Cohorts and Support for Marriage Equality

• In Event: Sexuality and Gender Identity: Individual Well-being and Public Opinion

Abstract

Support for same-sex marriage has doubled since 2004, when 13 states passed constitutional amendments banning it. Although people changing their minds explains most of the increase, cohort replacement accounts for one-fifth of the shift. In general, those born ten years later are 5 percentage points more likely to favor marriage equality. Although cohort patterns are roughly linear, however, those born in 1965 are no more likely than those born in 1952 to favor same-sex marriage.

Using data on over 200,000 respondents to over 200 surveys, this paper will document that plateau and explore its meaning. Preliminary findings indicate that this plateau holds not only for same-sex marriage in the past decade but in the late 1990s and early 2000s. The plateau also holds for attitudes toward homosexual sex and civil liberties for “admitted homosexuals” in the General Social Survey (GSS) and for employment rights and sodomy laws in other surveys. In contrast, the GSS shows a different plateau for approval of premarital sex and little evidence of cohort change for approval of sex among teenagers or of abortion (though cohorts born since 1943 are all about 10 percentage points more likely to support legal abortion for any reason than those born before 1937). That is, the pattern seems to hold for LGBT issues generally but not for other issues about sexuality or reproductive rights.

The paper will also explore reasons for the pattern. People born between 1952 and 1965 turned 18 during the early years of the gay rights movement, so their adult attitudes may have been shaped in a period that politicized homosexuality and LGBT rights in unique ways. Very preliminary analysis also suggests that people’s self-described conservatism increased in birth cohorts from 1951 to 1965, countering a general downward from at least 1940 to 1994, suggesting that the politics of the late 1960s to early 1980s may have shaped more general political attitudes as well. The analysis will also test the effects of cohort differences in education, religion, and race/ethnicity as possible explanations for the plateau and whether a similar plateau exists within each group.

Author

• Gregory B. Lewis, Georgia State University