What We Don't Know (or Refuse to Say) About Gender and Trade Policy Preferences

Abstract

Survey research on trade policy attitudes has frequently found that women are more protectionist than men. One possible reason for these findings that previous research has not considered is the fact that women are more likely than men to say they "don't know" in response to many survey questions, including those concerning trade policy. We consider two ways in which this pattern might influence estimates of the size of the gender gap. The first concerns the construction of dependent variables in analyses of survey data. Including the "don't know" respondents in a residual category alongside either supporters of opponents of trade protection can greatly influence the apparent size of the gender gap even if there are not many "don't know" responses. Second, the process generating the "don't know" responses could be related to the process shaping trade policy attitudes, creating sample selection bias when the "don't know" responses are treated as missing data. We use a selection model to test the extent of this bias and find no evidence of it in estimates of the gender gap. Like other research on this topic, we cannot entirely account for the gender gap.

Keywords: trade policy, public opinion, gender, non-response