An intervention in stereotype threat : does gender-affirming literature reduce female vulnerability in mathematics?

The current study examined whether the presentation of gender-affirming literature, in the form of information about the University of Richmond‟s Women, Gender, and Sexuality Studies (WGSS) program, mitigated female stereotype threat in the domain of mathematics. A pilot study tested 10 male and 10 female college-aged participants to determine whether a derogatory beer commercial produced the threat effect in women; results were inconclusive. A follow-up study used threat methods derived from previous research and information about the WGSS program at Richmond for the intervention manipulation. Results from 30 female college students indicated insufficient threat and a nonsignificant positive increase in performance following the intervention. Further research is needed to establish the supportive nature of gender-affirming literature. GENDER-AFFIRMING LITERATURE AND FEMALE THREAT 3 An Intervention in Stereotype Threat: Does Gender-Affirming Literature Reduce Female Vulnerability in Mathematics? The women‟s liberation movement has undergone a long and tortuous path in America. Although hiring practices, wage discrimination, and other forms of injustice are more carefully regulated in modern society, prevailing notions of female roles create a more subtle menace in the form of stereotype threat. Stereotype threat proposes that as members of a group are presented with a negative stereotype about themselves, they become preoccupied with the fear of reinforcing said stereotype (Steele, 1997). This threat produces many negative outcomes, such as decreased working memory capacity and underperformance in the stigmatized field (Schmader & Johns, 2003). The current research intends to address the significance of stereotype threat in relation to gender disparities in science, technology, engineering, and mathematics (STEM) fields. When assessing ability distributions across sex, previous research has found a higher percentage of males at both the higher and lower end of the spectrum – this has been attributed to not merely brain structure, but rather a range of sociocultural influences, including early experiences and educational policies (Halpern et al., 2007). Further reviews of national education results dating back to 1988 have found no significant differences between males and females in mathematic performance (Lindberg, Hyde, Petersen, & Linn, 2010). Despite these findings, there is still a marked gender gap in STEM occupations, with women making up 11% of professionals in engineering and 29% of professionals in the physical sciences (National Science Board, 2006). One possible factor behind such pronounced gender disparities involves the social messages implied in STEM areas; namely, women are stereotyped as not biologically suited to these GENDER-AFFIRMING LITERATURE AND FEMALE THREAT 4 particular fields, and those who defy that stereotype are viewed as „unfeminine.‟ These and other stereotypes may create a considerable threat that hinders female participation in the STEM fields. Manifestations of Stereotype Threat Past research has demonstrated that stereotype threat reaches across gender and racial boundaries (Steele, 1997). By distributing very difficult Graduate Records Examinations (GRE) problems to college men and women majoring in mathematics, Steele (1997) demonstrated that women underperform in the stereotyped field despite possessing comparable academic qualifications. Similarly, when researchers presented students with GRE problems described as „abilityand intellectually-diagnostic‟, African Americans of equivalent aptitude performed poorly in comparison to Caucasian counterparts (Steele). Steele claims that discrimination against women in male-dominated fields such as math and science, as well as against African Americans in academics, creates this threat and results in both underperformance and decreased self-confidence in these areas. As members of the group fear promoting and supporting a stereotype, they lose the motivation to identify with and achieve in the prejudiced domain; this has resulted in the underrepresentation of women in math and science fields, in addition to racial gaps in standardized and I.Q. testing (Steele, 1997; McKay, Doverspike, Bowen-Hilton, & Martin, 2002). Stereotype threat also affects cognitive and physiological functioning. When women were faced with stereotype threat in the form of false negative feedback on a difficult GRE math task, they exhibited increased aggression levels, increasingly unhealthy eating habits, risky decision making, and decreased attentional control as compared to non-threatened women (Inzlicht & Kang, 2010). Evidently, stereotype threat extends beyond the immediate GENDER-AFFIRMING LITERATURE AND FEMALE THREAT 5 consequences of impaired self-confidence and performance and can spill over into unrelated domains. As responding to stereotype threat depletes reserves of self-control, individuals are more inclined to experience negative effects in subsequent behaviors (Inzlicht & Kang, 2010). While stereotype threat generally produces negative behavioral outcomes, the presentation of said threat can also provoke an opposite, positive reaction in stereotyped groups. Hoyt, Johnson, Murphy, and Skinnell (2010) examined stereotypes of females in leadership roles across explicit and implicit boundaries. When exposed to a blatant threat (explicit condition), such as a derogatory article about a lack of women in upper-level business positions, women were more likely to display reactance, or an emotional reaction that strengthens the urge to contradict stereotypes, and take on a subsequent task involving leadership. However, when Hoyt et al. implemented a more implicit threat, such as manipulating the male-to-female ratio in participant groups, threatened women were subsequently more likely than non-threatened women to take a background role in a leadership task scenario. The current study employed both explicit and implicit manipulations of