Clergy-lay political (mis)alignment in 2019 – 2020

We use data from the new and nationally representative National Survey of Religious Leaders, supplemented with the 2018 General Social Survey, to examine the extent to which clergy are politically aligned with people in their congregations. Two assessments of alignment — clergy reports of how their political views compare to the political views held by most people in their congregations, and comparisons between clergy and lay voting preferences in the 2016 election — yield the same findings. Clergy in Black Protestant and predominantly white evangelical churches are much more likely to be politically aligned with their people than are Catholic or, especially, white mainline Protestant clergy, who often are more liberal than their people. Contrary to media reports suggesting that evangelical clergy are now likely to be less conservative than their people, the vast majority are either politically aligned with, or more conservative than, their members.

Scholars long have studied the political views and activities of clergy.At least since Jeffrey Hadden's now classic The Gathering Storm in the Churches, one major agenda in the study of clergy politics has been to examine how closely clergy political views match the views of people in their congregations (Hadden, 1969;Stark et al., 1971;Quinley, 1974;Guth et al., 1997;Djupe andGilbert, 2002, 2003;Olson, 2009;Smidt, 2016;Malina and Hersh, 2021).A remarkable feature of this literature is the consistent finding across several decades that clergy in predominantly white mainline Protestant churches have been much less likely than clergy in predominantly white evangelical Protestant churches to be politically aligned with the people in their churches (Guth et al., 1997;Smidt, 2016).Smidt (2016, 127-128), for example, found that, in 2009, only 27% of mainline Protestant clergy were politically similar to their congregants on both social and economic issues, while 55% were more liberal than their congregants.In predominantly white evangelical denominations, by contrast, 42% of clergy were politically about the same as their congregants, while only 23% were more liberal than their congregation.
Smidt (2016) also found that the political divide between clergy and their laity increased between 2001 and 2009 within both mainline and evangelical denominations, with clergy in both denominations becoming more liberal relative to their congregations.In 2001, 32% of mainline Protestant clergy were politically similar to their congregants on both social and economic issues, and only 27% were more liberal than their congregants (Smidt, 2016, 127-128). 1Among evangelicals, 53% of evangelical clergy reported being politically similar to their congregation in 2001, while only 14% were more liberal in that year.Smidt (2016, 129) suggested that this shift among evangelicals may reflect an evangelical laity becoming more conservative rather than their clergy becoming more liberal.Higher levels of clergy-lay political alignment among evangelicals may reflect a more uniformly orthodox evangelical clergy or a greater degree of clergy influence on members' views.
As the previous paragraphs suggest, research on this subject has been focused almost entirely on predominantly white Protestants, most prominently in the Cooperative Clergy Study (CCS) series of denominational surveys (Guth et al., 1997;Smidt, 2004Smidt, , 2016)).The 2001 CCS included surveys of Catholic priests, Jewish rabbis, and African Methodist Episcopal ministers, and there are published analyses of those data, but none that report clergy-lay alignment results.Some research on clergy political activity in predominantly Black churches has examined clergy-lay ideological alignment, but research on Black church and clergy political activity more commonly examines lay approval or disapproval of clergy political activity than clergy-lay ideological alignment. 2Only Malina and Hersh (2021) offer a direct comparison of clergy-laity political alignment across the religious spectrum.Using voter registration data, they found that clergy tend to be more politically partisan than laity from the same denomination.That is, clergy from more conservative traditions are more likely to register with the Republican party and clergy from more liberal traditions are more likely to register with the Democratic party than laity from their denomination.Our analysis, which examines both clergy assessments of political alignment with their own congregations and clergy-lay voting patterns, complements Malina and Hersh's examination of voter registration patterns.
We use data from the National Survey of Religious Leaders (NSRL), conducted in 2018-2019, and the 2018 General Social Survey (GSS), to update knowledge about clergy-lay political alignment among white Protestants, and to examine clergy-lay political alignment among Black Protestants and Catholics for the first time in a nationally representative sample of clergy.

