Another Issue Comes Out: Gay Rights Policy Voting in Recent U.S. Presidential Elections

Abstract

Drawing from the theory of policy voting, this study examines the impact of opinions about gay rights on voting for presidential candidates. Qualitative analysis of the major party platforms and candidate campaign rhetoric from the six presidential elections held between 1988 and 2008 indicates that Democratic and Republican presidential candidates began openly expressing opposing positions on gay rights issues in 1992. Quantitative analysis of public opinion shows that, starting in 1992 and continuing through 2008, gay rights issues became more salient to the public, and opinions about gay rights began to exert a significant effect on vote choice. The study concludes with a discussion of the partisan forces that shaped the electoral significance of gay rights issues during the period from 1988 to 2008 and speculation about the role of gay rights issues in shaping future partisan electoral strategy.

KEYWORDS:

• policy voting
• voting behavior
• public opinion
• gay rights issues
• same-sex marriage
• presidential elections
• election campaigns
• candidate strategy

Notes

1. The party platforms and nomination speeches are available at http://www.presidency.ucsb.edu.

2. HIV/AIDS, the health issue intimately, but of course not exclusively, associated with the gay community shared some of the spotlight with gay civil rights issues in 1992. Both major parties had briefly mentioned their concern with HIV/AIDS in their 1988 platforms, and they did so again in 1992 (and, with the exception of the GOP in 1996, in each subsequent presidential election platform). But especially notable in 1992 was the inclusion of convention speakers who had contracted the AIDS virus. The Democrats heard from Elizabeth Glaser, who had contracted it in a blood transfusion and passed the virus to her two children, and Bob Hattoy, Clinton’s openly gay environmental advisor (Schmalz, 1992). Mary Fisher, a “rich, white, heterosexual and high-caste Republican” spoke at her party’s convention, chastising the GOP for regarding the illness as a “self-inflicted plague earned by immoral behavior” (Kelly, 1992b). These speakers helped expand the public’s awareness of HIV/AIDS as a general public health issue, rather than one specific to gays or intravenous drug abusers. Because the HIV/AIDS issue is not a matter of exclusive concern to the gay community, I have chosen not to include it in my analyses of the partisan campaigns or voter behavior. My analysis of campaign news shows that the salience of the HIV/AIDS issue diminished between the 1988 and 1992 elections and virtually disappeared thereafter.

3. Levels of agreement were high. The coders initially agreed about whether candidates were explicitly connected to an issue for 449 of the 462 articles that mention at least one gay issue and at least one candidate. The remaining 13 articles were jointly coded by consensus of the two coders. That coding exercise yielded 200 articles that explicitly connect at least one candidate to at least one gay issue. Coders initially agreed about how to code the content of 182 of them. The remaining 18 articles were jointly coded by consensus of the two coders.

4. Codebooks compiled by the Inter-University Consortium for Political and Social Research (ICPSR) provide details on sampling procedure, study design, and question wording for the ANES items used in this article. The codebooks are accessible at http://www.electionstudies.org. ICPSR bears no responsibility for the analysis or interpretations presented here.

5. It is worth noting that the public’s support for each of these rights increased steadily over the period on which my analyses focus. In 1988, 55% of voters favored or strongly favored protecting homosexuals against job discrimination. Support rose steadily to 75% favorable in 2004 and 2008. From 1992 to 2008, the proportion of voters who favored or strongly favored permitting homosexuals to serve in the armed forces increased from 60% to 79%. The notion of allowing homosexual couples to adopt children also gained more support over time: 29% of voters favored or strongly favored the idea in 1992, while 52% favored it in 2008. The percentage of voters who favored same-sex marriage increased only slightly from 34% in 2004 (the year the question was first posed in the ANES survey) to 37% in 2008; but a more notable increase in support for civil unions occurred during that period: 6% of voters favored this option in 2004, while 28% favored it in 2008. These patterns of increasing support for gay rights are consistent with evidence other scholars report (Hicks & Lee, 2006; Lindaman & Haider-Markel, 2002; Loftus, 2001; Wilcox & Wolpert, 2000; Yang, 1997).

6. Opinion about same-sex marriage might also be expected to have a notable effect, but ANES did not provide a measure of opinion about this right for election years prior to 2004.

7. The models of voter behavior specific to American presidential elections usually include the following variables: party identification, ideological orientation, perceptions of the economy, opinions about various issues, and assorted demographic controls. The models may include additional variables to represent factors unique to the hypotheses being tested or the circumstances of a particular election year. The voting behavior literature is too vast to cite here, but interested readers can orient themselves by consulting the foundational work of Campbell, Converse, Miller, and Stokes (1960) and a more recent comprehensive work by Lewis-Beck, Jacoby, Norpoth, and Weisberg (2008).

8. Mockabee (2007) examined the effects of an index of various gay rights opinions on presidential vote choice in the elections of 2000 and 2004, and Abramowitz (1997) examined the effects of an index composed of opinions about gay rights and abortion in the 1992 presidential election. Other scholars examined the effects of same-sex marriage opinion in the 2004 presidential election, either as a stand-alone item (Hillygus & Shields, 2005; Lewis, 2005; Mulligan, 2008) or as part of an index of “morality issues” (Knuckey, 2007).

