be seen as typical members of the stereotyped group, individuals have been found to engage in counter-stereotypic behavior (e.g., African-Americans reporting to dislike jazz, hip-hop, or basketball; Steele & Aronson, 1995) or to disidentify from characteristics of the stereotyped group (e.g., women disidentified from typical feminine characteristics such as emotional, flirtatious, and planning to have children; Pronin, Steele, and Ross, 2004). Furthermore, individuals have been found to avoid or disidentify from the stereotyped domain (e.g., Davies et al., 2002; Davies et al., 2005).

Some attempts to externally eliminate stereotype threat effects go in hand with the manipulations used for the no stereotype threat conditions. For example, stereotype threat can be eliminated by rendering the stereotype irrelevant (e.g., Spencer et al., 1999). Others have shown that shaping individuals’ theories of intelligence as malleable (Aronson, Fried, & Good, 2002) or informing participants about stereotype threat effects (Johns, Schmader, & Martens, 2005) offers protection against stereotype threat effects on performance tests. In addition, various researchers showed that presenting positive models of members of the stereotyped group, who have succeeded in the stereotyped task or domain, can also buffer from the detrimental effects of stereotype threat on performance (e.g., Blanton, Crocker, & Miller, 2000; Blanton, Christie, & Dye, 2002; Marx & Roman, 2002; Marx et al., 2005; McIntyre, Paulson, & Lord, 2003).

1.1.2 Stereotype Threat and Other Outcome Variables

Although most studies on stereotype threat have investigated performance as the main dependent variable, a smaller number of studies have also shown that stereotype threat can influence other outcome variables some of which have already been reported above as coping mechanisms. Further outcome variables of stereotype threat besides performance are, for example, blood pressure (Blascovich, et al., 2001), anxiety and frustration (Marx & Stapel, 2006a), domain-specific self-efficacy (Aronson & Inzlicht, 2004), disidentification from characteristics associated with the stereotyped group (Pronin, et al., 2004; Steele & Aronson, 1995), but most importantly for the proposed framework disengagement or disidentification from the negatively stereotyped domain, aspirations for stereotype relevant careers, and motivation or avoidance of a task for which one’s own group is negatively stereotyped (e.g., Davies et al., 2002; Davies et al., 2005; Major & Schmader, 1998; Major, Spencer, Schmader, Wolfe, & Crocker, 1998; Osborne, 1995; Nussbaum & Steele, 2007). Nonetheless, compared to performance measures, these outcome variables have been largely understudied.

1.1.2.1 Disengagement and Disidentification

It has been theorized that members of stereotyped groups will disengage their self-esteem from self-evaluative feedback in the face of stereotype threat (e.g., Major et al., 1998; Steele et al., 2002). Disengagement can be temporary and a situationally specific response to stereotype threat or it can be chronic, i.e., when exposure to threat has been persistent. The chronic form of disengagement has also been referred to as disidentification (e.g., Steele & Spencer, 1992; Steele, 1997; Steele et al. 2002). Major & Schmader (1998) put forward two processes by which disengagement occurs: a) discounting (i.e., the validity or diagnosticity of feedback in the domain is questioned as a true indicator of one’s behavioral outcome), b) devaluing (i.e., the importance of the domain for one’s self is questioned). Both could be shown to be related to general disengagement (i.e., domain performance is seen as unrelated to one’s sense of self). However, discounting is mostly associated with situational disengagement whereas devaluing is associated with chronic disengagement or disidentification.

In line with the theorizing mentioned above, Major et al. (1998) provided correlational data showing that African Americans, who are negatively stereotyped in the intellectual domain, discounted and disengaged from the intellectual domain more than European Americans. Disengagement was further correlated with (higher) global self-esteem, underlining the assumption that when disengaging, one’s sense of self is not affected by the negative stereotype. In addition, it was shown that African Americans had higher self-esteem than European Americans when doing poorly in school and their self-esteem was equivalent to European Americans’ when doing well in school. Further, Osborne (1995) found a weakening of the correlation between self- esteem and academic outcomes of African Americans from 8th to 10th grade suggesting that African Americans disengage more from the academic domain the longer they are exposed to it.

Results of two experimental studies support these findings. In the first study, African-Americans’ self-esteem and performance was shown to be less affected by success and failure feedback than European Americans’. At the same time African Americans attributed their performance outcomes to a biased test (Major et al., 1998). In the second study, following a race-prime, African Americans showed higher self- esteem when receiving failure feedback than European Americans and African Americans who did not receive a race-prime. Furthermore, African Americans who scored high in a chronic disengagement measure also showed higher self-esteem after failure feedback than African Americans who scored low on the chronic disengagement measure (Major et al., 1998).