Data
The NSRL is a nationally representative survey of congregational leaders from across the religious spectrum.It was conducted in conjunction with the fourth wave of the National Congregations Study (NCS-IV) and the 2018 GSS.The GSS is an in-person survey of a nationally representative sample of non-institutionalized, English-or Spanish-speaking adults conducted by NORC at the University of Chicago (Smith et al., 2019). 3The 2018 GSS asked respondents who said they attend religious services at least once a year where they attend.The congregations named by GSS participants constitute a nationally representative sample of U.S. congregations.For the NCS-IV, NORC then contacted those congregations and interviewed a key informant, usually a clergyperson or other leader, about the congregation's people, programs, and characteristics.The religious leaders of these NCS-IV congregations constitute the NSRL sample.This design samples congregations with probability proportional to size, though we report results that weight each leader equally, no matter the size of his or her congregation. 4 The NSRL gathered data from congregational leaders between February 2019 and June 2020, primarily via an online self-administered questionnaire.It gathered data from 1,600 congregational leaders, but we focus exclusively on the 890 primary leaders in the sample.The NSRL collected data from leaders across all religious traditions, but the number of leaders from Jewish, Buddhist, Islamic, Hindu, and other non-Christian traditions in the sample is too small to make reliable inferences.Consequently, we focus on the sample of 846 primary leaders of Christian churches.The cooperation rate for this primary leader sample was 70%.Taking into account the NCS-IV's own 69% response rate, the NSRL's primary leader response rate is approximately 50%, with little nonresponse bias on key variables.See Chaves et al. (2022) for more methodological details about the NSRL.

Measures
We measure clergy-lay political alignment in two ways.We measure clergy perceptions of this alignment with an NSRL item that asked clergy, "When it comes to politics, how would you compare your own political views to those held by most people in your congregation?I am … much more politically conservative, somewhat more politically conservative, about the same as most people in my congregation, somewhat more politically liberal, much more politically liberal." We construct a second measure of clergy-lay political alignment by comparing clergy rates of voting for Donald Trump in 2016 within each of four major Christian traditions (using the NSRL) to lay person rates of voting for Trump 2016 within those same traditions (using the 2018 GSS).The NSRL asked clergy if they voted in the 2016 election and, if they voted, whom they voted for: Clinton, Trump, somebody else, or if they would "prefer not to say."The 2018 GSS asked their respondents similar questions though, when asking if the respondent voted, the GSS offered an additional "ineligible to vote" option.Also, when asking whom the respondent voted for, the GSS did not include the explicit "prefer not to say" option.To make the clergy and lay samples as comparable as possible, we restricted our analysis to respondents who said they voted and explicitly reported for whom they voted. 5 The four broad religious traditions in the NSRL are those indicated by the NCS religious tradition variable (TRAD3): Black Protestant, Roman Catholic, predominantly white mainline Protestant, and predominantly white evangelical Protestant. 6 Clergy are classified into one or the other of these traditions based on the denominational affiliation of the congregation they lead.See the online NCS codebook for details about how TRAD3 was constructed.We construct an analogous variable for the NSRL-GSS comparison using the denominational affiliation responses given by GSS respondents.Because we are interested in how clergy's political views align with those of the people they preach and minister to, we limit attention to GSS respondents who report attending religious services at least monthly.

