Bryonna Bowen and Dr. Jessica Preece, Political Science
The gender gap in participation, capability, and interest in politics has been widely researched. However, in this project we examined whether or not people implicitly identify men more strongly than women with politics. Furthermore, I was interested to determine to what extent everyday media exposure, particularly focusing on a specific gender, could influence an individual’s implicit biases. Mock newspaper articles, with either male or female political subjects, were used to further determine the effect of media content on people’s gender biases. While self-reported prejudices may be inconsistent and untrustworthy, I used the implicit association test (IAT) to assess the patterns of implicit association of Americans with gender and politics. I analyzed the difference between explicit and implicit association, male and female respondents, and the two article primes.
The implicit association test (IATs) is a measure of social psychology that has become a fairly recent and growing method for scholars to measure individuals’ implicit attitudes or the strength of association between concepts and stereotypes. The IAT is a computer-based test that gauges biases based on response times of an individual sorting words into specific categories. The individual will sort words belonging to two categories (politic and gender) with two levels of each factor (politics: non-political and political; gender: female and male). As a word appears in the center of the screen the individual must “sort” the word into the pre-determined correct category. Based on the timing in which it takes to sort the words into the respective categories, the implicit biases can be measured.
To administer this IAT we created a Qualtrics survey with an imbedded IAT. The survey was then distributed using Amazon’s Mechanical Turk for the adult population in the United States. Although there are limitations due to the methods in which we administered the survey, we take these into consideration and were able to eliminate many of the threats to reliability and variability through our methods as well as our analysis. Four treatments (control, placebo, male article, and female article) groups were analyzed with approximately 150 males and 150 females in each treatment group.
After receiving the sorting times of each individual in depth quantitative analysis was conducted to determine gender association with politics and media effect on these biases. Greenwald, Nosek and Banaji’s Understanding and Using the Implicit Association Test: I. An Improved Scoring Algorithm and other analysis literature was used to give the most reliable algorithm for analysis. Removal of individuals who incorrectly sorted too many words ( average of 6 out of the 32 per round wrong or a total of 20% wrong) eliminated people who may not have been attentive to the test and therefore the time would not have accurately measured their bias. Additionally, we looked at whether individuals were more likely to sort male, female words incorrectly in association with the political and non-political categories.
Analyzing the results we see, as expected that there was a stronger association with men and politics. This was seen by the time to sort words increasing when individuals were required to sort female words and political words together. Furthermore, the IAT results indicate that while the media articles influenced the attitudes; more analysis ought to be conducted to pursue the impact that media can have on implicit biases and associations.
The IAT was a successful psychological research to measure the strength of American’s associations with women and politics. Although the IAT has some limitations and weaknesses to conclude strength of association with implicit biases, this experiment shows that there is definitely a stronger association with men and politics than there is with women and politics. Recognizing that the gender gap and stereotypes in politics are implicit as much as they are explicit allows us a better understanding to know how we can confront the stereotypes.
I will be presenting this research at the Midwest Political Science Conference in Chicago in April. Furthermore, I believe that using the IAT to study self-association to determine biases of oneself would be a call for further research and interesting in the area of gender politics.