Editor’s note: We’re delighted to welcome Joshua Wu as he starts a new series using data snapshots of evangelicals to frame conversations about the intersection of faith with our public social and political lives. Stay tuned for more soon.
During this election season, hardly a day goes by without candidates citing Bible verses, touting their religious beliefs, and jostling to be America’s pastor-in-chief. Regardless if public religious fervor reflects a genuine personal faith, it is undeniable that candidates strategically and intentionally test the elasticity and compatibility of biblical doctrine with their political agendas to try and gain the support of evangelical voters.
But who is the “evangelical voter”? Do evangelicals have unique political attitudes and preferences? And are evangelicals homogenous enough to be grouped as a coherent voting bloc?
Using data from the 2014 General Social Survey (GSS), a biennial representative survey of Americans’ attitudes and behaviors, I challenge two myths about evangelicals. First, I show that evangelicals do not have unique political attitudes, but are all but indistinguishable from non-evangelicals when factoring for shared demographic and religious characteristics. Second, I challenge the myth that evangelicals have homogenous political preferences. Instead, I find evangelicals of different social and cultural backgrounds hold varying political attitudes.
A first myth about evangelicals is that they hold unique political attitudes. At first glance, this myth has strong evidential grounds. For example, 73% of evangelicals believe it is important for Americans to also be Christians compared to only 34% of non-evangelicals. Similarly, while 66% of evangelicals oppose gay marriage, 67% of non-evangelicals support gay marriage.
However, these unadjusted differences do not account for baseline demographic and social differences between evangelicals and non-evangelicals. Evangelicals are older, more likely to be women, less white, and more likely to be from the South than non-evangelicals. Comparing attitudes without accounting for demographic differences cannot reveal if political attitudes are attributed to respondents’ born-again experience or simply predicted by social and cultural predispositions.
To do so, I use a propensity score matching algorithm to pair evangelicals with a non-evangelical most similar to them in their social demographic and religious profiles. (For more details about this and all of the analyses below, please refer to my annotated statistical appendix here). The results summarized in the table below reveals evangelicals are indistinguishable from non-evangelicals with similar socioeconomic, cultural, and religious characteristics along some of the political debates salient in this election cycle.
| Table 1: Evangelicals indistinguishable from non-evangelicals with similar backgrounds | ||
| Unadjusted differences | Adjusted differences | |
| Identify as Republican | 16% | -2% |
| Important Americans are Christians | 39% | 3% |
| Oppose illegal immigration | 6% | 13% |
| Support gay marriage | -33% | -4% |
| Voted in 2012 election | 8% | 8% |
| Note: All unadjusted differences between evangelicals and non-evangelicals (in bold) are statistically significant (at p<.05 level). By contrast, all adjusted differences are indistinguishable from zero and not statistically significant (at p<.05 level). | ||
For example, while it appears evangelicals are 16% more likely to identify as Republicans, the adjusted difference is not statistically significant. Similarly, while it seems evangelicals are more likely to believe being a Christian is intrinsic to being an American, the adjusted difference is not statistically significant. This is not to say that evangelicals should be Republicans or more strongly agree that God and country are inseparable. But the evidence does challenge the popular misconception that evangelicals do fall into these categories; instead, on many political issues of the day, evangelicals appear no different than non-evangelicals.
A second myth about evangelicals is that they hold homogenous political preferences. Combined with their sizeable numbers, evangelicals are portrayed as a key political bloc or constituency politicians need on their side. However, that overlooks the significant differences in political attitudes among evangelicals. For example, compared to younger millennial and Generation X evangelicals, older baby boomer evangelicals are more likely to agree Americans must also be Christians (82% versus 66%), support stronger measures to exclude immigrants living in the country illegally (75% versus 57%), are less likely to support gay marriage (24% to 42%), and are more likely to vote given previous voting history (84% to 68%).
Political attitudes also differ between more educated and less educated evangelicals. College educated evangelicals are more likely to identify as a Republican (54% to 43%), less likely to agree Americans must also be Christians (64% to 85%), and more likely to vote given previous voting history (83% to 65%). By contrast, variation in religious beliefs and practices among evangelicals do not consistently predict differences in political attitudes.
These empirical snapshots challenge the myth of evangelical exceptionalism. They show that evangelicals are virtually indistinguishable from non-evangelicals with similar social, cultural, and religious profiles. Moreover, the same generational and socioeconomic divides that shape political preferences of non-evangelicals also produce differences in political attitudes between evangelicals.
