Tuesday, July 18, 2006

How important were ‘moral values’ issues in the 2004 election?

Here’s an interesting study by D. Hillygus and T. Shields. It’s on the importance of ‘moral values’ in determining voter preference for Bush vs. Kerry in 2004. The study used data on 2,800 individuals from a post-election survey designed to represent the entire US population. The study looked at the importance of five issues. The respondents rated each of the following five issues on a scale of 'strongly support' to 'strongly oppose' the Bush administration's position on the following, on a five point scale, with a 'neutral' choice allowed:

--Iraq war
--multilateral approach to terrorism
--jobs are secure
--abortion rights
--gay marriage

The study took into account each respondent’s sex, race, marital status, income, political ideology (conservative vs. liberal), and political affiliation (D, R, or Ind), So, the effect on each person's vote due to their position on each of the five issues can be estimated, keeping all the individual characteristics constant. The study estimated models for the whole country, the South, for independents only, battleground states only, and states with gay marriage initiatives only.

Below are the main results:

--gay marriage less important issue than Iraq war, terrorism, jobs and abortion,
--party loyalty determined vote more than any one issue,
--Republicans were more loyal to party than Democrats,
--Iraq war and terrorism position produced more defections from party than moral values issues,
--Iraq war and jobs were more important among independents than those with party ID.

The Iraq war, and job security were more important than moral values for every model estimated. Terrorism was more important in every model except for the gay marriage initiative states, where it was roughly tied for third most important issue. Abortion was more important in the South, but still less important than Iraq, terrorism and jobs.

The authors’ make an interesting observation regarding people without strong beliefs on the five issues: Bush generally found support among those who were neutral on the issues (except for those neutral on gay marriage, who usually voted for Kerry). They interpret this to mean that Bush received a second chance from people who did not have strong views on the major issues.

Voterlinks comments: This study is more reliable than looking at percentages from exit polls. The term ‘moral values’ is defined (‘abortion rights’ and ‘gay marriage’). The relationships to voter characteristics can be studied. We can see the relative importance of each issue (rather than just naming the most important one). Two main drawbacks. This was a post-election survey, so respondents truthfulness is always a question (truthfulness is always a question in these kinds of surveys even if the respondents don't intend to deceive). But it is hard to see what could be better, unless there is a data set that tracks individuals over time from beginning of campaign until right after the vote. The other big limitation is that it is the survey is limited to voters, so effects of issues on turnout cannot be studied.

Voterlinks has found a few more studies on moral issues and the 2004 election and will report on those later.

DS Hillygus and TG Shields. Moral issues and voter decision making in the 2004 presidential election. PSOnline (PS: Political Science & Politics) April 2005; 201-209.
http://www.apsanet.org/imgtest/PSApr05HillygusShields.pdf
(note: may require subscription, depending on from where you access.)