16

______

Patterns of In-Person Worship Service Attendance during the COVID-19 Pandemic

The Importance of Political and Religious Context

BENJAMIN R. KNOLL

“Easter’s a very special day for me. Wouldn’t it be great to have all of the churches full? You’ll have packed churches all over our country. I think it would be a beautiful time.” This optimistic prediction came from President Donald Trump on March 24, 2020, in an interview with Fox News. Earlier that month, spiking levels of the novel coronavirus pandemic had prompted a critical mass of states and localities to issue shutdown orders and for public schools to transition to online learning. At the time, many Americans (including President Trump) were hopeful that the lockdown measures would be short-lived and that Americans would be able to get back to their regular lives within a few weeks, perhaps even by Easter on April 12.

Of course, these early hopes were dashed as the spread of COVID-19 infections only intensified in the coming weeks and months. Public health experts consistently warned against large, in-person gatherings, religious gatherings included. States took varying approaches to regulating religious gatherings. An analysis by Jenkins (2020b) showed that in April, seven states exempted religious services from the state’s shutdown protocols while another ten had limited exemptions. Another eleven states, in contrast, deliberately did not exempt religious gatherings from shutdown orders. Given these varying approaches, what did individual worshippers do? Did they continue to worship in person, or did they stay home and take advantage of online religious services? Who were the ones who decided to attend in person against the guidelines offered by public health officials, and how did they differ from those who decided to stay home? These are the central questions of this chapter.

These questions are important because they speak to longer-term trends in American religious behavior. Even before the COVID-19 pandemic, religious service attendance patterns had been in decline for decades. Research has shown that regular participation in religious congregations is strongly habitual and that significant life changes (marriage, relocation, etc.) are the most likely time when these habits can be interrupted, leading adults to make changes in their religious patterns (Putnam and Campbell 2012, ch. 6). The COVID-19 pandemic was a disruptive event in many ways, leading the vast majority of regular worshippers to suspend in-person participation for at least a few weeks in April 2020. While in-person worship experiences decreased as a result of the pandemic, other religious behaviors, such as frequency of prayer, increased substantially (Bentzen 2020; Dallas 2020; Dein et al. 2020). Religious behavioral patterns during the COVID-19 pandemic, especially attendance at religious services, might offer clues about longer-term trends in American religious patterns. In what follows, I draw on available data to track religious behavior during critical periods of 2020.

Late March

Before the COVID-19 pandemic hit, 31 percent of Americans reported that they attended religious services once a week (or more), according to extensive polling by Pew Research, which combined results of multiple surveys from 2018 to 2019. Another 13 percent said that they attended a few times a month, and 20 percent, a few times a year. Those who rarely attended made up 17 percent of Americans, and those who never attended comprised another 17 percent (Pew Research Center 2019). Our baseline, then, is that before the 2020 COVID-19 pandemic hit, about one-third of Americans said they attended religious services on a weekly basis while about half said that they attended only occasionally or not at all.

The March 2020 survey fielded by Djupe, Burge, and Lewis (see Chapter 1) polled 3,065 Americans about their religious activities in the first few weeks of the pandemic shutdown, specifically between Monday, March 23, and Friday, March 27. At this point it had been a full week since the Trump administration’s guidelines to limit gatherings to ten people or fewer and included the first weekend that worshippers would have had the opportunity to attend religious services after these guidelines were released. When this question was asked, respondents were likely reflecting on their activities from the previous weekend (a week after the ten-person gathering-size guidelines) and their plans for the upcoming weekend (two weeks afterward). Further, these questions were given only to survey respondents who indicated that they attend religious services at least “seldom” (excluding the 24.8 percent who said they “never” attend religious services aside from weddings and funerals).