These experimental data provide further evidence that when a negative stereotype about one’s group’s performance exists an individual tends to disengage from the domain by devaluing the test outcome as biased. This disengagement shows to protect the individual’s self-esteem and performance from performance feedback. Thus, one could draw the conclusion that disengagement following stereotype threat can further protect from stereotype threat. Empirical support for this argument can be seen in studies that show that only those high in domain identification are affected by stereotype threat. Individuals who did not identify with the domain, i.e., are chronically disengaged from the domain, did not show any performance decrements under stereotype threat. However, their performance was shown to be quite low to begin with (Aronson et al., 1999; Cadinu et al., 2003; Stone et al., 1999). As Steele et al. (2002) argue both domain identified and disengaged individuals may suffer from low performance under stereotype threat for different reasons. The performance of the highly domain identified might suffer from stereotype threat due to frustration and experienced pressure whereas the performance of the disengaged might decrease due to a decline in their motivation. This reasoning implies that stereotype threat would lead to a decrease in motivation.

In fact it was shown in a study by Davies et al. (2005) that stereotype threat can decrease women’s leadership aspirations. This study is of particular relevance, because it has investigated stereotype threat effects on women’s leadership aspirations and thus has partially inspired the methodology for the studies presented later in this thesis. Therefore, I will briefly introduce the study by Davies et al. in the following.

Davies et al. (2005) exposed male and female participants to either stereotypic commercials (stereotype threat) or neutral commercials (no stereotype threat). Then they were asked to take part in an alleged role play for which they were told they could either take over a leader role or a problem solver role. Specifically, they were told that “both the problem solver and the leader will be given a written description of a series of complex problems to be solved. The leader, however, will also be supplied with the answers to those problems. It’s the leader’s job to guide the problem solvers to the solution without explicitly telling them the answers.” (Davies et al., p. 279). Further, the participants were told that for the role play they would join a group of other participants down the hall to avoid effects of the particular group composition. They were then asked to indicate their interest for each role on a scale from 1 (no interest) to 7 (strong interest). No significant effects were found among the male participants. However, for female participants it was shown that when exposed to gender-stereotypic commercials the problem solver role was preferred over the leader role whereas no difference was found in the interest for the problem solver and the leader role for female participants in the neutral commercials condition. Further, female participants in the neutral commercial condition had more interest in the leader role than those in the gender-stereotypic commercials condition. Conversely, female participants in the gender-stereotypic condition expressed more interest for the problem solver role than female participants in the neutral commercials condition (see Figure 2).

Figure 2. Women’s role motivation as a function of stereotype threat and role type (adapted from Davies et al., 2005).

Abbildung in dieser Leseprobe nicht enthalten

Davies et al.’s (2005) study shows that stereotype threat can have an impeding effect on motivation. However, the overall empirical evidence concerning stereotype threat effects on motivation is not as clear. In fact motivation and aspirations have sometimes, but not always, been shown to decrease from stereotype threat.

1.1.2.2 Domain Aspirations and Motivation or Avoidance?

Davies et al. (2002), for example, have shown that stereotype threat can lead to a decrease in motivation and an avoidance of the domain among highly domain identified individuals. Nussbaum and Steele (2007), on the other hand, found that stereotype threat caused a situational disengagement which helped to maintain domain identification and increased motivation.

Davies et al. (2002) exposed highly math identified male and female participants to commercials which were either gender-stereotypic (stereotype threat condition) or neutral (no stereotype threat condition). Then they gave them an aptitude test containing the same number of verbal and math items. Nussbaum and Steele (2007) gave European and African American participants a test of anagrams which were allegedly either diagnostic (stereotype threat condition) or non-diagnostic (no stereotype threat condition) of academic ability. Then they gave them a second test containing the same number of anagram and verbal analogy items. In both studies the numbers of chosen or attempted items for the stereotypic domain (i.e., math or anagram items) were measured. The results showed that female participants in Davies et al.’s study choose fewer math items under stereotype threat whereas African Americans in Nussbaum’s and Steele’s study choose more anagram tasks under stereotype threat compared to all other conditions (see Figure 3 and 4).

Figure 3. Math items attempted as a function of stereotype threat (adapted from Davies, et al., 2002)

Abbildung in dieser Leseprobe nicht enthalten

Figure 4. Anagram items selected as a function of stereotype threat (adapted from Nussbaum & Steele, 2007)

Abbildung in dieser Leseprobe nicht enthalten

Furthermore, results in Nussbaum’s and Steele’s (2007) study were mediated by situational disengagement but not by domain devaluation, suggesting that as in Davies et al.’s (2002) study participants did not differ in their chronic domain identification.

