Measuring church attendance is tricky. People have a tendency to either exaggerate or not remember things very well. In one famous study, researchers surveyed a sample of respondents from an isolated geographic area and asked them if they had attended a religious service in the past seven days. Unbeknownst to the respondents, the researchers had also literally gone to the churches and counted the number of people actually sitting in the pews, estimating that about half of people who say they were in church weren’t actually in church.

So people aren’t super accurate (or maybe honest?) when asked about their religious behavior. However, to some extent that is a weakness of all survey research that forms the bread and butter of the social sciences and has remained relatively unchanged for the past half-century, and there’s little we can do to get around it — until now.

By now, most people know that your smartphone is following you wherever you go; your smartphone uses your location data for apps and sells that data to advertisers to be able to push location-based ads. Cutting-edge researchers are starting to use these data to measure not what people claim to do, but what they actually do, using location-based data to study myriad topics ranging from depression to mass protests to how Republicans were less likely to adhere to government stay-at-home COVID-19 mandates.

Now, in one of the first studies measuring religious behavior using smartphone data, University of Chicago economist Devin Pope studied the religious attendance and other behaviors of over 2 million Americans right before COVID-19 in 2019. With the little trackers almost everyone carries in our pocket, Pope was able to get the real picture on religious attendance that is obscured by self-report surveys.

The pandemic changed many churches. But did it change church attendance?

Of course, not everybody brings their phone to church. Pope anticipated this problem, so he compared smartphone location predictions in areas with “ground truth” numbers to see how off he was. For example, his smartphone data suggested that about 26 million people visited Six Flags during the window, while Six Flags themselves reported an attendance of 23 million — so not razor-sharp precise, but probably much more accurate than the numbers we get from self-report that can be off by half. (He also performed a similar analysis on AMC Entertainment, McDonald’s, Olive Garden, Planet Fitness, Lowe’s, Target, Home Depot and Walmart. Plus, for questions involving comparisons between groups and not baseline averages, the leave-the-phone-at-home effect shouldn’t matter, since there is no fundamental reason to think that, say, Catholics bring phones to church any less than Protestants.)

There are of course young children who attend church without smartphones and individuals who are homebound or otherwise physically unable to attend church, but consider themselves regular religious participants. Latter-day Saints, for example, will sometimes go to the homes of such individuals to administer the sacrament to them. Individuals from various faith traditions have long watched church services on television. Indeed, the term televangelism was around long before the advent of streaming church online.

So what did Pope find in his groundbreaking study?

  • About 1 in 345 Americans are weekly-attending Latter-day Saints (again with some error bars). While there are many, many more Catholics, their activity is so much lower that there are actually more weekly attending Latter-day Saints in the pews every Sunday than there are frequent Catholic attenders. However, when non-frequent attenders are included there are still about three times as many Catholics overall in the pews on any given Sunday as Latter-day Saints.
  • Overall, about 5% of people attend religious services weekly. Even allowing for some undercounting, this is much less than the 22% of people who self-report the same.
  • The location data was able to pinpoint which neighborhood each parishioner was from, allowing Pope to derive wealth statistics for faiths. He found that Latter-day Saints and Jehovah’s Witnesses had the most economic equality within their congregations, while Muslims had the least.
  • Not a lot of worshipers are deterred by bad weather. “For every 200 people going to church, only 1 is deterred when it rains on the day of the service.” Also, about 5% of people who would go to church on a greater than 35 degree day would not go church on a less than 25 degree day.
  • About half as many weekly religious service attenders visited strip clubs as never-attenders (.9% vs. 2%, respectively).
  • Of all the faith groups, Latter-day Saints are the least likely to visit strip clubs, liquor stores and tobacco stores, but are actually somewhat typical when it comes to visiting casinos.
  • Jewish synagogue attenders visit gyms much more than any other faith group.

There are other findings, and those interested can read the whole study here.

Demographer Lyman Stone and others have raised questions regarding the accuracy of the smartphone data, given that cell towers and cell reception still remain unreliable in certain areas and buildings. People can turn off their smartphone or put on airplane mode. Others pointed to religious traditions that reject technology altogether or avoid using smartphones during sabbath, such as Orthodox Jews, a point Pope concedes in his study. Again, there are complexities, hedges and caveats, but there are in every study, including those that use more traditional methods. The use of smartphone data is opening up research topics that would be difficult with the old-time pen and paper approach, and new ways to study not only what we say we do, but what we actually do when we think no one is watching.