The March survey revealed that of the three-quarters of American adults who attend religious services, whether routinely or occasionally, only 12 percent said that their congregation was still open for in-person worship services. Among those whose congregations were still open, roughly one-third (37.5 percent) said that they were still attending in-person services. Interestingly, 18.3 percent of those whose congregations had closed reported that they were still worshipping in-person. While some of this is likely overreporting (common with surveys of church attendance; see Presser and Stinson 1998), it could also mean that some worshippers chose to attend another congregation that was still open during the last two weeks of March. Overall, then, 20 percent of American worshippers reported that they were continuing to attend in person during the last two weeks of March. Who were they?

In terms of religious identity, those with the lowest rates of in-person attendance were mainline Protestants (16.4 percent) and those who do not claim any religious affiliation (16.6 percent). Those with the highest rates of attendance in late March were, interestingly, the “other” category—non-Protestant/Catholic/Jews—at 27.8 percent.¹ While much of the public focused on the public refusal of some evangelical congregations to comply with the shutdown orders, the March survey showed that they were only slightly more likely (17.6 percent) than mainliners and “nones” to report still worshipping in person and lower than either Catholics (22.9 percent) or Black Protestants (19.6 percent). Given the approximately 2 percent margin of survey error, though, evangelicals were statistically indistinguishable from some other major religious groups.

Politically speaking, survey respondents were asked to rate, on a scale of 0 to 100, how supportive they thought that their congregation’s “head clergyperson” was of Donald Trump. Those in congregations where the pastor, priest, or other leader was perceived by the respondent to be strongly supportive of Trump (a score of 67 or higher), one-quarter (26.5 percent) were still worshipping in person compared to only one in ten (11.1 percent) among those who perceived their clergyperson to be least supportive of Trump (a score of 33 or lower).

March survey results also revealed a somewhat counterintuitive finding regarding the issue of religious freedom. Much of the litigation surrounding the shutdown orders throughout 2020 centered on religious freedom arguments, arguing either that a prohibition on in-person religious worship violated the First Amendment or that houses of worship were unfairly restricted compared to other businesses or venues. Survey responses, though, revealed no statistically significant difference in worship patterns between those who said that their clergyperson had addressed the topic of religious freedom that year (21.8 percent) and those who had not (19.3 percent).

In contrast, the strongest contextual variable seems to be direct messaging from congregational leaders. One-quarter of respondents (25.7 percent) said that their congregations had encouraged them to continue to attend in-person worship because of the virus. Among them, a full half (50.8 percent) continued to do so compared to only 10 percent of those whose congregations had not encouraged them to continue to worship in person. Similar to the findings of other research on the effect of religious messaging from congregational leaders (see Knoll and Bolin 2019 for a review), I find here that worshippers were very responsive to the recommendations of their congregations about whether to continue to meet in person, at least initially.

While there was a great deal of uncertainty at the beginning of the pandemic about how the virus affected different communities, one of the more reliable patterns was that older people were more susceptible to contracting COVID-19 and experiencing worse symptoms, including morbidity, than younger people. There seemed to be widespread awareness of this pattern in our survey responses, as only about 7 percent of those over the age of fifty-five said that they were attending worship services in person after the nationwide restrictions on gatherings were put in place the weekend of March 15–16. In contrast, nearly one-third (29.3 percent) of those in their late twenties and thirties continued to attend, and one-quarter (26.2 percent) of those in their late teens and early twenties (remember that these questions are asked of those who said they attend more often than never).

Further analysis showed other patterns that defy conventional wisdom about worship patterns during the early weeks of COVID-19 in the United States. As with so much in contemporary American politics, attitudes toward COVID-19 were politically polarized from the outset. In the second half of March, though, there was a not a strong partisan difference between Democrats/leaners and Republicans/leaners in terms of their worship patterns (although the small difference was statistically significant). While 22.4 percent of Democrats reported continuing to worship in person, 19.1 percent of Republicans did so as well. In comparison, 14.9 percent of self-identified pure independents (excluding those who lean toward one party or another) continued to worship in person.