stereotype threat to examine its effects on math problemsolving ability and math self-efficacy in college women. It was expected that a combination of explicit and implicit components of stereotype threat would create vulnerability and not reactance in women. Reducing Stereotype Threat Proposed methods of reducing stereotype threat include building self-efficacy, or confidence in a specific field, through improved teacher-student relationships (Steele, 1997). Increases in self-efficacy have been correlated with improved performance across a variety of domains (Stajkovic & Luthans, 1998; Multon, Brown, & Lent, 1991). By improving academic GENDER-AFFIRMING LITERATURE AND FEMALE THREAT 6 relationships, it is theorized that authorities can create a safer environment in which stereotype threat is reduced. A pilot study using questionnaires and correlational methods was conducted at the University of Richmond (Dowd & Berry, 2010). I assessed the nature of mentoring relationships in college students and faculty members, with a focus on how these relationships affect academic self-efficacy. Analyses showed that students seem to prefer a mentor of the same sex, as indicated by a   p = .067. This suggests a positive association between student and mentor gender; although nonsignificant, it is trending in the correct direction. This might support the theory that mentoring of women by women will lead to increased participation of women in fields such as math and science where high-performing women may be stigmatized (Nielsen, Marschke, Sheff, & Rankin, 2005). Furthermore, the perceived effectiveness of mentoring was positively correlated with academic self-efficacy, r = .31, p < .05, suggesting that strong mentoring relationships increase academic confidence. The results from this preliminary investigation of sex differences in academic achievement and self-efficacy provided a conceptual bridge to the current study. Another method for reducing stereotype threat involves improving identity safety (Davies, Spencer, & Steele, 2005). An experiment on negative female stereotypes in leadership had women view television commercials, such as an ad for a local college that presented a female applicant as a vapid blond, which affirmed inferior roles for women. This caused them to be less likely than non-threatened women to take on a leadership role in a subsequent task. However, a statement that affirmed equal gender capability in leadership roles reduced the effects of stereotype threat in a comparable female participant pool. Through the promotion of positive identity environments that assert equal abilities between the sexes, women can more ably resist threatening stereotypes that prime for lesser gender roles (Davies et al.). The current GENDER-AFFIRMING LITERATURE AND FEMALE THREAT 7 study expanded upon this experiment by promoting identity safety through educational literature on a Women, Gender, and Sexuality Studies (WGSS) program and a Women in Living and Learning (WILL) program at the university from which participants were drawn. Research Overview and Hypotheses The current study applied previous findings on stereotype threat in mathematical and scientific fields to college women. Research on female undergraduates in male-dominated domains, such as engineering, computer science, and mathematics, found that these women were more likely to report discrimination, greater feelings of intellectual vulnerability, and a desire to change majors (Steele, James, & Barnett, 2002). A pilot study applied the experimental method of Davies et al. (2005) in determining how a threatening video, specifically a beer commercial, affects math self-efficacy and performance in women. It was hypothesized that women who view a threatening commercial prior to a math task will report lower levels of self-efficacy and will perform worse than women who do not view a threatening commercial. Overall, math selfefficacy should be highly correlated with math performance regardless of threat manipulation. The primary study, also on female stereotype threat in mathematics, assessed the effect of an intervention involving gender-affirming literature on reported math self-efficacy and math performance. It was hypothesized that female participants who are presented with positive literature derived from the WGSS and WILL programs after stereotype threat will experience reduced threat and subsequently report higher math self-efficacy and perform better on a math task. GENDER-AFFIRMING LITERATURE AND FEMALE THREAT 8 Method – Pilot Study Participants Ten male and ten female participants were recruited at the University of Richmond. The age range was 18-23 years, with a mean age of 20. The ethnic composition of participants was 70% Caucasian, 15% African American, 10% Hispanic, and 5% Asian. Participants were recruited by posting flyers in academic buildings and sending out campus-wide e-mail announcements. Participants were offered $10 as compensation for taking part in a 45 minute study. The experiment was described as a „short-term memory‟ study, as mentioning the use of stereotype threat could prime participants and affect results. Both the experimental and control trials included consent forms and a debriefing of information. In order to ensure confidentiality, each participant‟s data packet was assigned a random number and kept separate from the consent form. Materials All participants completed a general questionnaire reporting demographic information, Scholastic Aptitude Test (SAT) Math scores, and previous math courses (see Appendix 1). The experimental, threat-inducing beer commercial was a 30-second Heineken advertisement which portrayed an attractive female figure that robotically provided beer through a keg dispenser. Researchers chose this commercial due to its promotion of negative gender stereotypes, specifically the passive and beautiful female. The control commercial was a Budweiser advertisement, also 30 seconds in length, which centered on the Budweiser Clydesdale and presented no gender stereotypes. A short survey on the beer commercials provided