Implicit Association Test as an Indicator of Gender Bias in Computer Fields
Although gender bias in technical fields was observed by the author in the early days of her career, thirty years ago, it was her expectation that attitudes of current students would be much more gender neutral. To test this hypothesis, she assigned students in a Computer Ethics class to write about the Gender Gap in computer fields, then take the Harvard Implicit Association Test (IAT) relating gender and science, and finally to reflect on the results of the IAT. It was the expectation of the author that students today would perceive themselves to be completely unbiased, but would, in fact, be slightly more biased than they believed. Results, however, were both surprising and concerning. This paper discusses the assignment, the results, and reflections on the current state of perceptions of women in the computer field.
The unpublished paper by Smyth, Greenwald, and Nosek  indicates, based on two IAT studies including over 110,000 college graduates and students, that the male students who have the strongest automatic association between men and science are the most likely to major in these fields, while women with similarly strong associations were least likely to do so. This would be consistent with the results in the UTC course, as most of the respondents were men in technical fields. The study  also indicated that women who majored in scientific fields tended to have the least automatic association of men with science. Researchers found that “For women in both studies, implicit stereotyping was more strongly related to majoring in STEM [Science, Technology, Engineering, and Mathematics] than was SAT-math performance, an indicator of math ability that is often prominent in models of post-secondary STEM interest and achievement.” This could present questions that are difficult to answer: if the male students in the STEM fields are more likely than the general population to hold stereotypes about the appropriateness of male and female roles in those fields, how does this affect the female students in these majors? And if stereotypical ideas of appropriateness of females in science are an indicator for which women will go into those fields, how can the number of women in STEM areas be significantly increased, both at the university level and in the profession?
This is especially concerning, as some studies indicate that women’s beliefs of male/female abilities affect women’s performance appreciably. In a study by Spencer, Steele, and Quinn, cited in , male and female first year psychology students at the University of Michigan were given a test based on the math portion of the Graduate Record Exam. Although all of the students had similar math abilities and backgrounds, the results varied significantly based on what students were told before the test was taken, as seen in Figure 3. If the students were told that “there were no gender differences in test performance,” male and female performance was very similar, and differences were not statistically significant. However, if the students were told that “men tend to do better,” male performance increased slightly,and female performance decreased very significantly. Although this is only one study, it does suggest that student expectations of male and female performance can affect the level of accomplishment of female students in particular.