Would election outcomes be different if the distribution of religious preferences in the U.S. were different? The graph below shows that the hypothetical predicted percent of two-party vote received by the Republican candidate for each election (1992-2008). With different religious configurations in the U.S., the outcomes would be different, but the differences are not as large as one might expect. Even making all voters the max religiously traditional or secular does not lead to a scenario where Republicans gain 80 or 90 percent and 10-20 percent of the vote, respectively. That said, presidential races are decided by small margins; therefore, changes in the religious configuration of the electorate would certainly matter.
Why does this matter? Both political parties have to attract individuals who are religious to their political party. But, the future does suggest that more individuals are becoming religiously tradition, but rather the younger generations are becoming more secular. It is yet to be seen if this pattern will remain. But if it does, in the future we might see the Republican Party trying to attract more secular voters. If they do not, assuming the same voting patterns remain, the future would look bad for the Republican Party.
The GOP may slowly be starting to take notice with the recent switching of positions on the gay marriage issue.
Data comes from the ANES cumulative data file and is limited to presidential election years from 1992 to 2008. Religion is measured as an interval level variable and was constructed from a series of religious indicators. Principal components analysis was used to construct the religion index. Hypothetical scenarios are estimates by changing the values of religion of the respondents, making them more or less religious. Then I counted the number that had a higher than .5 probability of voting Republican and those who had a less than .5 probability of voting Republican. The graph summarizes this data.














