You’ve no doubt seen the reports that 37% of Alabama evangelicals were more likely to support Judge Roy Moore after the accusations from the four (now five) women who were teenagers when he was in his early 30s. The data in question came from a poll conducted by a firm called JMC Analytics between November 9 (the day the story broke) and Saturday, November 11.
After reading several references to this statistic (and sharing much twitter outrage), I took the time to actually look at the poll results themselves. It’s somewhat helpful in making sense of this and other statistics. Here are some things I learned.
First, the poll seems to reflect a party affiliation. The analysis section on page 3 of the poll results contains an underlined section describing why “Republicans should be concerned.”
Second, the makeup of this poll (compared to the undated previous poll) was skewed more heavily toward evangelicals, up from 53% to 58%. I could write a separate post about whether it really makes sense to suggest that six in ten Alabamians are evangelicals, but I’ll let it go.
Third, the top-line data on the key question on how the scandal changes votes shows that the plurality (38%) are less likely to vote for Moore. And this was before the continued drumbeat of further reports, the comments from McConnell and other senators, and the RNC pulling funding.
Fourth, the subgroup comparisons show that most other subgroups are less likely to vote for Moore than more likely. One exception is that Whites are barely more likely to support Moore (and “other” is much more so). The evangelical subgroup is the only one that stands out — but as I’ve been arguing, the overlap between whites and evangelical self-identification is pretty great (and especially so in Alabama).
Finally, the most important story in the data is the substantial move of voters toward Doug Jones. Males are split, females are +6. Jones wins every age group but the oldest. A third of evangelicals say they would support Jones. The result of these splits is that Jones is up 6 percentage points from the previous poll (whenever that was) moving from a tied race to Jones ahead.
To review, there were 575 people interviewed via landline in four regions of Alabama. About 333 of those claimed to be evangelicals. As the Hill reports, 38% of that 333 (123) claimed they were more supportive of Moore (as of Saturday). But that’s still only about a fifth of those polled.
The PRRI data from last year (that I cited in my last post) on the remarkable shift in how important morality is in determining political leaders is still nothing short of striking. But every poll doesn’t necessarily represent that same trend.