Arizona State University | Ranked #1 for innovation in the ...



Running head: EFFECT OF STEREOTYPE THREAT 1

Research Proposal

How Expert Influence Moderates Stereotype Threat in Adolescent Girls

Denise T. Kaeo

PSY 290-Research Methods

Dr. Lorraine Festa

Arizona State University

April 21, 2012

Abstract

Although research has shown that stereotype threat has been linked to lower performance in math among females with higher levels of mathematics identification, it is unclear whether reinforcing disconfirming stereotypic math beliefs moderates their susceptibility to stereotype threat. This study examines how exposure to expert influence disconfirming the negative stereotype of girls’ ability in mathematics affects the math performance of adolescent females with stereotypic and counter-stereotypic beliefs. One hundred adolescent females from diverse ethnic and socio-economic backgrounds will be randomly assigned to an experimental or control group and be required to complete a math test. It is predicted that the participants in the experimental group with counter-stereotypic beliefs, who are exposed to expert influence, will earn significantly higher math scores than the participants in the control group who are not exposed to expert influence. Findings from this study will advance our knowledge about the importance of appropriate intervention when aiming for true gender equality in science, technology, engineering and math achievements. Implications for future research will be described.

How Stereotype Threat Effects Adolescent Girls

“For the past few decades, women in the US have tended to underperform men by 50-80 points on the mathematics sections of the SAT and GRE” (Halpern, 1992; Wah & Robinson, 1990) and they are less likely than men to obtain majors in math or engineering fields in college (National Science Foundation, 2000). In addition, a dramatically larger proportion of women have entered the labor force, and while they have made significant gains in paid employment (National Science Foundation, 2002); studies reveal that women are underrepresented in math and science related careers (Meece, 2006). According to the National Science Foundation (2008) men earned a majority of the bachelor’s degrees in engineering, computer sciences and physics (80%, 78%, and 79% respectively). Whereas, approximately 90% of domestic workers, nurses, secretaries, receptionist, and elementary/middle school teachers are women (US Department of Labor, 2003). These statistics clearly demonstrate that significant gender disparity exists in the US workforce. Women, including those who are equally skilled in mathematics and experience as their male counterparts, still avoid majors involving medium to high levels of math (Lefevere, Kulak, & Heymans, 1992). This gender disparity is of great concern because women’s employment and career choices have a significant impact upon their family income and life satisfaction (Meece, 2006); the fact that they are underrepresented in some of the fastest growing and highest paying occupations is detrimental for their well-being since single women are more likely than men to have incomes below poverty line, and 39% of children who live in poverty come from female single-headed households (Lichtenwalter, 2005). The career decisions of women will also have serious long-term effects since “women have longer life expectancies than men and fewer than 30% of older women in the US are receiving

retirement incomes” (US Department of Labor, Women’s Bureau, 2004b).

There is widespread concern in our society over the consequences of this gender disparity and much research (Bonnot & Croizet, 2007; Brown & Pinel, 2003; Cherney & Campbell, 2011; Davies, Quinn, Gerhardstein, 2002; Keller, 2007; Meece, 2006) is being done to find answers for this gender divide. The causes for this underrepresentation may stem from complex interactions between biological and environmental factors; however, strong consensus among researchers indicate that a distinct social process known as stereotype threat may have a strong influence upon the lack of female advancement in math and science careers and negatively impair women’s math performance (Brown & Pinel, 2003; Cherney & Campbell, 2011; Krendl, Richeson, Kelley & Heatherton, 2008; Schmader, 2002, Schmader & Johns, 2003; Spencer, Steele & Quinn, 1999; Steele, 1997). An influential study by Steele (1997) offered a stereotype threat theory rooted in the power of social stereotypes to influence thought and behavior. According to Steele and colleagues (1995), stereotype threat is defined as the increased performance pressure experienced by individuals who must perform a task on which they are socially stereotyped and the added concern that poor performance could be seen as confirming a negative stereotype often leads to a decreased level of performance (Steele, 1997; Steele & Aronson, 1995). Subsequently, a great deal of research has surfaced to explore the effects of stereotype threat on women’s math performance and has uncovered information about various methods of inducing stereotype threat (Davies, Spencer, Quinn, & Gerhardstein, 2002; Inzlicht & Ben-Zeez, 2000; O’Brien & Crandall, 2003). Due to growing evidence from various research studies, there is strong consensus that stereotype threat has a strong influence on women’s math performance (Brown & Pinel, 2003; Cherney & Campbell, 2011; Krendl, Richeson, Kelley &

