Which Identity is More Important: Race, Gender, or Religion?

Or are we just trying to put smoke in a box?

There’s this line from Walt Whitman that I think about a lot when I am doing survey work, it’s from Song of Myself, 51:

Do I contradict myself?

Very well then I contradict myself,

(I am large, I contain multitudes.)

There’s this inherent tension in doing survey research. We are trying to get people to explain their thinking about things. I don’t know if that’s always possible. As many of you know, I worked on a book project called The Great Dechurching with two pastors - Michael Graham and Jim Davis. We asked folks a whole bunch of questions about why they stopped attending religious services and why they might come back. In writing that book we would always come back to the possibility that people themselves don’t fully understand why they do the things that they do. They contradict themselves. They contain multitudes.

I’ve described it like trying to put smoke in a box. Trying to create order from chaos. Trying to make sense out of randomness. But maybe that’s all that we can do, really. The alternative is worse, in my estimation. It’s just a whole bunch of blind conjecture based on anecdotes and personal experiences.

I was thinking about that a bunch when I was poking around the Cooperative Election Study’s search tool in search of questions about religion. I found this nice little module that was fielded by the terrific Lilliana Mason at Johns Hopkins. She asked people, “How important are each of the following to your identity and how you think of yourself?” Then they were given three options to respond to: your gender, your race, and your views about religion. It’s definitely an angle that I hadn’t really considered much - asking people to reflect on how they construct their worldview. Maybe that’s trying to quantify smoke in a box, but it’s definitely worth some exploration.

Here’s how the entire sample answered those questions.

I don’t know what I was expecting to see in these results before I generated the bar graph, but I was surprised to see how much gender really popped in these results. Half the sample said it was “very much” important to their identity. For those wondering - there was a huge gender gap on this question. Among women, 58% said it was very important compared to just 41% of men. (That’s probably why this result was so interesting to me - I’m a guy, I don’t think about that very much.)

About 30% of folks said that their race was very much important but then 25% said it was not at all important. Of course, the respondent's race is crucial here. For white folks, just 19% said their race was very much important compared to 55% of non-white respondents. I think we can see a pattern forming here - white males don’t see their race or gender being that central to their identity.

What about the religion question? What I like about this is how it’s phrased. It’s not “Your religious faith” it’s “Your VIEWS about religion” which means that this question is still helpful to understand the perspective of atheists, too. In the entire sample, 36% said it was very important. But there was a huge gap between the religious and non-religious. For the nones, 40% said it was “not at all important” compared to 10% of the religious sample. So, despite a smart question design, it still seems to be excluding the non-religious. Or, the non-religious just don’t have strong feelings about religion.

Of course, I know what you all wanted to see - how that religion question broke down based on religious affiliation. So, here’s what that looked like.

Man, this graph is a great example of the peculiarity of American evangelicalism. In this sample, nearly eight in ten evangelical Protestants said that their view of religion was “very much” important to their identity. That was forty points higher than any other religious group in this survey. For non-evangelical Protestants, only 30% said it was central to their identity and it was 38% of Catholics. You can just see how much difference there is between evangelicals and other types of Christians. For evangelicalism, it’s an all encompassing religious framework. That’s just not the case for the mainline or Catholics.

I couldn’t do any analysis of groups like Buddhists, LDS, or Hindus here, though. I know someone wants to ask. These modules have a total sample size of 1000 folks, so I only have 11 Mormons and 2 Hindus. But what about the non-religious? You can see a clear divide between atheist/agnostics and nothing in particulars. Nothing in particulars just don’t care about religion, while there’s evidence that some atheist/agnostics do have some strong feelings. Nothing in particulars are just apathetic about the whole thing.

I am guessing that partisanship may play a role in how these other identities can shape responses to these three questions. Let me test that out.

Well, those results surrounding the gender question just aren’t that much different. Which is interesting. For both Democrats and Republicans, about half of them say that their gender is very much important to their personal identity. I don’t really know what to do with that, honestly. For race, there is a bit more variation, though. For both Democrats and Independents, about two-thirds say that their race is central to who they are as a person. That’s about ten points higher than it is for Republicans. I’m guessing that’s because the Republican sample is a whole whiter.

But on the question of religion there is definitely a huge gap. For Democrats and Independents, religion is central to about a quarter of them. In fact, for both groups the “not at all” and “very much” percentages are nearly the same. That’s not true for the Republicans - half of them indicate that “their views about religion” are very much important. When I write about the God Gap, this is exactly what I’m talking about.

I thought that the generational differences were helpful to consider, too.

The one huge finding that jumps out for me is that Generation Z doesn’t seem that concerned with identity politics compared to Boomers or Gen X. For older folks, about half of them say that their gender is key to their identity, but it’s about ten points lower for Generation Z. When looking at the question about race, there’s only one generation where the “not at all” share is clearly larger than the “very much” share - it’s Generation Z. It looks like young adults are choosing to not define themselves based on race and gender. Which is super interesting to me considering they have grown up in a world that placed a lot of emphasis on questions related to diversity, equity, and inclusion. It seems like they are reacting to that discourse pretty significantly.

Now, the religion result is definitely intriguing to me. It makes perfect sense that Gen X places less importance on religion compared to Boomers and Millennials are lower than both of them. That just comports with what we know about religious decline across generations. But then Gen Z is this puzzle. They are actually slightly more likely to say that religion is “very much” important to their identity compared to Millennials (29% vs 26%). Maybe a bit of religious revival there? But then, the share who chose “not at all important” is right in line with prior generations. Not a super strong signal there one way or the other.

But let me close this with a bit of a coding exercise. I wanted to try to create a rank ordering system between these three identities. For instance, how many people clearly think that religion is more important to their identity than their race or gender? Or what percentage of folks rank all three of them exactly the same. Just for a bit of explanation - if someone said that their gender, race, and religion were all somewhat important to them that would be an “all tied” result. If they put religion as “somewhat” and gender and race as “a little” then in that case religion would be the most important.

The most popular combination was a person ranking all three as just as important. That was true for about 28% of the sample. It didn’t vary too terribly much by partisanship, either. The next most popular response was gender being the most important - that was 20% of the sample. Clearly this was chosen with less frequency for Republicans than Democrats or Independents. I was very surprised to see that just 4% of the sample clearly saw their race as the most important of the three. That number was higher for Independents than Democrats or Republicans. Meanwhile, just 13% of the sample ranked religion by itself as the most important identity. That was seven points lower than gender! Of course, Republicans scored highest on that metric. Which should come as no surprise based on the prior analysis.

What combinations show up a bunch? Race and gender was fairly popular, but notice the huge partisan gap on this one. While 24% of Democrats said race and gender were key to their identity, it was only 12% of Republicans. I think this is pretty strong evidence of why the Democrats seem to be constantly talking about identity politics. Meanwhile, gender and race did show up with quite a bit of frequency, too. However, the partisan gap was reversed with 20% of Republicans choosing this combination compared to only 11% of Democrats.

People contain multitudes. They aren’t easily sorted into simple binary categories. That’s the beauty of the human experience. People contradict themselves. They constantly change how they think about themselves and the world around them. Frankly, they don’t make much sense sometimes.

But it’s my job to put smoke in a box. I wouldn’t have it any other way.

Code for this post can be found here.