McIntyre, R., Paulson, R., & Lord, C. (2003). Alleviating women’s mathematics stereotype threat through salience of group achievements. Journal of Experimental Social Psychology, 39, 83-90.
- The study’s objective was to examine the impact of alleviating women’s stereotype threat—that is, the idea that men outperform women in mathematics—on women’s subsequent performance on a difficult mathematics test.
- The study contained two experiments in which participants were randomly assigned to either a treatment condition, which received an intervention to alleviate women’s stereotype threat, or a control condition, which did not receive the intervention. The primary outcome of interest was an adjusted score for a study-administered mathematics test consisting of difficult questions from the Graduate Record Examination (GRE), an entry test for graduate school.
- The study found that women in the treatment group scored 2.0 to 2.5 points higher on the math test than women in the control condition. These results were statistically significant.
- The quality of causal evidence presented in this study is high because it was a well-conducted randomized controlled trial. This means that we are confident that the estimated effects are attributable to the intervention, and not to other factors.
Features of the Study
The study was a randomized controlled trial, including 162 college students in the first experiment and 106 in the second. The study was implemented in mixed-gender groups, but about 70 percent of the study participants were female.
The study focused on two methods for alleviating stereotype threat—that is, the idea that women perform worse than men in mathematics—for female college students. In both experiments, an experimenter, a female in the first experiment and a male in the second, first activated the stereotype threat by telling participants that men generally perform better than women in math. In the first experiment, the female experimenter then distributed one of two types of booklets of GRE math problems to the students. The instructions in the treatment booklets stated that students would be invited to participate in additional experiments later in the semester and that only women could participate in these later experiments because “… women produce more reliable and valid data, comprehend the task requirements better, and produce better results in all types of psychological experiments [than men].” The control test booklets said that students would be invited to participate in additional experiments later in the semester, but made no mention of women’s performance in these experiments or that only women would be invited to participate.
The second experiment had an additional component: before completing the math problems, treatment group members were asked to read a series of four essays focused on women who overcame gender biases and succeeded in various professions, none of which was mathematical or scientific. Control group members also read four brief biographical essays, but they focused on successful corporations, not individuals. Both the experimenters and the study participants were blind to participants’ study group assignment.
In both experiments, researchers analyzed adjusted scores from a mathematics test that included 34 difficult quantitative items from sample GRE tests.
- In the first experiment, women in the treatment condition scored approximately 2.0 points higher on the sample GRE math test than women in the control condition, a statistically significant difference.
- In the second experiment, women in the treatment condition scored approximately 2.5 points higher on the sample GRE math test than women in the control condition, also a statistically significant difference.
Considerations for Interpreting the Findings
In both experiments, the authors conducted a 2 (sex) x 2 (intervention) analysis of variance (ANOVA) with planned contrasts comparing men and women within the same condition, across conditions, and the interaction between sex and condition for women in the control condition—that is, whether women in the control condition scored worse than the average participant. Because the study was a randomized controlled trial with no attrition, ANOVA is an appropriate analysis; however, the authors did not present the coefficients that isolate the impact of the intervention on women, which is the focus of the Clearinghouse for Labor Evaluation and Research (CLEAR) review. Therefore, CLEAR estimated the intervention’s impact on women by calculating t-tests using the means and standard deviations provided in the study.
The study activated the stereotype that women might perform worse in math for all participants before presenting a strategy to counteract it. Thus, the implications for how such an intervention would work outside the laboratory—in other words, without the artificial activation of the stereotype threat—are unclear.
Causal Evidence Rating
The quality of causal evidence presented in this report is high because the study has consistent probabilities of assignment to the research groups and has no sample attrition. This means we are confident that the estimated effects are attributable to the intervention, and not to other factors.