Skip to main content
. Author manuscript; available in PMC: 2014 Jun 13.
Published in final edited form as: AJS. 2008 Jan 28;114(2):408–446. doi: 10.2139/ssrn.1010098

Fig. 10.

Fig. 10

Trends in issue partisanship and alignment by region and religious attendance. Estimates from the multilevel regression models with varying intercept, varying slope, and correlations grouped by issue domain. Each box compares the correlation trends in an issue domain for two mutually exclusive subgroups. The x-axis is time (1972–2004), and the y-axis is correlation (–.05 to .4).In the religious attendance models, high attendance means twice a month or more, and low attendance means less than twice a month.