by Ryan P. Burge, Eastern Illinois University
Featured Image Credit: Debate.org
I started graduate school over a decade ago and I’ve forgotten most of the things that I read and discussed in those first few years. However, a few nuggets of wisdom stayed lodged in my memory for the long haul. One of those is depressing: the American public is woefully ignorant about what is going on in the world of government and/or politics. The seminal work that elucidates this comes from Delli Carpini and Keeter (1996) who conclude that the average American is poorly informed but not uniformed. They asked respondents a series of basic questions about government, and just 40% of respondents could answer at least six out of ten correctly. Overall, Delli Carpini and Keeter say that about half of Americans could be seen as politically knowledgeable.
While their book does a deep dive into political knowledge, it leaves religion out of the equation. Other work in social science has shown that religious people indicate higher levels of social capital and civic skills, but does that also translate into having higher levels of political knowledge? I turned to the 2016 Cooperative Congressional Election Study to find out.
The CCES asks respondents a series of questions that try to assess their general political knowledge. Here are the five that I included for this index:
- Which party has a majority of seats in the House of Representatives?
- Which party has a majority of seats in the Senate?
- Which party has a majority of seats in your state’s Senate?
- Which party has a majority of seats in your state’s Lower Chamber?
- Which political party is your governor affiliated with?
Each correct answer was counted as one point and each incorrect answer would receive zero points. Each respondent’s points were summed to create a scale from 0 (getting no questions correct) to 5 (getting all questions correct). The graph below displays the average score for a number of religious groups/denominations with 95% confidence intervals indicated by the horizontal bars. In addition the dashed vertical line is the average score for the entire sample (3.3 out of 5.0).
As can be quickly ascertained, there’s a tremendous amount of variability across the religious spectrum. The religious group that had the highest level of political knowledge were Episcopalians, followed by those in the Presbyterian Church (USA). Note that both of these are flavors of mainline Christianity. Atheists clock in at the third position, and agnostics are ranked fifth in mean political knowledge. Note that the only denomination that could be considered “evangelical” in the top 8 is Missouri Synod Lutherans. Southern Baptists are ninth on the list, just above the sample average of 3.3. Groups that score below average are Catholics, Black Protestants, and “Nothing in Particular.” It’s interesting that atheists/agnostics are some of the most politically knowledgeable and nothing in particular are some of the last knowledgeable yet they are often placed in the same general category: religious nones. Notice that both Buddhists and Muslims score poorly on this measure, as well.
The insightful Djupe offered an interesting suggestion: try to estimate political knowledge using a simple regression model and see which religious groups “out kick their coverage” and which ones underperform. I specified a regression model with just a few independent variables: level of education, level of income, and a dichotomous variable for African Americans. The graph above displays the point estimates in the orange and the actual knowledge scores in the blue. As is evident, many of the denominations do much better than the model predicts – where something about their religious affiliation adds value. At the top part of the graph, the only group that the model performs relatively well for is Jewish respondents. It is interesting to look at the bottom portion of the visualization where the estimate is higher than the actual score – where something linked to religion appears to impede knowledge acquisition. For instance there is a huge gap for Muslims –the model says that their knowledge score should be ~3.2, when they actually scored closer to ~2.4. There are many possible explanations, but perhaps the most likely because many of them are recent immigrants to the country and have not been enmeshed in the political culture of the United States. For instance, of those who identify as Muslims in the sample, 60% indicate that they are immigrants themselves (citizen or non-citizen) or that they are first generation immigrants. In comparison just 10% of Protestants and 25% of Catholics indicate such a recent migration to the United States.
I was interested to see if there was any variation among political partisanship among religious groups. For instance, Time Magazine ran a story: “Study: Are Liberals Smarter than Conservatives?” Is that really true here? The answer is clear from the graph above: no. There is absolutely no clear discernible pattern to be found. In fact, for most groups there is no statistical difference between those who identify as Republicans, Independents, or Democrats. It seems that Republican atheists are less knowledgeable but that’s tempered by the fact that there weren’t many of them in the sample. It seems that Roman Catholic Democrats have lower levels of political knowledge than the rest of Catholics, but that may be a function of racial differences with the same pattern emerging for Southern Baptists.
Academics deal with young people, a lot. One thing that we can all attest to is that some of them can be pretty clueless about what is going on in the world. There are a bunch of reasons why this could be true. Most have had very few interactions with government. Many have never voted. Lots of them pay no taxes. So, is there hope for them in the future? The answer is clearly yes. The graph above displays the mean level of political knowledge by larger religious traditions across the age spectrum. The good news is that there is an unmistakable positive relationship between age and political knowledge regardless of religious group. The average 18 year know scores ~2.2 on the scale. The average 75 year old is at ~4.0. That is relatively consistent across all religious traditions, however it’s worthwhile to note that those of no faith do break away from the pack in their late 60s and continue to gain knowledge at a steady rate.
Finally, I wanted to understand if there was some sort of connection between church attendance and political knowledge. The above graph displays mean knowledge for each attendance level broken down by several major religious traditions. There is a small negative relationship between frequency of attendance and political knowledge. However, it’s crucial to not oversell this relationship. For instance, evangelicals who never attend have an average political knowledge score of 3.5. For those who attend weekly it drops to ~3.3. An interesting aside: those of “no faith” who attend weekly (and yes, those people do exist) seem less aware of what is going on in the religious and political worlds.
In 1988 Wald, Owen and Hill published “Churches as Political Communities,” an article that argues convincingly that while most people think that churches are largely devoid of political matters, signalling still happens through political posters around election time, through bumper stickers in the parking lots and discussions in the pews after the service has concluded. While it is clear that these discussions are happening, what is not clear is that they are actually leading people to want to learn more about their political environment. Much could be said about the polarization of the American electorate, and the overwhelmingly toxic media environment in which we live. What may be the most depressing about these results is that church going does not lead to more political awareness. If churches are to fulfill their social purposes, then creating the desire in their congregants to become engaged in both the life of their church and the life of their community should become a higher priority.
Full coding syntax for this analysis is available on my Github.