Prop 8 and Race: More Complex Than First Reported
January 15th, 2009
Guest columnist Clayton Critcher emailed me (Jim) a few days ago with comments on my critique of the NGLTF report on Prop 8. Since he had some very pertinent observations — that the relationship between religion and race with regard to African-Americans and Prop 8 is more complex than reported — I invited him to write up a guest post for Box Turtle Bulletin.
Clayton Critcher is a summa cum laude graduate of Yale University, and is now a PhD candidate in social psychology at Cornell University, where he is a National Science Foundation Graduate Research Fellow. His research and publications include experimental work on political ideology and unintended consequences of anti-gay public policy.
The recent NGLTF-sponsored report on Prop 8 and race has reignited the discussion about the determinants of support for Prop 8. Unsurprisingly, most of the interest in this report has been on what it says about African American voters’ support for the amendment. In my opinion, the report does an impressive job of demonstrating that exit polls likely overstated Black voters’ support. Instead, just under 60% of both African Americans and Latinos supported Prop 8, while Whites and Asians were barely against it.
But what explains the gap between Blacks/Latinos and Whites/Asians? This is where things get controversial. The report suggests that the remaining gap between these ethnic groups can be explained by differences in religiosity. On Monday, Jim suggested that this analysis may have suffered from a low sample size, which can make real differences difficult to observe (statistically). Was the NGLTF report too quick to claim that racial differences were entirely explained by religious differences?
I set out to answer this question by doing my best to reconstruct the results of the poll on which the race and religion results were based. According to a comment on BTB by Jamie Grant of NGLTF, there were 149 African Americans sampled. I used other information from the report about the poll’s sample size, data about the demographics of the sample, and the study’s assumptions about the California voting population, to complete a “best-guess” reconstruction of the racial composition of the remaining sample.
Report Modification #1: Although it is true that there is no significant effect of race after controlling for religion, there was not a significant effect of race before controlling for religion. This suggests that Jim’s point about high margins of error was right on. If we could not find racial differences before controlling for religion, it is not very impressive that we cannot find them afterwards.
Because the first point in the report was that African Americans’ support had been overstated, and that it was African Americans and Latinos together that showed (modestly) more support for Prop 8 than Whites and Asians, I then dichotomized people racially. I identified each person as a minority (African American or Latino) or not. By not dividing into as many small groups, we help to bypass the sample size problem identified by Jim.
Report Modification #2: This analysis produced an unexpected finding, one that has not been considered in the discussion. The influence of race depended on whether one was religious. Among those who were highly religious, support for Prop 8 was equally high across the races. But among those who were less religious, African Americans and Latinos were more supportive of Prop 8 than Whites and Asians. The NGLTF report misses this effect because neither Latinos or African Americans by themselves show significantly higher support (in the low religiosity subsample) because of Jim’s high margins of error.
This suggests that the relationship between race, religion, and Prop 8 support is slightly more complicated than has been discussed. Being religious was associated with increased support for Prop 8, but among those who were not religious, being African American or Latino was associated with support for Prop 8.
These conclusions aside, I must say I have been confused by the intense interest in whether religiosity can “explain away” racial differences in support for Prop 8. Unless one believed that the skin color gene also produced support for Prop 8, racial differences in support for Prop 8 would have to be “explained through” some cultural factor. If that factor is religion, the question simply becomes, “Why do some racial groups show more interest in homophobic religious institutions than others?”, and I do not see why this would be any less troubling to those who seek to shift this discussion away from race. Nevertheless, my new analyses suggest that among the non-religious, an unidentified explanation for racial differences remain.