The last 30 years have seen a massive change in the American religious landscape: more and more people are deciding not to affiliate with any particular religion at all. What’s changed? Answers range from claims that religion is slowly dying out to listless and wandering millennials having no clear value system. When sociologists study these opinions over time, however, they find a number of changes and consistencies that might point to a more precise explanation.

In 2002, Michael Hout and Claude Fischer published a paper showing that the growth in religiously-unaffiliated Americans actually didn’t have much to do with individuals’ religious beliefs. Instead, political views and changes between generations had more of an impact on whether respondents identified with a religion. Now, about a decade later and with new data from the General Social Survey, they update their findings with some new trends. First, the growth in religious disaffiliation has remained steady since the late 1980s; while this group was only 8% of the population in 1990, it has since increased to 20% in 2012. At the group level, this is primarily due to young generations replacing older cohorts. But while the trend includes adults born in the 1960s, the authors emphasize that “young people who have become adults since 2000 express even less religious preference than any of the previous cohorts.”.

Using a new panel component to the GSS which tracked the same respondents over time, Hout and Fischer go on to show that respondents who shift towards liberal political views are more likely to drop a religious affiliation (and those who lean conservative are less likely to disaffiliate). The political effect dovetails with younger generations’ preferences for autonomy—the belief that it is important for individuals to think for themselves rather than obey traditional authorities. Hout and Fischer argue that growing up in a generation that values autonomy and codes conservative politics as “religious” is the best predictor of religious disaffiliation. Further, the authors emphasize that they do not see large spikes in atheism or other changes in beliefs. Rather, the most common belief among these younger generations is that there is truth in many religious traditions.

The standard story, then, is backwards. For younger generations, religious identities don’t determine values; they are an outcome of political beliefs. Thus, young people haven’t lost the faith—they have simply learned to express their values in voting booths instead of pews.