Results
Figure 1 uses NSRL data to examine clergy-lay political alignment.It shows clergy perceptions of how their political views compare to their congregants' views within four types of Christian churches.There are several remarkable features of this figure.First, a large majority of clergy within predominantly white evangelical and predominantly Black Protestant denominations reported that their political views were about the same as most people within their congregation (74 and 70%, respectively), with the remainder roughly equally divided between feeling more liberal and more conservative than most people in their congregations.Second, Catholic priests and mainline Protestant leaders, by contrast, are much less likely to be politically aligned with their people, with the misalignments overwhelmingly in the direction of a more liberal clergy.Only 28% of Catholic priests who lead parishes and 33% of mainline Protestant head clergy say their political views are about the same as most people in their churches, and a majority-53% in both groups-say they are more liberal than most of their people.Third, consistent with prior research (Hadden, 1969;Guth et al., 1997;Smidt, 2016) mainline Protestant leaders stand out as being especially misaligned politically with their people, with over half (53%) saying they are at least somewhat more liberal, and one in five (21%) saying they are much more liberal than their people. 7 Figure 2 shows that comparing lay and clergy voting in the 2016 presidential election leads to a very similar picture as the one painted by clergy perceptions of political Note: The percentage of Black and evangelical leaders saying they are politically about the same as their people is statistically significantly higher than the percentage of mainline and Catholic leaders saying that ( p < 0.001).The percentage of mainline clergy saying they are much more liberal than their people is statistically significantly higher than that of any other group at least at the p < 0.001 level.
alignment with their congregations.Among lay Catholics and mainline Protestants who attended services at least monthly, voted, and reported their vote preference, nearly half (49%) reported supporting Trump in 2016.In contrast, clergy within these traditions were much less likely to have voted for Trump.Among leaders who voted and reported their vote preference, a quarter (24%) of Catholic primary leaders voted for Trump, and only one in six (16%) mainline Protestant primary leaders did so.Among evangelicals, a large majority of both leaders and laity voted for Trump, with evangelical leaders somewhat more likely to have voted for Trump than their congregants.About two in three (66%) regularly attending evangelical congregants who voted and divulged their vote preference reported voting for Trump, compared to eight in ten (80%) primary leaders.There are no meaningful differences in vote preference between Black Protestant congregants and clergy, as in both cases very few reported voting for Trump (1 and 5% respectively). 8 These results are consistent with Malina and Hersh's (2021) findings about leaders' and lay people's political party registrations.They found that, depending on the specific denomination, 60-80% of evangelical leaders and 15-30% of mainline Protestant leaders registered as Republican and that, while there was a positive correlation between leader registration and laity registration, laity from most traditions were collectively less one-sidedly partisan than their pastors.This is similar to our voting results.The only exception is among Catholics, where Malina and Hersh found that both Catholic laity and leaders are roughly evenly split in registration between the Democratic and Republican parties, while we find that Catholic leaders were far less likely to vote for Trump than Catholic laity.This likely reflects a measurement difference (i.e., party registration versus voting behavior), suggesting that party registration may not be as strong a proxy for voting behavior among Catholic priests as it is for clergy in other groups.
Note that the share of evangelical Protestants voting for Trump reported here is smaller than figures commonly reported in the media, which are as high as 80% (e.g.Gjelten, 2020).This is because the measure of "evangelical Protestants" used here is a categorization based on the kind of congregation the respondent reports attending, while the commonly cited "80%" figures are typically based on people who self-identify as both "white" and "born-again evangelical." 9Given debates around the validity and reliability of the self-identified "born-again evangelical" measure (Smith et al., 2018;Margolis, 2022;Smidt, 2022), and since we are concerned with the voting patterns of people who actually attend evangelical Protestant churches, we believe that the measure used here is the most appropriate.However, even if 80% of evangelicals voted for Trump in 2016, that would only match, and not exceed, the percentage of evangelical clergy who did so.By either measure, there is no evidence that white evangelical clergy are generally less politically conservative, or were less likely to vote for Trump, than their people.

