Voters with high levels of religiosity favor Mitt Romney, while voters who are moderately religious or nonreligious favor President Barack Obama, according to a new Gallup poll.
Fifty-four percent of registered voters categorized as "very religious" said they are currently supporting Romney, while 37 percent said they would vote for Obama, in the presidential race. Among the "moderately religious," 54 percent support Obama and 40 percent support Romney. Among the "nonreligious" Obama's support is even greater, 61 percent, while Romney gets only 30 percent.
Voters are more closely split when comparing Catholics and Protestants. Protestants favor Romney, a Latter-day Saint, 48 to 43 percent. Catholics favor Obama, a Protestant, 51 to 45 percent.
Among "very religious" Catholics, though, the advantage tilts to Romney, 50 to 46 percent. The gap is even wider among "very religious" Protestants, who favor Romney 54 to 35 percent. When factoring in race and ethnicity, the gap becomes wider still. Romney has a 41 percentage point advantage (64 to 23 percent) among very religious non-Hispanic white Protestants.
Religiosity has been a reliable predictor of vote choice in presidential elections at least since the 2000 election. Those with high levels of religiosity have favored Republicans while those with low levels of religiosity have favored Democrats. This "God gap" has been a stronger predictor of vote choice than the much ballyhooed "gender gap," in which females favor Democrats and males favor Republicans.
In Gallup's poll, the "very religious" are those who said that religion is an important part of their daily life and they attend a church, synagogue or mosque at least every week or almost every week. They comprised 41 percent of the sample.
The "nonreligious," 32 percent of the sample, said that religion is not a part of their daily life and they seldom or never attend a church, synagogue or mosque. The remaining 27 percent of the sample was categorized as "moderately religious."
The poll of 2,157 registered voters in all 50 states was conducted April 19-23. The margin of error for the full sample is plus or minus three percentage points.