Heatherton, 2008; Schmader, 2002; Schmader & Johns, 2003; Spencer, Steele & Quinn, 1999; Steele, 1997); however many parameters of this effect have yet to be defined and tested. For example, in a study conducted by Schmader (2002) it was determined that those who highly identified with the negative stereotype of their social group experienced a greater degree of the threat and exhibited decreased performance. Similar results were found in a Brown and Pinel (2002) study where female undergraduate students with low stigma consciousness showed no relation to performance. These results suggest that women with a higher awareness of the stigma of female inferiority in math will perform worse than women of low stigma consciousness. Evidently, social-psychological research has demonstrated that knowledge of negative gender stereotypes can influence academic outcomes of stigmatized individuals, but what happens when participants are exposed to disconfirming stereotype beliefs? A study by Selimbegovic, Chatard and Mugny (2007) revealed that young women with counter-stereotypic beliefs or low stigma consciousness increased their intentions to pursue math/science-related careers.

The present findings provide ample evidence that social-processes, such as ‘stereotype threat’ significantly impair the math performance of women who highly identify with the negative stereotype of their social group. However, most of the previous studies included participation from undergraduate students with an average age of 18 and a diverse pool of ethnic and/or socioeconomic backgrounds was not an important factor when selecting participants (Bonnot & Croizet, 2007; Cherney & Campbell, 2011; Crisp, Bache & Maitner, 2009; Davies, Spencer, Quinn & Gerhardstein, 2002; Schmader, 2002). Moreover, fewer studies have been launched to explore the effects of stereotype threat upon adolescent girls in middle school

where studies have failed to show any overall gender gap on math tests (Hyde & Linn, 2006) and investigate ways induce change in stereotype beliefs. For this research study, the hypothesis would be that expert influence moderates stereotype threat and positively influences the math performance of adolescent females with counter-stereotypic beliefs.

The primary objective of this study is to explore expert influence as a means to induce change in stereotype beliefs and further support the evidence that stereotype threat has little effect on females who do not strongly identify with the negative stereotype of their social group. Therefore, the results of this study should provide the additional evidence that supports the efforts of educators to continue enforcing practices and policies that refrain from sex-biased instruction and promote gender equity; thereby, reinforcing the need for expert influence in school settings as a possible means to encourage girls’ ability in mathematics and encourage mobility toward STEM related careers. The research would be conducted as a cross-sectional study. Participants would be selected from three public middle schools located in an urban and suburban area of Southern California with a diverse ethnic population. One hundred middle-school aged female students (ages 13 – 14) with advanced math abilities will be selected and randomly assigned to an experimental or control group. Participants will be required to take a math test and take a pretest to measure gender stereotyping in mathematics.

Method

Participants

As a cross sectional study, participants in this study will include 100 middle-school age, 8th-grade, female adolescent students (ages 13 – 14) from three middle schools located in the Redlands Unified School District, an urban and suburban area of variable socioeconomic status

in Southern California. Thirty-five percent of the participants will be Caucasian, 7.9% African-American, 18% Asian, 35% Hispanic and a marginal percentage will be multi-racial. This ethnic mix is the ethnic composition of the school district. In a study by Aronson et al. (1999) research revealed that stereotype threat is greatest for people who value the domain being tested. Subsequently, to reduce the number of participants with limited math ability or who are unconcerned with math, only participants who scored above 425 (advanced) in mathematics on the California Standardized Testing and Reporting (STAR) would be eligible for this study. Experimenter will obtain permission to conduct a study of “student’s academic motivation” from school administrators. Participation from students who meet the math eligibility score will be voluntary and parental consent to participate will be required.

Design

This design is a two-group, posttest-only, randomized experiment because it utilizes both a control group and a means to measure the mathematics performance change that occurs in both groups. The independent variable is exposure or non-exposure to expert influence and the dependent measure for this study is the level of stereotype threat and math performance. Adolescent females would be randomly assigned to the experimental or control condition, and

they would be required to complete a mathematics test (MAT) along with answering one additional question to assess gender stereotyping in mathematics. Participants’ math abilities would be assessed with the Mathematics Achievement Test (MAT) created for the purposes of this experiment and modeled after California Standardized Testing and Reporting (STAR), which are described in more detail below.