distracter marketing questions about the product, such as “How persuasive was the ad?” and “How likely are you to buy this product?” (see Appendix 2). GENDER-AFFIRMING LITERATURE AND FEMALE THREAT 9 The Math Attitudes and Perceptions Scales (see Appendixes 3 and 4), the Math Task Self-efficacy Scale (see Appendix 5), and the Math Performance measure (see Appendix 6) were all derived from previous research (Pajares, 1996). The Math Attitudes Scale included questions such as “I get really uptight during math tests” and “I have usually been at ease in math courses” – these were rated on a 1-5 Likert scale, 1 being “strongly disagree” and 5 being “strongly agree”. The Math Perceptions Scale included questions such as “I am quite good at mathematics” and “Mathematics makes me feel inadequate” – these were rated on a 1-8 Likert scale, 1 being “Definitely false” and 8 being “Definitely true.” The Math Task Self-efficacy scale asked participants to rate how confident they were that they would be able to solve each of the specific math problems that were later presented in the Math Performance task. Confidence ratings were made on a scale of 0-100%. The math performance measure included 18 questions such as “ABOUT how many times larger than 614,360 is 30,668,000?” and “3 4 /5 1 /2 = _____” (see Appendix 6). Procedure Participants were randomly assigned to either the experimental or the control group, with 5 men and 5 women in each level. We hypothesized that the effect of the threat would be present in women but not men; therefore, male participants were included to test relative effects between experimental and control groups. Sessions were held in single-sex groups of up to 3 people, with only male or female participants in each trial. Participants completed the general questionnaire before viewing either the experimental or control commercial. In order to avoid conscious priming, participants were informed that this commercial was presented in order to assess product marketability. After completing the marketing survey questions, participants completed the Math Attitudes and Perceptions Scales. GENDER-AFFIRMING LITERATURE AND FEMALE THREAT 10 In order to ensure that participants would not attempt to solve the math problems while completing the Math Task Self-efficacy Scale, the instructor then led the group by reading the questions aloud from a PowerPoint slide presentation. Participants viewed each question for as long as the instructor took to read it, and then had eight seconds to rate their confidence at solving each problem correctly. After participants had provided their confidence ratings for each of the math problems, they were then given 20 minutes to solve as many of the 18 problems as possible. A single piece of scrap paper was provided, and participants were not allowed to use a calculator. After the 20 minute time period ended, the instructor debriefed the participants as to the purpose of the study. Results – Pilot Study Self-reported math self-efficacy was highly correlated with number of problems correctly solved, r = .72, p = .001 (see Figure 1). Women presented with the threatening video reported slightly higher self-efficacy (M = 80% confidence, SD = 5.91) than women presented with the control video (M = 80% confidence, SD = 19.99), but this was nonsignificant. Likewise, women presented with the threatening video performed better on the math task of 18 questions (M = 12.67 questions correct, SD = 1.63) than women presented with the control video (M = 9.50 questions correct, SD = 5.74), but this was nonsignificant (see Figure 2). Male participants showed similar trends, with the threatened condition reporting higher self-efficacy and higher performance. Discussion – Pilot Study The hypothesis that math self-efficacy is highly correlated with math performance was supported. However, the hypothesis that women who view a threatening commercial prior to a GENDER-AFFIRMING LITERATURE AND FEMALE THREAT 11 math task would report lower levels of self-efficacy and would perform worse than women who do not view a threatening commercial was not supported. To the contrary, women in the experimental group demonstrated higher levels of both math self-efficacy and math performance. These results do not support previous research by Davies et al. (2005), which found that stereotype threat presented through video creates sufficient vulnerability in women and has a negative impact on their likelihood to accept a subsequent leadership task. Several limitations within the pilot study can account for these results. A small sample size of 20 participants, only 10 of which were female, restricts the reliability of the data. The Heineken beer commercial may also have presented too explicit a threat. Previous research has demonstrated that, when presented with a single explicit stereotype threat, women are more likely to display reactance and perform more positively in the discriminated field (Hoyt et al., 2010). As a result of the chosen commercial, the experimental group may have experienced the threat too explicitly and gained positive resolve from the stereotype. Furthermore, analysis of the testing environments revealed that the primary female control group was tested by a male experimenter. Past experiments have shown that the presence of an assertive male as instructor can produce sufficient implicit threat and decrease performance (Hoyt et al., 2010; Steele et al., 2002). While additional pilot studies may have determined a more implicitly threatening commercial, project resources did not allow for the gradual development of a valid and reliable threat manipulation. In order to apply the gender-affirming literature as an intervention, the primary study used an implicit form of stereotype threat supported by published research (Hoyt et al., 2010; Steele et al., 2002). Also, researchers shortened the allowed time limit for the math GENDER-AFFIRMING LITERATURE AND FEMALE THREAT 12 task in order to increase task difficulty and optimize the probability of obtaining group differences (Steele, 1997). Method – Primary Study