First published: August 2017
What are the determinants of homophobic beliefs in the US?
“To what extent is homosexuality justifiable?” is the question asked to a representative sample of 1000 Americans recently and to which they had to pick one of the following answers: “Never”, “Sometimes” or “Always”.
The framing of the question is worth spending some time thinking about what a “Sometimes” may mean in this context. Maybe, for some, homosexuality is OK from Monday to Saturday, but not on Sundays!? Or is it maybe OK from Saturday night to the following Friday night? Or maybe do they have in mind some kind of restrictions on the number of homosexual acts in someone’s life? “Not too often…once a year, not more…”. Or maybe it is just meant at capturing all the people who don’t have extremely negative views about homosexuality but for whom this is not totally OK either.
1. A first look at the aggregate
| Is Homosexuality justifiable? | Percent |
| Never | 33.9% |
| Sometimes | 22.92% |
| Always | 43.18% |
| Total | 100% |
To this question, 33.9% answers “Never”. Adding up “Nevers” and “Sometimes”, one gets a worrisome 55%. Read carefully: it means 55% thinks homosexuality is not always OK in this survey.
2. Determinants of Homophobia
For each individual in the survey we have information about their income, education, gender, region, degree of religiosity, political orientation, race and their age. I run an ordered logit regression of the answer to the previous question onto these socio-demographic variables.
(Jump to the next paragraph if you don’t want details about the underlying variables)
All variables are categorical. Most of them are ordered (income, education, degree of religiosity from “very religious” to “not at all religious”, political orientation from “very conservative” to “very progressive”, age) and some of them are not (gender is equal to 1 for a male and 0 otherwise, do not see any sexism in this arbitrary choice ; region contains Midwest, Northeast, South and West ; race contains Asian, Black, Hispanic, Others, White), I drop the first category for each variable (and Midwest for region and Asian for race). Therefore the reference individual is a poor, below high-school, very conservative, very religious, Asian woman living in the Midwest. The results are contained in the following pdf.
The results show that female, younger individuals, Hispanic and White, progressive, not religious and Northeast individuals are more likely to be less homophobic than male, Asian, Black, Other race, very conservative, very religious and from the Midwest individuals. Regarding magnitude, the most striking variables are political orientation and religion: being very conservative rather than very progressive makes one 15 times less likely to be tolerant towards homosexuality rather than not – the odds-ratio is 7 between very religious and not all religious individuals. Interestingly, education and income are not significant. I don’t know what to take away from this, but there is still a long way to go.