In the experimental condition, expert influence disconfirming the negative stereotype of girls’ ability in mathematics would be communicated to students before taking the mathematics exam, thereby promoting feelings of competence in mathematic ability and suppressing the stereotype threat effect. In the control group, participants would receive basic instructions for the exam. Both groups would be given a pretest to measure gender stereotyping in mathematics, participants would be asked to indicate whether, in their personal opinion, boys are better than girls, girls are better than boys, or boys and girls are equally good at mathematics. Students would respond to a 3-point scale ranging from -1 to +1 with a neutral point (0) to indicate judgment of equal abilities in boys and girls. As a continuous variable in the design, individual differences in the belief of gender stereotypes (initial position) would be examined. In an attempt to control for all confounding variables, or at least consider their impact while attempting to determine whether the expert influence (independent variable) is causing the change, random assignment to an experimental group or control group has been incorporated into this research design. To eliminate plausible alternative explanations, all measures will be completed in one day. The research design incorporates a control group to rule out additional plausible alternative explanations and establish covariation of cause and effect. Evidence of temporal precedence will be met by administering the exam before measuring the effects.

There are no real multiple-group threats to internal validity because exam will be administered within one day, utilizing one math test instrument, and a pretest design will not be employed.

Design Notation:

R X O

R O

Measures

Mathematics teachers from the three middle schools in the Redlands Unified School District would create the Mathematics Achievement Test (MAT) used to assess the participants’ math ability for the purpose of this experiment. The teachers would be asked to provide a list of topics that had been covered in the Grade 8 curriculum during the present academic year. The researchers would review the list, and common topics already covered that typically appear on

the California STAR math test would be selected as topics for the exam. The final test would consist of 15 difficult multiple-choice items randomly selected from each set of test questions created by the mathematics teachers. The test content areas to be assessed would be mainly comprised of Algebra One subjects such as: Number Properties, Linear Equations, Graphing, Quadratics, Polynomials Functions, Rational Expressions and Geometry, all standard topics for grade 8 Mathematics in the State of California (California State Board of Education, 2010). Students would complete the mathematics test on computer coding sheets that would be graded electronically.

Procedure

This study is designed to determine if exposure to expert influence disconfirming negative stereotypes of girls’ ability in mathematics would affect the math performance of

Grade 8 adolescent females. Once the 100 eligible, ethnically diverse, participants are obtained and school administrators and parents have given their written consent to allow their child to participate in the study, the exam would be scheduled to take place at one of the middle school campuses (in a natural setting) during the last quarter of the school calendar year. Upon arriving at the exam, participants would be randomly assigned to either the control or experimental

group. Once students were in their respective test-taking classrooms, a professionally dressed female research staff member would read aloud instructions for the exam. In the experimental condition, the disconfirming stereotype in the form of bogus scientific findings showing past success rate of girls would be presented in a convincing manner before students take the test. The following description would be used in the experimental group to introduce the exam:

You are about to take a Math Achievement Test (MAT) and it is a very reliable indicator of one’s math ability and is commonly used to test mathematical skills. According to our past research, girls’ scores have been significantly higher than boys’. Please answer the questions provided to the best of your ability. You will have 45 minutes to complete 15- multiple-choice questions. Please select the answer you think is correct for each question by shading in the circle that corresponds to the correct answer. Your performance on this exam will be compared with past performance of students.

In the control condition, participants would receive the following basic instructions for the exam:

This exam consists of 15-multiple-choice questions and you will be given 45 minutes to complete the exam. As you complete the exam, please use the worksheet provided to calculate your answers. Each questions has four possible answers and you must choose

one answer that you feel is correct. Please select the answer that you think is correct by shading in the circle that corresponds to the correct answer.

As outlined in the design section above, a pre-test question to measure gender stereotyping in mathematics would be included in the beginning of the test for both the experimental group and the control group. Before the researcher has gone over the exam instructions the researcher

would ask the participants to indicate whether, in their personal opinion, boys are better than girls, girls are better than boys, or boys and girls are equally good in mathematics. Students would answer on a 3-point scale ranging from -1 (girls are better than boys in mathematics) to +1 (boys are better than girls in mathematics); the neutral point (0) indicated perception of equal abilities in boys and girls – Scale question adapted from a study by Muzzatti & Agnoli

(2007) to determine gender stereotypic beliefs in mathematics. Once the students from each group completed the pre-test question the researcher would read aloud the exam instructions and students would be instructed to begin the exam. Each participant would be required to write down his/her full name and school in the blanks provided on the computer coding answer sheet.

