The gender gap and torture: Opposition to torture among men and women in the USA

Abstract

The study of women in politics includes the analysis of both elite women and the masses in regard to their voting, political activity, and attitudes. Research shows that women hold different political attitudes. This gender gap is explained in relation to a number of attitudes, but few studies analyze whether the gender gap is still evident in an issue area that women might be torn in two hypothesized directions. The purpose of this study is to conduct an analysis of the gender gap in individual's attitudes toward the appropriateness of torture to prevent future terrorist attacks.

Keywords:

• Gender and politics
• Gender and public policy
• Gender and political behavior
• Torture policy
• Gender gap

Notes

1 Source: American National Election Study, 2008: Pre- and Post-Election Survey. International norms encouraging or accepting the use of torture possible has a long term negative impact on the state of human right in the world; however, most American will not think this far into the future and likely will believe the only individuals that suffer from this policy are foreign terrorists.

2 The response “neither favor nor oppose” is coded as missing. This response is not truly a middle category in the responses; rather, it represents ambivalence. With such a sensitive question it is assumed that there are different reasons why individuals may identify themselves as ambivalent toward this policy making this response invalid. Analyses including this response as the middle category between “favor a little” and “oppose a little” are consistent with the results where this value is coded as missing.

3 This variable is somewhat flawed, it ends at 17 years even though some individuals in the analysis do have terminal degrees. The other available education variable includes the degrees an individual has received, yet it contains a large amount of missing data making the years of education variable the best possible measure of education available.

4 The ANES (2008) Survey oversampled Latino and African-American respondents. The descriptive statistics in the Appendix are not weighted, rather they are the descriptive statistics for the sample, not the population. The analysis presented in the model is weighted.

5 The coefficients and significance levels of the coefficients of the interactions are calculated using STATA's Lincom command.