By now, people who follow news about the role of religion in the 2008 elections may feel as though they have tumbled down the rabbit hole.

The Republican front-runner, Rudolph Giuliani, has consistently defended gay and abortion rights and is currently in his third marriage—a civil union that, unlike his first two marriages, was not performed under the auspices of the Catholic Church. The former New York City mayor, estranged from his son and daughter, speaks little about family and religion on the campaign trail. “The mayor’s personal relationship with God is private between himself and God,” his spokesperson has said.

Although his motivations for putting religion in the background may be very different, Republican rival and former Massachusetts governor Mitt Romney, a Mormon, echoes Giuliani’s sentiments about religion and politics. Romney recently delivered a major speech in which his main point was to outline the ways that his Mormonism would not shape his presidency: “Let me assure you that no authorities of my church, or of any other church for that matter, will ever exert influence on [my] presidential decisions. Their authority is theirs, . . . and it ends where the affairs of the nation begin.” During past political campaigns in Massachusetts, Romney showed that he could indeed divorce his Mormon beliefs from his policies: he adopted a position of tolerance for abortion and promised that he would be “a stronger advocate for gay rights” than Ted Kennedy.

Despite the fact that the latecomer to the Republican race, former Tennessee senator Fred Thompson, has positioned himself as the social conservative, he is married to his second (and, yes, much younger) wife and has spent more of the past ten years playing a public servant on television than being one in real life. A 1994 Associated Press article identified Thompson as a “pro-choice defender in a party with an anti-abortion tilt,” though he now publicly opposes Roe v. Wade. He’s been criticized for not being a regular churchgoer. When recently asked by a voter in South Carolina to speak more fully about his religious beliefs, Thompson replied, “I guess I am one of those people who feel a little bit uncomfortable getting too inside your person and personality and so forth.”

Until recently, the odd man out in the Republican field was former Arkansas governor Mike Huckabee, who seems at times to wear his Christianity not merely on his sleeve but on his forehead. (In one of his recent television ads, the words “Christian Leader” are superimposed over a picture of the candidate.) This ordained Baptist minister cites so many biblical passages in his stump speeches that Time has likened them to sermons. Huckabee has publicly called for Christians to “take the nation back for Christ,” and as a candidate for Arkansas governor in 1992, he advocated isolating AIDS patients as a public health measure.

One way or the other, the party of “family values” is clearly not what it used to be.

Meanwhile, all three Democratic front-runners are still married to their first spouse. New York senator Hillary Clinton speaks of religion often and openly in her two books, attends church and prayer groups on a regular basis and has hired a strategist—an evangelical Baptist from Mississippi—to help shape her message to Christians and other values-driven voters.

Former North Carolina senator John Edwards is fluent in the language of faith and has toured the country with a focus on the poor, the sick and the outcast, fully aware of the biblical parallels. He has said, “The hand of God today is in every step of what happens with me and every human being that exists on the planet.”

Meanwhile, Illinois senator Barack Obama told an audience last year: “If we scrub language of religious content, we forfeit the imagery and terminology through which millions of Americans understand both their personal morality and social justice.” He concluded: “Secularists are wrong when they ask believers to leave their religion at the door before entering into the public square.”

What sense is to be made of this topsy-turvy world in which day is night, left is right, and Republicans seem a lot like Democrats (and vice versa)? The befuddled observer has a new tool in the search for answers: John C. Green’s insightful book The Faith Factor. Green, a senior fellow at the Pew Forum on Religion and the Public Life and a political science professor at the University of Akron, takes a detailed look at the poll results from recent presidential elections and overturns a number of preconceptions about the role religion plays.

We are conditioned to look at the role of religion in American elections too simplistically. Over the past decade, pollsters have commonly divided the U.S. population into just five large religious groups: evangelicals, mainline Protestants, black Protestants, Roman Catholics and Jews. The problem with this approach, Green stresses, is that there is immense diversity within each category. Mainline Protestants include urban Presbyterians in Connecticut and rural Methodists in Alabama; Roman Catholics include both old-school Irish Catholics in Boston and recent Latino immigrants in Albuquerque.

Even so, Green points out, the conventional categories do have predictive power. The popular press was quick to credit fundamentalists with President Bush’s 2004 reelection, and there is some truth to the claim. After all, evangelical Protestants, spurred on by opposition to gay-marriage initiatives and constituting almost 22 percent of the total vote, supported Bush over John Kerry by 78 percent to 21 percent. In contrast, 86 percent of black Protestants chose Kerry, but they constituted only 8 percent of the total vote.

But what if we were to be more precise about the religious categories we employ? What if we also factored in, for instance, regularity of church attendance? When Green adds factors of religious “behaving” to those of religious “belonging,” new contours emerge. For the total voting population, if you attended a religious service in the previous week, there was a 61 percent chance that you cast a vote in favor of Bush. If you were an evangelical Protestant and attended church weekly, there was an 82 percent chance that you supported Bush. On the other hand, if you were a churchgoing black Protestant who didn’t attend church regularly, there was only an 8 percent chance that you voted for Bush.

What about the nature of voters’ religious beliefs? If we ask voters questions about their views of God, prayer, the Bible and evolution, then place them on a ten-point scale from the most traditional to the least traditional believers, we see tendencies that were previously invisible. In the 2004 election, Roman Catholics were almost evenly split in their support for Bush or Kerry. But 72 percent of traditional Catholics supported Bush while only 31 percent of nontraditional Catholics did the same. Add ethnicity? Sixty-nine percent of Latino Catholics favored Kerry compared to fewer than half of non-Latino Catholics. Gender and income also affect political leanings.

What Green’s detailed (and admittedly complicated) statistical analysis shows is that religion does indeed matter, but in ways that we have only begun to fathom. Equipped with Green’s more nuanced appreciation of religion’s role in voting preferences, we can map out new features in the political landscape. The fact that 50 percent of white evangelicals identified themselves as Republicans in 2004 but only 40 percent do so today should be enough to give the Republican leadership serious pause. On the other hand, that statistic may help to explain why the pro-choice Giuliani is leading polls among likely Republican voters. The rapid growth in the number of Latino Catholics in the American Southwest may have similarly profound implications for the Democratic primaries, shifting the number and priorities of voters, and may tip battleground states such as New Mexico toward the Democrats in the general election.

In a context in which secular pollsters too often set the agenda for political analysis, Green’s fascinating and challenging book offers an important corrective. Money, race, class and region are all important. But at times, Green shows us, Americans really do vote on the basis of what they believe.