All student participants would be debriefed upon completion of the exam. Participants would be informed of the fictitious information provided to the control group and general information regarding the focus of the study would be explained. All participants would receive hard copies of the debriefing so they could discuss the experience with their parents or guardians.

Analysis

To test whether stereotype threat effects were moderated by expert influence, a

2 (condition: expert influence vs. control) x 2 (belief: stereotype vs. non-stereotype) one-way analysis of variance (ANOVA); using mathematics test scores as the dependent variable, would be conducted. Test scores from the control group and experimental group would be analyzed. Significantly higher scores from the experimental group compared to the control group, specifically the girls in the experimental group identified as participants with counter-stereotypic beliefs, would confirm the hypothesis that expert influence disconfirming the

negative stereotype of girls’ ability in mathematics positively effects the math performance of adolescent females with counter-stereotypic beliefs.

Results

The results of this study are expected to show that participants in the experimental group, who are exposed to expert influence disconfirming the negative stereotype of girls’ ability in mathematics, will earn significantly higher math performance scores than the participants in the control group who are not exposed to expert influence before taking the math test. The results of the ANOVA will be analyzed; and it is expected that the adolescent females with counter-stereotypic beliefs, who are exposed to the expert influence condition, will have the highest scores. Because advanced female math students often experience a high degree of stereotype threat in situations that could confirm gender-based limitations; specifically, when challenged with math testing situations that involve difficult math problems (Steele, 1997), it is predicted that the female adolescents in the control group who are not exposed to expert influence disconfirming negative math stereotype, and who believe in the gender stereotype that boys are better in math than girls, will have the lowest scores.

Discussion

A gap in literature with respect to research conducted on possible ways to suppress the stereotype effect formed the basis for this study. How expert influence disconfirming the negative stereotype of girls’ ability in mathematics impacts the math performance of adolescent females with advanced math abilities, will be examined. The use of a two-group, posttest-only,

randomized experimental design in this study allows researchers to take an in-depth look at the mathematical performance change of adolescent females when exposed or not exposed to expert influence. Keeping a design focus on middle school aged females, where studies have failed to show an overall gender gap in mathematics (Hyde & Linn, 2006), will be an improvement over past research because it may possibly provide evidence of ways to prevent the stereotype effect during the important transitional age of adolescence. In addition, the ethnic and socio-economic diversity of the population sample will provide more information for a currently under-researched population of adolescent females. One limitation of this study is that it will be conducted with grade 8 participants only, and repetition with different age groups would be necessary before reaching confident conclusions that expert influence moderates the effects of stereotype threat. Also, findings from this study will advance our understanding of the possible role that expert influence may have upon negative stereotype beliefs in mathematics and introduce ways to promote female advancement in science, technology, engineering and math (STEM) related careers. It is also quite plausible that the negative gender stereotype beliefs of the participants will prove to be resistant to expert influence (Devine, 1989) and that math scores of the experimental and control group will show little or no differences – In this case, much more research will be needed to determine the age at which stereotype threat effects begin to emerge and what practices can be utilized to suppress the negative effects on academic achievement. Future research should utilize ANOVA models to test for demographic differences in ethnicity and age. Also, the restricted range of the sample limits the scope and interpretation of the role of expert influence and future research should include participants with lower math abilities from a wider demographic sample population.

References

Aronson, J., Lustina, M.J., Good, C., Keough, K.A., Steele, C.M., & Brown, J. (1999). When white men can’t do math: Necessary and sufficient factors in stereotype threat. Journal of Experimental Social Psychology, 35, 29 46.

Bonnot, V., & Croizet, J.C. (2007). Stereotype internalization, math perceptions, and occupational choices of women with counter-stereotypical university majors. Swiss Journal of Psychology, 169-178. doi:10.1024/1421-0185.66.3.169

Brown, R. P., & Pinel, E.C. (2003). Stigma on my mind: Individual differences in the experience of stereotype threat. Journal of Experimental Psychology, 39, 626- 633.

California State Board of Education: Common core state standards for Mathematics, Adopted August 2010.

Cherney, I.D., & Campbell, K.L. (2011) A league of their own: Do single-sex schools increase girls’ participation in the physical sciences? Springer Science Business Media, 65:712-724. doi:10.1007/s11199-011-0013-6

Crisp, R., Bache, L., & Maitner, A. (2009). Dynamics of social comparisons in counter- stereotypic domains: Stereotype boost, not stereotype threat, for women engineering majors. Social Influence, 4(3), 171-184.

