4 Types Of Nones: Study Explores Spiritual Yearning In Post-Religious America
(ANALYSIS) As some of you know, Tony Jones and I won a grant from the John Templeton Foundation. It was part of the foundation’s Spiritual Yearning Research Initiative. Our specific project is titled “Making Meaning in a Post-Religious America,” the centerpiece of which is a huge survey of the non-religious with questions specifically tailored to explore the nuances of the nones in a way that we haven’t been able to do with previous survey methods.
About a year ago, we put that survey instrument into the field using Qualtrics, which is widely used in both the business and academic worlds to provide high quality data. The total sample size for this data collection effort was 15,296 Americans: we gathered responses from 12,014 folks who identified as atheist, agnostic, or nothing in particular (the nones).
We also polled 3,282 people who said that they were a member of any religious group. It was primarily Protestants and Catholics, but there were also some Muslims, Buddhists, Jews, etc. as well.
The reason for including that part of the data collection effort was simple: We need a group to compare the nones to on all these metrics. They are our “reference case.” Just so we have full transparency on this, this effort cost about $110,000. Doing good data work is not cheap!
To create our new typology of the nones, we used a bit of machine learning. In this case, it was k-means clustering. It’s a pretty simple process, really. You pick some variables that you think that might be meaningful in creating categories and then let the algorithm find commonalities in the dataset.
We chose this process because it achieves an important goal in this type of work: It’s atheoretical and (theoretically) unbiased. The computer has absolutely no idea what any of the numbers mean — it just sees a bunch of values that range from zero to one.
We didn’t want to bias the process with our own assumptions of how the nones should look. In the end, it didn’t take that many survey questions to create our categories and we will actually publish a full writeup of the variables we included and the code to recreate this analysis on other survey datasets in the future.
We landed on four categories of the non-religious for both statistical and practical reasons. The data really did say that four types of nones is the optimal number (using the elbow method). But we also know that it’s hard for people to remember a dozen different groups. You’ve got to find the sweet spot between simplification and over-simplification. We think we’ve found that by creating the four types of nones.
You can read the rest of Ryan Burge’s post on his Substack page.
Ryan Burge is an assistant professor of political science at Eastern Illinois University, a pastor in the American Baptist Church and the co-founder and frequent contributor to Religion in Public, a forum for scholars of religion and politics to make their work accessible to a more general audience. His research focuses on the intersection of religiosity and political behavior, especially in the U.S. Follow him on X at @ryanburge.