Discussion
These results advance knowledge about clergy-lay political alignment in several ways.Regarding the white Protestants who have been well studied on this subject, these results show that the large political gap between mainline clergy and their people that has been observed since the 1960s continues to this day.While regular attendees in mainline congregations were split in their 2016 voting preference, a large majority of clergy in these congregations voted for Clinton, and one fifth reported that they were much more liberal than their congregants were.Moreover, there does not appear to be any substantial change in the last decade in the level of clergy-lay political alignment within the mainline, as our finding that 33% of mainline clergy are politically aligned with their congregation is comparable to the 27% Smidt (2016) found in 2009.We also find no evidence that the trend toward increasing political misalignment among mainline clergy that Smidt observed from 2001 to 2009 continued between 2009 and 2020.
In contrast, evangelical clergy generally are politically aligned with their people, though, consistent with Malina and Hersh (2021), evangelical clergy were more likely than attendees at evangelical churches to vote for Trump in 2016.Compared to Smidt's findings in 2009, we find that evangelical clergy are more likely to report being politically similar to their congregants.While Smidt reported that only 43% of evangelical clergy were politically aligned with their congregants in 2009, we find that fully 74% of evangelically clergy are politically "about the same" as their congregants.Some of this difference likely reflects differences in sampling or measurement between the CCS and the NSRL, but the size of the difference suggests that some real movement may have occurred.Notably, this trend is the opposite of what Smidt observed from 2001 to 2009, a decade in which it appears that that evangelical clergy became more politically mismatched with their congregants.Future research might assess whether the differences between Smidt's findings and ours are because the NSRL contains a wider range of evangelical clergy, measures political alignment slightly differently, or because the political alignment patterns among evangelicals have changed.Perhaps people in evangelical pews have grown more conservative over time to better match a conservative clergy.If this is the case, then it would further highlight the importance of religious leaders' ability to influence the views of their congregants (Djupe and Gilbert, 2003).
It is worth emphasizing that a large majority of white evangelical clergy are either politically aligned with or more conservative than their members.Only small minorities said they were more liberal than their members (12%) or reported voting for Clinton in the 2016 presidential election (20%).Media reports about evangelical leaders being pushed out of their congregations by more conservative laity (Wehner, 2021;Graham, 2022), reports that sometimes suggest that it is typical to find evangelical clergy who are less conservative than their people (New York Times, September 30, 2022: A2), should not be taken as documenting a common phenomenon.Only a small minority of white evangelical clergy are more liberal than their congregations.They are the exception, not the rule.
Beyond the well-studied white Protestant clergy, the NSRL allowed us to examine the political alignment between clergy and their people in other religious groups.Finding that Black Protestant clergy mainly are politically well matched with their people may not be surprising.However, it may be more surprising to find that, when they are not politically matched, Black clergy are about as likely to be more conservative than their people as they are to be more liberal.Even more surprising is the finding across both sets of analysis that Catholic priests are not particularly conservative, and in fact a majority of priests consider themselves to be more liberal than their congregation.This is despite the well documented fact that the Catholic priesthood has been shifting in a more theologically and politically conservative direction because younger cohorts of priests are significantly more conservative than older cohorts (Hoge and Wenger, 2003;Levesque and Siptroth, 2005;Sullins, 2013;Vermurlen et al., 2021).It may be that Catholic laity have become more conservative over time to match shifts in the clergy.Alternatively, perhaps the recent trend among Catholic priests toward being more conservative means that Catholic leaders were even more liberal than their laity in previous years.Further research is needed to further investigate these possible sources of change over time.
This study is not without limitations.The NSRL's assessment of clergy-lay political alignment comes only from the clergy's perspective and does not include perspectives from congregants on how closely aligned their leader's politics are to their own.While congregational leaders often have a sense of their congregants' viewpoints, the congregants might have a different perspective on how well matched they are with their leader.Unfortunately, the data do not allow us to make that comparison.Because we do not have information about the people within leaders' congregations, we also are unable to assess division and polarization within these congregations.Congregations are not always ideologically monolithic, but the survey question asks leaders to overlook the diversity of views in their congregation in the interest of presenting a single, homogenous assessment.Leaders who say they are not politically well-matched might be reacting to the presence of loud minorities who out-shout quieter majorities.Our data cannot address these issues, which must be left to future research.