There were similarly surprising findings when it comes to the respondent’s most trusted source of news about current events and politics. Contrary to what many might expect, those who put their trust in Fox News were initially much more likely to stop worshipping in person, with only 16.7 percent reporting that they were continuing to do so. This is less than for those who put their trust in outlets with more liberal reputations, such as MSNBC, where 30.7 percent continued to worship in person, or Comedy Central’s The Daily Show or John Oliver’s Last Week Tonight, where a full 37.5 percent continued to worship in person. This may be more of an age effect, though, as the Fox News audience skews older, and Comedy Central’s audience skews younger (Pew Research Center 2012). For their part, those who trust NPR the most reduced their in-person worship to 6.9 percent.

Worship patterns were also not strongly related to supposed “anti-science” skepticism in March. Survey results showed no statistically significant difference between those who said that they agreed that the virus was a “major threat” (19.8 percent) and those who did not (18.1 percent). There also was only a relatively small and statistically nonsignificant 7-percentage-point difference in worship patterns between those who say that they “trust the medical professionals and scientists” (19.6 percent) and those who do not (26.8 percent).

What did make a clear difference, though, were levels of political paranoia around the coronavirus issue. Those who thought that the “hysteria” around the coronavirus was “politically motivated” were 23 percentage points more likely to report continuing to worship in person than those who disagreed with that statement (31 percent to 7.8 percent, respectively) (see Jamieson and Albarracín 2020; and Chapter 2 in this collection by Orcés, Huff, and Jackson for more about the link between conspiracy theory acceptance and COVID-19 attitudes). Further, one of the strongest effects was the respondent’s level of belief that the Democratic Party and its leaders were actively persecuting Christians for their religious and political beliefs. Respondents were asked if they believed that a Democratic president would likely “ban the Bible,” “force you to pay for abortions,” or “take away your guns.” They were also asked if they believed that they would “lose their religious freedom if Democrats control the federal government.” These responses were combined into a collective index variable of “perceived persecution from Democrats.” Among those with the strongest levels of perceived persecution from Democrats (the top third of the scale), a full 40.1 percent continued to worship in person in the few weeks after President Trump’s shutdown order, compared to 16 percent of those with moderate levels of perceived persecution and 9.1 percent of those with the lowest levels of perceived persecution. (These figures remain virtually unchanged if we exclude non-Christians from the sample.)

In reporting these various trends in church attendance in late March, it is also important to keep in mind that many of these factors overlap. We saw, for example, that Fox News viewers were actually less likely to continue worshipping in person. We also know, though, that Fox News tends to draw an audience that is disproportionately older, and older people were more likely to begin stricter social distancing given that it was clear at the time that they were at a greater risk for contracting COVID-19. Given that reality, Table 16.1 shows the independent effect of each of the various factors considered above.

In this more sophisticated (logistic regression) analysis, we see that there were three key factors that each independently predicted whether someone continued to attend in-person religious worship services in the early weeks of the nationwide shutdown: congregational encouragement to attend, age, and perceived religious persecution from the Democratic Party. Specifically, those whose congregations encouraged them to continue to worship in person were 23 percentage points more likely to do so than those whose congregations encouraged them to stay home, even controlling for a variety of other potential factors (described above). The youngest people in the survey (age eighteen) were about 21 percentage points more likely to say that they were continuing to worship in person compared to the oldest (age ninety). Finally, those who strongly perceive that the Democratic Party in the United States is actively seeking to ban the Bible and eliminate religious freedom for Christians were about 20 percentage points more likely to continue to worship in person compared to those who strongly disagreed that Christianity is being actively persecuted by Democrats in the United States.

Table 16.1 also shows that a few other factors mattered, although not to the same extent as the three described above. For example, biblical literalists were about 12 percentage points more likely to continue to worship in person than those who believe the Bible is not divinely inspired; those who attended religious services at least weekly before the pandemic were about 7 percentage points more likely to say that they were continuing to worship in person compared to those who rarely attended. There are also small effects with men, Republicans, those who think that COVID “hysteria” is politically motivated, and, interestingly, those who say that CNN/MSNBC are their most trusted news sources, each being about 4–5 percent more likely to report continuing to worship in person.