There are legitimate reasons and justifications that can and should motivate differences and perhaps even disagreement in how evangelical precepts translate to social and political priorities. But if evangelicals are not distinct from our neighbors, socio-economic peers, or cultural circles, can we really conclude that our faith shapes our social and political worldviews? Or have we, in an effort to avoid controversy or discomfort, allowed our faith to be disconnected if not entirely dismembered from our social and political worldviews?
In his 1947 treatise The Uneasy Conscience of Modern Fundamentalism, Carl Henry urged evangelicals to work out the social and political implications of their faith and acknowledgement of Jesus Christ as Savior. His thoughtful critique and forceful imperative remains relevant today.
If, as the evidence suggests, our social and political worldviews are not that unique from non-evangelicals, how articulate and effective can we be as witnesses that reflect the transformative effects of our personal relationship with Christ in all aspects of our lives? And how can we (re)balance being in the world and not of the world so that our worldviews do not fragment along the same cultural and socio-economic divides as non-evangelicals?
These are no quick or easy responses to these questions. But the challenge and importance of formulating answers should draw us more humbly before our God. It should move us to prayerfully meditate and reflect on how (if) our faith is shaping (or not) our social and political worldviews. And it should motivate a gracious engagement within and across our faith communities to determine how evangelical commitments such as common grace should be articulated in our engagement with the social and political world outside our church walls.
If we have not yet begun, this election season is a good time as any to start.
Joshua Su-Ya Wu is a husband, father, pastor’s kid, and social scientist seeking to faithfully reflect Christ in all aspects of his life. He has a doctorate in Political Science from The Ohio State University, works in analytics and data science, and writes about data analytics at Reasonable Research and the intersection of faith and culture at Stuff I Didn’t Learn in Church. He currently lives in Rochester New York with his wife and two kids, and can be reached on Linkedin or on Facebook
Thanks Joshua, what a wonderful analysis! I especially love how you used that analysis to link that analysis with a prompt for personal and collective reflection on the role that our faith plays in our social and political views.
PS, I see that you used the “born again” question to define evangelical. If you’re curious about other methods, you might find Conrad Hackett’s Scientific Study of Religion paper on Measuring Evangelicalism (pdf) to be interesting (though as a political scientist of religion, I imagine you’ve already seen and wrestled with this question).
Hi Nathan,
Thanks! Ultimately, I believe that social science should be more than an objective exercise in “knowing more,” but it should inspire and prompt reflection.
Thanks for the Hackett link. I sorta remember going through his work, or subsequent scholars who cite his work!
Ultimately, as I describe in my comment below, we must only do the best we can with what religion question(s) are available on publicly-available surveys. There are adjustments that can be made (for example, not just using the REBORN question but also church attendance, identification with a religious tradition, minimal belief in God, and belief in the afterlife to further restrict or refine an original question).
But I guess my approach is that having some operationalization (with justifiable assumptions and sensible amendments) is better than saying “we don’t have an ideal measure, so we can’t say anything about this population/topic”
Really interesting. Like Matias above, I’m curious if you know to what degree this analysis holds if you use reltrad (still using with your behavior variables–attend, god, and postife, right?) instead of reborn. Also curious if you have thoughts on reltrad vs. reborn as a starting place.
Hi Ted, great questions! Unfortunately, the reltrad variable was not available in the GSS 2014 wave. There is a religious denomination variable, but it does not include evangelical as a choice.
However, I do further exclude respondents as not evangelicals if after answering yes to the “reborn” question, they do not belong to a Christian denomination (factoring for this incomplete religious tradition variable), never attend church, do not believe God exists, or do not believe in life after death. Though my more restrictive criteria reduces the number of respondents who qualify as evangelicals (30% instead of the original 40%), I think this better captures the concept of an evangelical (as someone who needs to have some minimal evangelical beliefs and greater than zero religious practice).
Regarding on what is better, I think it depends on the application/conceptualization of evangelical. If the conceptualization of evangelical in the study is to make respondents choose either to be (mainline) Protestant or evangelical, the RELTRAD variable is better. Conversely, if the focus is on evangelical as a core evangelical belief, I would say the REBORN variable is better (with the additional qualifications of minimal religious practice and belief, as I do here). Unfortunately, most of the time (at least in my experience where I have had to use “off-the-shelf” data), the researcher does not have the choice but must best operationalize evangelical given the question(s) included in the survey.
Hope the response is more enlightening than (further) confusing!
Let us consider for a moment the possibility that your analysis is correct: Evangelical voters are really just voting along geographical, ethnic, and ideological lines, independent of their faith.