Davies, P.G., Spencer, S.J., Quinn, D.M., & Gerhardstein, R. (2002). Consuming images: How television commercials that elicit stereotype threat can restrain women

academically and professionally. Personality and Social Psychology Bulletin, 28, 1615 – 1628.

Devine, P. G. (1989). Stereotypes and prejudice: Their automatic and controlled components. Journal of Personality and Social Psychology, 56(1), 5-18. doi:10.1037/0022- 3514.56.1.5

Halpern, D. F. (1992). Sex differences in cognitive abilities. Hillsdale, NJ: Lawrence Erlbaum Associates.

Hyde, J. S., & Linn, M.C. (2006). Gender similarities in mathematics and science. Science, 314, 599-600.

Inzlicht, M., & Ben-Zeev, T. (2000). A threatening intellectual environment: Why females are susceptible to experiencing problem-solving deficits in the presence of males. Psychological Science, 11, 365-371.

Keller, J. (2007). Stereotype threat in classroom settings: The interactive effect of domain identification, task difficulty and stereotype threat on female students’ math performance. British Journal of Educational Psychology, 77, 323- 338.doi:10.1348/000709906X113662.

Krendl, A. C., Richeson, J. A., Kelley, W. M., & Heatherton, T. F. (2008). The negative consequences of threat: A functional magnetic resonance imaging investigation of the neural mechanisms underlying women’s underperformance in math. Psychological Science, 19(2), 168-175. doi:10.1111/j.1467-9280.2008.02063.x

Lefevre, J., Kulak, A., & Heymans, S. (1992). Factors influencing the selection of university majors varying in mathematical content. Canadian Journal of Behavioral Sciences, 24, 276-289.

Lichtenwalter, S. (2005). Gender poverty disparity in US cities: Evidence exonerating female- headed families. Journal of Sociology and Social Welfare, Volume XXXII(2), 75-96.

Meece, J.L. (2006). Trends in women’s employment in the early 21st century. Educational Research & Evaluation, 12(4), 297-303. doi:10.1080/13803610600765539

Muzzatti, B., & Agnoli, F. (2007). Gender and mathematics: Attitudes and stereotype threat susceptibility in Italian children. Developmental Psychology, 43(3), 747-759. doi:10.1037/0012-1649.43.3.747

National Science Foundation. (2000). Women, minorities, and persons with disabilities in science and engineering: 2000. Arlington, VA: NSF 00-327.

National Science Foundation. (2002). Higher education in science and engineering: Increasing global capacity. Washington, DC: Author.

National Science Foundation. (2008, February). Science and Engineering indicators. Retrieved from

O’Brien, L.T., & Crandall, C.S. (2003). Stereotype threat and arousal: Effects on women’s math performance. Personality and Social Psychology Bulletin, 29, 782-789.

Schmader, T. (2002). Gender identification moderates stereotype threat effects on women’s math performance. Journal of Experimental Social Psychology, 38, 194- 201.

Schmader, T., & Johns, M. (2003). Converging evidence that stereotype threat reduces working memory capacity. Journal of Personality and Social Psychology, 85, 440-452.doi:10.1037/0022-3514.85.3.440

Selimbegovic, L., Chatard, A., & Mugny, G. (2007). Can we encourage girls’ mobility towards science-related careers? Disconfirming stereotype belief through expert influence. European Journal of Psychology of Education, 22(3), 275-290. doi:10.1007/BF03173426

Spencer, S. J., Steele, C. M., & Quinn, D. M. (1999). Stereotype threat and women’s math performance. Journal of Experimental Social Psychology, 35(1), 4-28. doi:10.1006/jesp.1998.1373

Steele, C.M. (1997). A threat in the air: How stereotypes shape intellectual identity and performance. American Psychologist, 52(6), 613-629. doi:10.1037/0003- 066X.52.6.613

Steele, C.M., & Aronson, J. (1995). Stereotype threat and the intellectual test performance of African-Americans. Journal of Personality and Social Psychology, 68, 797- 811.

US Department of Labor, Bureau of Labor Statistics. (2003). Facts on women workers. Washington, DC: Author.

US Department of Labor, Women’s Bureau. (2004b). Quick stats 2004. Washington, DC: Author. Retrieved February 14, 2006, from

Wah, D. M., & Robinson, D. S. (1990). Examinee and score trends for the GRE general test: 1977 78, 1982 83, 1986 87, and 1987 88. Princeton, NJ: Educational Testing Service.

................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download