One of the most important revelations from this analysis is the strong influence of context on a person’s decision to continue to worship in person during a global pandemic. The context of a person’s congregation (whether the congregation chose to specifically encourage people to continue to attend) and the context of the person’s information environment (whether they believed that Christians were actively under attack by the Democratic Party and its leaders) were two of the three strongest and most consistent predictors of continuing to worship in person.

While there are, of course, legitimate conversations to be had about the important trade-offs between individual liberty and public safety, these results illustrate the strong effect that opinion leaders (whether in religious congregations or political contexts) can have on the public’s decisions on matters of public health and safety in the midst of a pandemic. Those whose congregations did not encourage them to continue to attend and who did not perceive an active political attack on Christianity were significantly more likely to stay home and therefore helped prevent further spread of the coronavirus in the early days of the pandemic. In contrast, those whose information ecosystems were actively cultivating a perceived persecution of Christianity and encouraging them to continue to worship in person were much more likely to do so, likely accelerating the early spread of the virus in the United States.

April and Easter Sunday

Easter Sunday took place on April 12, 2020, four weeks after the Trump administration’s initial recommendation to limit gatherings to ten people or fewer. The week before Easter, the Public Religion Research Institute fielded a national poll (N = 1,007), which included a question asking respondents who say they usually attended worship services at least “a few times a year” whether “the place at which you primarily attend religious services is currently holding in-person gatherings for Easter or other religious occasions in the coming days or weeks?” (PRRI 2020). This survey revealed that fewer than one in ten Americans (9 percent) reported that their congregations were still planning in-person worship services for Easter or other occasions (down a little from the 12 percent who said their congregations were open in the March survey two weeks prior). When asked about their personal plans, only 3 percent of Americans who ordinarily attend religious services at least a few times a year said that they were planning to celebrate Easter or other religious occasions by worshipping in person. Most of the remainder (63 percent) said they planned to celebrate with some sort of online service, and 33 percent said they did not plan to celebrate one way or the other.

The PRRI results further showed that the vast majority of American worshippers planned to either celebrate Easter or other occasions online (or not at all), and the survey results once again show the importance of context: among the 9 percent of regular worshippers who said that their congregations were still planning in-person worship services, nearly one-quarter (23 percent) reported that they were planning on taking advantage of the opportunity and worshipping in person. Only 1 percent of those whose congregations were closed to in-person worship said they planned to worship in person (presumably with a congregation other than their usual one). In other words, when a congregation chose to stay open for in-person worship, one in four planned to attend; when a congregation chose to close for in-person worship, nearly all of its congregants stayed home for Easter or other religious services. This shows once again the strong influence of congregational leaders in a time of a public health emergency—their choice to stay open for in-person worship or to move exclusively online has a strong influence on the choices of worshippers in their congregations and thus also the spread of the coronavirus in their communities (and beyond).

Late October

By late fall of 2020, the United States was well into the third wave of COVID-19 infections, resulting at the time in approximately 80,000 new cases per day (double the rate from only one month previous). In the middle of this spike, the October survey by Djupe, Burge, and Lewis (see Chapter 1) polled 1,800 Americans between October 20 and November 3. Findings from this survey showed consistency in religious behavior from the summer through the fall—roughly one-third (34.2 percent) of American worshippers (i.e., those who say they attend religious services at least “seldom,” about 70 percent of all adult Americans in the survey) said that they were back to attending in-person worship services.

In terms of individual religious identification, mainline Protestants were least likely to say that they were worshipping in person in late October, with only one in five (20.2 percent) doing so, again perhaps because mainline Protestants are, on average, older and at a higher-than-average risk for COVID-19. Views of the Bible made a difference, with biblical literalists roughly 30 percent more likely to be worshipping in person than those who say that the Bible is “not the word of God” (49.2 percent to 18.1 percent, respectively). Typical worship service attendance was also strongly related to whether someone is attending in person or not in late October. Of those who attended worship services weekly or more prior to the pandemic, nearly 40 percent more were likely to say that they are worshipping in person than those who attended a few times a year or seldom prior to the pandemic (50.7 percent to 12.3 percent, respectively).