Your second to last paragraph bemoans this, “If, as the evidence suggests, our social and political worldviews are not that unique from non-evangelicals, how articulate and effective can we be as witnesses that reflect the transformative effects of our personal relationship with Christ in all aspects of our lives?” The suggestion seems to be that Evangelicals really should be pursuing some common political agendas together, and the failure to do so is evidence of some pathology.
I fear this assessment encourages misguided efforts to organize the Church under a single political agenda (like the Religious Right or Moral Majority), without actual reflection into the ethnocentric (primarily white) and culturally exclusionary (often anti-immigrant) roots of many of these platforms. I am reminded of Jesus’ disciples, which included at least one tax collector (that collaborated with Rome), one Pharisee (who saw Roman currency, with its graven image, as idolatry), and at least one Zealot (who advocated violent opposition to Rome). Somehow, following Jesus did not lead to unity in this political issue, though Jesus unsettles all of them by dodging a direct question about taxes, when asked by Pharisees and Herodians, and by offering a spiritual revolution instead of the political revolution for which they hoped.
Let me present another option to your assessment. Perhaps the real problem is not this failure to have a common agenda, but the persistent claim that Evangelicals make that their personal political agendas are a direct consequence of their faith. The conflation of political agenda with spiritual devotion is corrosive, directing our witness to human ideas, efforts, and movements instead the world turning significance of Jesus.
Of course Jesus should affect our politics, but we should not expect this to be evident in a common political agenda. Rather, I think, He unsettles the central importance of politics in our world, and unsettles any worldview rooted here too. To this end, I point to James Hunter in To Change the World for a more comprehensive exploration of that notion.
What are your thoughts?
Hi Joshua, thanks for your thoughtful reply!
The implication of my analysis is not that evangelicals should have a common set of political priorities. Instead, I believe that evangelicals should be united in the process of ever-struggling and seeking how our faith shapes and is manifest in our social and political lives.
Given genuine disagreements among evangelicals about theological issues (for example, predestination, baptism, the role of women, etc), so the intentional articulation of our faith into politics should lead to different evangelical preferences on alleviating poverty, extending social justice, the use of force, and so forth.
This is not to say that evangelicals with different theological perspectives cannot have some shared political attitudes and preferences. However, I would not expect an emergence of a common or comprehensive “evangelical” political platform or agenda.
Where I may disagree with you is how (if) an evangelical should participate in politics, beginning within his or her local community. If I understand Hunter’s argument (disclaimer: I did not read the book myself, but a few reviews of the book from ESN and Christianity Today), he is skeptical if not dismissive of the role that evangelicals can play within the political realm. Here I would disagree. While there are many reasons why evangelicals should not be involved in politics, there are also strong reasons why they/we should.
Again, not every evangelical should or need be involved in politics the same way. I agree that politics is of greater importance among non-evangelicals (and perhaps some evangelicals) than it should, but I would counter that means there are more opportunities for Gospel infusion and engagement than would otherwise be possible if politics remained an aloof social exercise that affected only a few in limited areas of life.
I will actually be exploring if/how evangelicals can (or should or perhaps even must) be political in my next post, so would look forward to your thoughts again then!
Thanks for the reply. I’ll look forward to reading your next post.
Regarding Hunter, I think his ideas are important to engage here. I read him differently than you.
In my reading, it is not that we should withdraw from politics entirely. Rather Hunter is making an observation about the pitfalls of focusing on changing the world through political influence. The US is so strongly politicized that it reorders how we as Christians think about the world. The world is portrayed as a grand struggle between the right and left, conservatives and liberals, Republicans and Democrats, where everything of significance turns on political battles. We reflect this over-politicized culture when we, as the Church, when political agendas in the church take outsized importance.
Hunter’s point is that political effort is not effective, because political activism isn’t really how change happens. Rather change happens through networks of thought leaders working together, and incarnate communities. Moreover, most Christian engagement in politics merely reflects the false liberal/conservative mythology, acting as if political parties are the real standards of public morality.
I would add, this political syncretism—that we see in the Christian right and the Christian left—sets Christians down a theologically questionable path. Does God’s plan in this world really need political power? Or faith does really well when it is out of power, and our faith withstands even murderous persecution. The power of God is more clear in political weakness. The truly compelling thing about our faith is that it makes sense even without political power.
I think one call of Jesus in our world is to reject this sharp divide between right and left as being the definitive moral guide. One thing I love about Intervarsity and my local church is that I run into people across the political spectrum, and we are still family together. This is a bizarre reality, not usually encountered in our world, or even (unfortunately) most churches.
