Everything I do on social media is a trade-off.

That’s the nature of data visualization. You have to take an incredibly complex and often messy social world and distill it into a rather straightforward graph that the average person scrolling Twitter can understand in five seconds or less. Not an easy task.

One of the ways in which social scientists have tried to make the concept of age more palatable is through the use of generations — Boomers, Gen X, Millennials, etc. There have been several pieces published in the last few years that have exhorted social scientists to stop using these concepts because they mean very little from an empirical perspective.

One reason is that they are completely arbitrary. Born in December of 1979 — you are a Gen X. Born just one month later — you’re a Millennial. And consider the fact that you can be born in 1981 or 1995 and are part of the same generation. That’s just nonsensical the more you think about it.

One way around is to use five-year birth cohorts.

Instead of 15 or 20 year spans for generations, a birth cohort can be just those born between 1940 and 1945. This helps to further isolate the impact of age on religious trends. I’ve been making graphs using this cohort strategy for a while now and they can provide a lot of illumination about American religion and politics.