Despite their similarities in design and procedure, these studies show diverging evidence for stereotype threat effects on motivation or avoidance for the stereotyped domain. Taken together the results suggest that there are conditions under which participants’ motivation for the stereotyped domain decreases under stereotype threat and are in line with results of stereotype threat effects on performance (e.g., Davies et al., 2002). It is also shown that there are other conditions under which participants’ motivation for the stereotyped domain increases and participants might feel motivated to disconfirm the stereotype present under stereotype threat (e.g., Nussbaum & Steele, 2007). Thus, the remaining question is when motivation might increase or decrease as a function of stereotype threat.

Although not explicitly addressing the when question, Smith et al. (2007) provided empirical evidence pointing to the joint effect of performance goals and achievement motivation (cf. Elliot & Church, 1997) as a potential moderator of stereotype threat effects on task interest. Smith et al. induced stereotype threat by explicitly activating group differences on a computer task for female participants who were high or low in achievement motivation. Then performance avoidance vs. performance approach goals (cf. Elliot & Church, 1997) were manipulated via the following instructions: “Some students stand out because they do quite poorly (vs. well) on the [task]. For instance, if you do worse (vs. better) on the [task] than a majority of students, you will demonstrate that you have poor (vs. good) computing aptitude.” (Smith et al., 2007, p. 103). As a dependent measure, participants’ interest in the task was assessed via three items (e.g., “This task is fun to do”). The researchers found that under stereotype threat women who were high in achievement motivation showed higher computer task interest under a performance approach goal than a performance avoidance goal whereas women who were low in achievement motivation showed higher computer task interest under a performance avoidance goal than a performance approach goal (see Figure 5).

Figure 5. Women’s interest in a computer task as a function of achievement motivation and performance goal under stereotype threat (adapted from Smith et al., 2007).

Abbildung in dieser Leseprobe nicht enthalten

I will discuss the results of Smith et al.’s (2007) study in the following in an attempt to answer the question when stereotype threat can lead to increased or to decreased motivation. Someone high in achievement motivation as described by Smith et al. is “someone who aspires to accomplish difficult tasks, maintain high standards and is […] willing to put forth effort to attain excellence” (p. 101). Someone low in achievement motivation on the other hand wants to “avoid demonstrating incompetence” (p. 100). Not surprisingly, it has been argued that high achievement motivation goes along with domain identification and low achievement motivation is associated with chronic disengagement (see Crocker & Major, 1989; Steele et al., 1992; Steele, 1997). Thus, participants in Smith et al.’s study who were low in achievement motivation might also have been psychologically disengaged from the domain, i.e., they did not consider performance in that domain as important. Participants high in achievement motivation on the other hand most likely cared a lot about their performance in that domain.

When introducing a performance approach goal, where the focus is on attaining success (Elliot & Church, 1997), participants high in achievement motivation were probably very motivated to show that their performance is among the best and showed an increase in interest. Participants low in achievement motivation on the other hand possibly did not care much about achieving high performance and thus their interest could not be enhanced by the approach goal. When introducing an avoidance goal, where the focus is on the avoidance of failure (Elliot & Church, 1997), however, participants low in achievement motivation may have wanted to ”avoid demonstrating incompetence” or, in other words, avoid confirming the negative stereotype present under stereotype threat and thus demonstrated an increase in interest in the task. Participants high in achievement motivation, however, might not have been as sensitive to the avoidance goal, because their aim is to attain excellence and not to avoid failure. Accordingly, individuals should be most sensitive to the goals in line with their motivational orientation. Returning to the when question motivation should increase under stereotype threat when a person’s goal is in line with his or her motivational orientation and should remain stable or decrease when it is not.

[...]

^[1] Conversely, other authors have argued that stereotype threat does not activate a prevention focus and no stereotype threat a promotion focus, rather prevention-focused individuals are thought to be particularly sensitive to the negative expectancies present under stereotype threat whereas promotion- focused individuals are thought to be particularly sensitive to the relatively positive expectancies present in no stereotype threat conditions (see Keller & Bless, 2008). The approach taken in this thesis integrates both accounts by postulating that stereotype threat can temporarily activate a prevention focus and no threat a promotion focus, because the presence or absence of negative outcomes (stereotype threat) or positive outcomes (no stereotype threat) are salient in the situation (cf. Higgins, 1997)

^[2] According to regulatory fit theory (Higgins, 2000) regulatory nonfit refers to the mismatch of a person’s regulatory orientation and his or her goal.

^[3] Words in brackets were added to the definition offered by Wheeler and Petty (2001).