Conclusion
Clergy play a prominent role in political life through their cue-giving and mobilization efforts.It is important to know the political alignment of clergy with their people because the degree of alignment provides clues about how effective clergy in different religious groups might be as political opinion leaders and mobilizers.Clergy who are well-aligned with their congregants presumably would find it easier to mobilize laity for political action.Clergy who are politically different from most of their congregants, on the other hand, may be wary of discussing political topics to avoid making political divisions salient.Documenting the extent to which clergy within various religious groups are politically aligned or misaligned with their people, and, when they are misaligned, showing the direction of that misalignment, contributes to the larger agenda of understanding which clergy are better positioned to motivate-or fail to motivate-their people for political action.
Financial support.The NSRL was funded by the John Templeton Foundation.Data were gathered by NORC at the University of Chicago.The NSRL dataset and codebook will be available from the Association of Religion Data Archives (http://www.thearda.com)and also will be available from the Inter-university Consortium for Political and Social Research (www.icpsr.umich.edu).The codebook also is available at the NSRL website (https://sites.duke.edu/nsrl).
Competing interests.The authors declare no competing interests.Notes 1. Guth et al. (1997) found that as many as three-quarters of clergy within some mainline denominations were more liberal than their people in 1989.However, this very high level of clergy-lay mismatch was not evident in each mainline denomination examined by Guth et al. (1997), and they used different measures than those used by Smidt (2016), so the Guth et al. and Smidt numbers are not directly comparable.2. On Catholic priests, for example, see Calfano (2010Calfano ( , 2014)), Jelen (2003Jelen ( , 2004)), or Smith (2008).On rabbis, see Djupe and Sokhey (2003) or Sokhey and Djupe (2004).See McDaniel (2008) on clergy-lay ideological alignment in Black churches and Brown et al. (2021) or Harris (1999) on lay approval of Black clergy political activity.3.All analyses of GSS data were conducted using WTSALL, the standard GSS survey weight variable.4. Specifically, we use the variable WT_NSRL_PRIMARY_DUP that limits analysis to just primary leaders and weights cases inversely proportionate to the number of regularly attending adults at their congregation.For more details consult Chaves et al. (2022) and the NSRL codebook and documentation available at: https://sites.duke.edu/nsrl/. 5.In total, 16% of the Christian head clergy interviewed in the NSRL preferred not to say when asked whom they voted for: 21% of Catholics, 13% of evangelicals, 27% of Black Protestants, and 8% of mainline Protestants.6.Among predominantly white Protestant denominations, the mainline versus conservative/evangelical distinction captures differences in views about the Bible and attitudes about adapting religious traditions to cultural change.Mainline denominations such as the United Methodist Church, Episcopal Church, Presbyterian Church (USA), Evangelical Lutheran Church in America, United Church of Christ, and Christian Church (Disciplines of Christ) are more liberal in these areas by interpreting the Bible in light of the historical and social conditions in which it was written and expressing openness to other religions and the secular world.Evangelical/conservative denominations such as the Southern Baptist Convention, Assemblies of God, Lutheran Church-Missouri Synod, and most nondenominational Christian congregations are more conservative in these areas by resisting historical contextualization of the Bible and viewing other religions and secular institutions as targets of conversion rather than as partners in efforts to make a better world.7. Weighted t-tests with bootstrapped standard errors show that the percentage of Black and evangelical leaders saying they are politically about the same as their people is statistically significantly higher than the percentage of mainline and Catholic leaders saying that ( p < 0.001).The percentage of mainline clergy saying they are much more liberal than their people is statistically significantly higher than that of any other group at least at the p < 0.001 level.There are not enough rabbis in the sample to report percentages with confidence, but there is a hint in the data that these leaders are somewhat more likely than mainline or Catholic leaders to be politically aligned with their people but, when they are politically different, it is mainly in the direction of a more liberal clergy.There are not enough Buddhist, Hindu, or Muslim leaders in the sample to draw meaningful conclusions.8. Weighted two-sample t-tests with bootstrapped standard errors show that the differences in vote preference between clergy and laity are statistically significant at least the 0.05 level among Catholics, evangelicals, and mainline Protestants.There are not enough Jewish, Muslims, Buddhist, or Hindu respondents in the GSS or NSRL samples to offer meaningful comparisons.9.While no measure of "born-again evangelical" self-identification exists in the 2018 GSS, there is such a measure in the 2021 GSS.Using the 2021 GSS, 79% of self-identified "born-again evangelicals" reported voting for Trump in 2016 (consistent with commonly reported figures, which range from 76 to 81%), compared to 69% of monthly attenders who go to evangelical congregations (consistent with the 66% reported in this article).The consistency in these results suggests that differences between the percentage of evangelicals voting for Trump reported here and the percentages commonly reported in the media result from slight differences in the concept being measured and not from idiosyncrasies in the GSS.

Figure 1 .
Figure 1.Clergy perceptions of their political views relative to most people in their congregations Source: National Survey of Religious Leaders, 2019-2020.Note: The percentage of Black and evangelical leaders saying they are politically about the same as their people is statistically significantly higher than the percentage of mainline and Catholic leaders saying that ( p < 0.001).The percentage of mainline clergy saying they are much more liberal than their people is statistically significantly higher than that of any other group at least at the p < 0.001 level.

Figure 2 .
Figure 2. Comparison of 2016 vote preference between clergy and laity.