This, obviously, is a much longer discussion, and maybe you will touch on this in your next post.
A very interesting study. A few questions do arise for me, though:
1) At times I regret not taking a class in statistics, particularly given the extent to which they are referred to in political questions (and, of course, keeping in mind Mark Twain’s dictum about lies and statistics). Maybe I need to look for some Khan Academy videos…. But I find it hard to understand why an unadjusted difference of 6% regarding opposition to illegal immigration is statistically signficant, but the larger 13% adjusted difference is not. The unchanged 8% difference for those who voted in the 2012 election raises the same question.
2) You compare evangelicals with non-evangelicals who are as close as possible in other characteristics, and see little difference in political positions. There are, it seems to me, two alternative explanations: a) Your explanation that their faith commitment isn’t changing the way they (we) think politically, or b) evangelicals are having an effect on non-evangelicals and convincing them to respond politically in similar ways. In communities where there is a strong ethos of religiosity (as seems to be the case in parts of the rural South, for instance), one might expect the latter to be significant. One might call this a chicken-and-egg question.
3) Who are the “non-evangelicals” that you’re comparing with? Are they traditional Catholics, cultural Catholics, mainline Protestants, “nones”? These different groups all supposedly show considerably different political positions, at least on certain issues, from the hypothetical typical evangelical, and sometimes from each other. If “nones” with similar cultural and socioeconomic characteristics are similar politically to evangelicals, that’s significant news (though again the question is whether it means that evangelicals are being drawn to secular positions, or that nones who come from evangelical-type backgrounds are maintaining their politics even while jettisoning their faith). If devout Catholics and mainline Protestants are similar to evangelicals, that may mean that there’s far less here than seems at first, and that all of them are acting on their common Christian faith.
Are there other political issues where differences remain after adjustment? Abortion is one issue that comes to mind that has a very strong religious connection. Why didn’t you include that in your analysis?
Hi Kelvin,
Thanks for your thoughtful replies! Let me reply to each of your questions in turn.
1) The adjusted differences are not statistically significant because the standard deviation (or how much variation are there in average differences) is greater than the variation in the unadjusted difference. Tests of statistical significance takes that into account, which is why adjusted differences are not statistically significant and there is a greater 5% likelihood that the average difference we find could have happened by chance.
2) Your argument that perhaps there are no differences because evangelicals have already so influenced their neighbors is interesting! While hopeful that you are right (and definitely aware that there are examples where this has happened), I am skeptical for two reasons. First, until the past few decades, evangelicals have not been very active in engaging the political sphere (choosing what some call a “fundamentalist” approach of withdrawal from the political sphere). Having a more inward focus makes it unlikely that evangelicals were persuading or seeking to change the culture around them.
Second, there may be a sorting effect where people move to communities where people already are similar to them (or form new communities of like-minded people). So without more analysis, it can be difficult to determine if evangelicals did change their neighbors’ minds, or if they became neighbors with those they are likely to agree with (already, before having a persuasive or culture-shaping effect).
3) Your point about non-evangelicals is a very interesting question as well. There is a lot of disagreement about who an evangelical is (and is not). Please see my online appendix for some more description of who are considered evangelicals in the analysis here (https://joshuaswu.wordpress.com/2016/02/14/myth-of-the-evangelical-voter-annotated-statistical-appendix/).
4) Regarding other issues, I picked 5 issues or dimensions that were the most salient during this election year and dominant themes on the campaign trail. There are definitely other issues (like abortion) that I could have analyzed (and may in a future post actually).
Thanks again for your questions!
Thank you, and thank God, for such a scholarly, readable, polite, thoughtful, articulate, and inspiring contribution! Also, thank you to the equally excellent comments. I echo the experience of being in a politically diverse congregation that, nonetheless, is a family together in the Lord, who is worshiped as sovereign and mighty above all else, and each believer as a broken, but redeemed, child of God in this world. We might consider the political spectrum as healthy, and perhaps even on the level of callings (and spiritual gifts) for different ministries and audiences, in order to meet the spectrum of moral and institutional evils in this world. God, not we, is in charge, but we each have a calling and duty to respond to the Lord’s leading.
Hi Joanne,
Thank you for the kind words! Indeed, to God be the glory! It is great to hear that you are able to attend a church that has social/political/cultural diversity. I am curious, how does politics come up (if at all?), and did you pick a church like this because it was diverse (or did you go there and “happen” to find out that it was diverse?)