9. Including an abortion rights variable provides some assurance that any effects the gay rights opinions exhibit are not absorbing the effects of opinions about moral issues omitted from the model.

10. I use multinomial logit rather than multinomial probit because the logit model is susceptible to fewer estimation problems, offers more intuitive answers to the theoretical questions raised here, requires less computational resources, and conforms well to the structure of the data. Dow and Endersby (2004) asserted that these are strong arguments in favor of the logit model. A frequent criticism of the logit model is that it assumes the independence of irrelevant alternatives (IIA). Although there are a variety of tests that purport to determine whether the assumption is valid in a given set of data, they yield conflicting and sometimes erroneous results (Cheng & Long, 2007). Current advice indicates that concerns about IIA in the context of voter choice in multiparty elections are exaggerated; under most circumstances, the logit model performs as well or better than the probit model (Dow & Endersby, 2004).

11. This sets up a conservative translation of the impact of gay rights opinions on the probability of voting for a candidate. A respondent who strongly opposes (or favors) one right is likely to oppose (or favor) the others. Taking this into account in the calculation of the probabilities creates even greater differences between opponents and proponents of gay rights than the differences reported here.

12. The comparisons are made separately for strong partisans, weak partisans, and leaning partisans (e.g., strong Democrats who oppose gay rights compared to strong Republicans who favor gay rights).

13. To create the index, each individual opinion item was recoded so that strongly oppose = −2, oppose = −1, depends = 0, favor = 1, strongly favor = 2. Values above zero are treated as positions favorable to gay rights; values below zero are treated as positions opposed to gay rights. Respondents at the zero point are not included in the analysis. Findings for the individual opinion items are nearly identical to the findings reported for the index, and the index counteracts the conservatism of the single-item estimates reported in Table 3.

14. When asked in 2004, “Do you think it should be legal or illegal for gay and lesbian couples to get married?” 72% of Republicans, 46% of Democrats, and 50% of Independents said, “illegal” (http://www.washingtonpost.com/blogs/the-fix/wp/2013/03/18/gay-marriage-support-hits-new-high-in-post-abc-poll/).

15. For example, during the 2012 election campaign likely voters consistently told pollsters that “the economy” or “economic issues” would be the factor most likely to affect their vote (http://www.pollingreport.com/prioriti.htm).

16. When asked in March 2013, “Do you think it should be legal or illegal for gay and lesbian couples to get married?” 58% of Americans said, “legal,” and 36% said “illegal.” In 2004, 41% said, “legal,” and 55% said, “illegal” (http://www.washingtonpost.com/blogs/the-fix/wp/2013/03/18/gay-marriage-support-hits-new-high-in-post-abc-poll/). In a poll conducted a few months later, support for same-sex marriage among 18- to 29-year-olds registered at 68%, support among 30- to 44-year-olds, at 54%, support among 45- to 64-year-olds, at 49%, and support among those 65 or older, at 32% (http://www.nytimes.com/interactive/2013/06/06/us/new-york-times-cbs-news-poll-june-2013.html?ref=politics). Taking partisan identity into account, 52% of 18- to 49-year-old Republicans think it should be legal for gay and lesbian couples to get married, 37% of 50- to 64-year-old Republicans think so, and 25% of Republicans 65 or older think so. The corresponding figures for Democrats in the three age groups are 73%, 73%, and 64% (http://www.washingtonpost.com/blogs/the-fix/wp/2013/03/18/gay-marriage-support-hits-new-high-in-post-abc-poll/). The usual caveats apply to comparing results from two polls conducted by different organizations at different times.

17. Same-sex marriage is probably the toughest test of support for gay rights. Washington Post-ABC News Polls conducted from 2004 to 2013 indicate that the percentage of Democratic identifiers who think it should be legal for gay and lesbian couples to get married rose from 50% in 2004 to 72% in 2013 (http://www.washingtonpost.com/blogs/the-fix/wp/2013/03/18/gay-marriage-support-hits-new-high-in-post-abc-poll/). Among high-profile Democratic elites, President Barack Obama, former Presidents Bill Clinton and Jimmy Carter, Vice President Joe Biden, former Vice Presidents Al Gore and Walter Mondale, and 49 of the 52 Democratic U.S. Senators in the 113th Congress have stated their support for same-sex marriage.

18. In March 2013, 34% of Republicans said they think it should be legal for gay and lesbian couples to get married; 72% of Democrats said so (http://www.washingtonpost.com/blogs/the-fix/wp/2013/03/18/gay-marriage-support-hits-new-high-in-post-abc-poll/). Same-sex marriage has been the leading topic in recent surveys of opinion on gay rights, but there are more traditional civil rights issues that elicit greater support from Republicans. The 2008 ANES survey, for example, shows that 66% of strong, weak, or leaning Republican identifiers supported or strongly supported laws that protect gays against job discrimination; 34% opposed or strongly opposed such laws. For comparison, 76% of strong, weak, or leaning Democratic identifiers supported or strongly supported such laws; 24% opposed or strongly opposed them.