The Gender Gap in Ideology

For men the percentage of “don’t know” responses to the ANES ideology question has dropped from 23% to 16%, and for women from 32% to 24%.

NES has a second follow-up question on ideological self-identification: “If you had to choose, would you consider yourself a liberal or a conservative?” However, who was asked this second question varied over the years with no follow-up for the 1980 survey, in other surveys the question is asked of those responding “don’t know” or “haven’t thought much,” and in still other surveys moderates are also asked the second question. The follow-up question limits the number of response categories to three: liberal, moderate, and conservative. More importantly, the centrality of a self-identification that arises only on a second round of questioning is more suspect. Thus, we use responses only to the first ideology question.

The increased liberal identities among women come mathematically from the sum of three insignificant interactions—that women are becoming conservative slightly less quickly than men, that they are becoming moderate slightly less frequently, and that they are slightly more likely than men to adopt an ideological identity over time. None of these interactions are in themselves significant.

The income variable was not asked in the 2002 survey and the urban/rural distinction is not included in the 2002 or 2004 surveys, so the time series covers only 1972 to 2000. The following variables and codings were used: age cohort (VCF0103);dummy variable for white (code 1 in VCF0105); education (VCF0110); rural (VCF0111); coast (VCF0112) recoded with Northeast and West as coast (1) and Midwest and South (0); income (VCF0114); union (VCF0127) recoded with (1) as someone in household belong to union and (0) no member; married or widowed (VCF0147) recoded with married or widowed combined for code (1) and all others as (0); employed (VCF0116) separating code (1) “working now” from all other categories coded to (0); church attendance (VCF0130) with the scale inverted and collapsed with the lowest category combining codes 5 “never” and 7 “no religious preference” and the highest code combining categories 1 “every week” with 2 “almost every week); and evangelical denomination indicated by a dummy coding from VCF0129 codes 123, 130 through 149 for cases from 1972 to 1988 and 2002 and VCF0152 codes 120 through 149, 180 through 199, 221, and 250 through 269 for cases from years 1990 through 2000 and 2004.

This may be an artifact of the way the union question is asked in ANES. The questions asks if the respondent is a member of a union household, so it maybe that a higher portion of women than men are actually merely married to a union member.

The Bible question was not asked in 1982, the equal role for women question was not asked in 1986 and the defense spending question were not asked in 1998. Most of the issue questions were not asked in the 2002 survey, so that year was eliminated from the analysis. All questions were recoded to vary between 0 and 1 and for 1 to indicate more conservative leanings. All questions other than abortion were asked on a seven-point scale. Abortion is a four-point scale. Questions used in the issue analysis were government guarantee of jobs (VCF0809), government aid to Blacks (VCF0830), equal role for women (VCF0834), when should abortion be allowed (VCF0838), and defense spending (VCF0843). The direction of the abortion question was reversed so that all issue questions would indicate conservatism with a higher value.

The Bible question format changed in 1990. The question used through 1990 had four categories (VCF0845), while the post 1990 question has three categories (VCF0850). The difference is that the 1990 format had two categories for the Bible being written by man: (3) “The Bible is a good book because it was written by wise men, but God had nothing to do with it” and (4) “The Bible was written by men who lived so long ago that it is worth very little today.” These two categories were collapsed as one to be equivalent to the post 1990 category “The Bible is a book written by men and is not the Word of God.” The other two categories for both question formats also have some wording changes but the most conservative always implies a literal interpretation while the second indicates that the Bible is the word of God but not everything should be interpreted literally. The direction of the Bible question was reverse, so that a literal interpretation of the Bible is the highest code, and the three categories were recoded to vary between 0 and 1.

All of the issue questions were recoded to range over a scale from 0 to 1. The median values for each of the issues are as follows: government jobs, aid to blacks and defense spending (0.51) which are the original category 4 on the 7-point scales; abortion (0.33) which is allowing abortion in cases other than rape and incest “after the need for the abortion has been clearly established”; women’s role (0.17) which is the second most liberal position on the 7-point scale, and the middle position on Biblical beliefs (0.5). The values for year run from 0 to 24, and these two extreme values are used to calculate the influence of each issue in 1980 and 2004.

No ideological change occurred for the remaining group of those allowing abortion for reasons other than rape and incest. Their ideological mean in both the 1980s and 1990s was 4.24.