Similar to patterns shown in late March, religious context was an important predictor of in-person worship in October. Those who had heard their pastor mention the topic of religious freedom in his or her sermons were about 20 percentage points more likely to be attending in person than those who had not (48.1 percent to 29.4 percent, respectively). A similar pattern is evident among those who say that they believe that their congregation’s clergy are highly supportive of President Trump compared to those whose clergy are not supportive (43.1 percent to 26.4 percent, respectively). Those who attend larger congregations (101 or more on a typical worship service) were about 15 percentage points more likely to be worshipping in person (43.7 percent) than those in smaller congregations (100 or fewer) (27.3 percent)—notable especially because COVID-19 had consistently been shown to spread especially in large gatherings, including religious services.

Of course, one of the most direct contextual factors is whether a congregation chose to remain open for in-person worship services or whether they chose to provide these services online or to suspend services temporarily. By late October, only 27.6 percent of American worshippers said that their congregations were still closed to in-person worship, meaning the other three-quarters were hosting in-person services. Of those whose congregations were open, nearly half (48.5 percent) said that they were attending in person. Interestingly, though, nearly one-third (31.2 percent) of those who said that their in-person worship services had been suspended said that they were attending worship services in person. As we saw in the March data, some of this inconsistency is likely due to respondents overreporting their actual in-person worship behavior due to social desirability concerns, but it is also possible that some of those whose primary congregations had suspended in-person services decided to attend other congregations that were open. Indeed, in the late October survey, 12.1 percent of those who said that their primary congregations had suspended in-person worship had switched to a new congregation sometime in 2020.

Given the strong correlation between age and COVID-19 morbidity rates, it is also notable that baby boomers and silent generation worshippers, while ordinarily some of the nation’s most consistent worshippers, continued to curtail their in-person worship activities throughout the fall. Less than one-quarter (23.5 percent) of those born before 1965 reported attending in-person worship activities compared to nearly half (46 percent) of Gen Xers and a third (24.1 percent) of Millennials. For their part, only 28.3 percent of Gen Z worshippers were attending in person (see Chapter 12 for more on this group).

Consistent with our March survey results, certain social/political attitudes were strong predictors of in-person worship in the October 2020 survey. For example, those who believe that “hysteria over the coronavirus is politically motivated” were about 35 percentage points more likely to be worshipping in person than those who disagreed with this (50.5 percent vs. 15.7 percent, respectively). Similar to March, a perceived sense of religious, and specifically Christian, persecution from the Democratic Party is especially correlated with decisions to worship in person or online. Nearly three in five (56.4 percent) of those with strong perceptions of Democratic religious persecution were choosing to worship in person compared to nearly one in five (17.4 percent) of those who disagreed with these perceptions.

Politically speaking, there was some partisan difference when it came to in-person worship behavior in October. Republicans (including independents who lean Republican) were about 12 percentage points more likely to be worshipping in person than Democrats, who themselves were about 7 percentage points more likely than pure independents (42.1 percent, 30.5 percent, and 23.9 percent, respectively). As might be expected, Fox News viewers were more likely than others to be worshipping in person (40 percent), but as might not be expected, they were attending at the same rate as those who list CNN as their most trusted news source (39.7 percent). We also see, again, that those who look to the Daily Show or Last Week Tonight were even more likely to be worshipping in person (44.8 percent), again perhaps related to the younger audiences that these programs tend to attract. In contrast, the media group least likely to be worshipping in person, though, were NPR listeners, with 17.2 percent.

As with the data in March, it is important to keep in mind that many of these factors are correlated with each other, which could distort our interpretation of the effect of each individual factor on the decision to worship in person or not. For example, those who believe that Democrats are actively persecuting Christians in the United States are also less likely to identify as Democrats; which of the two factors is the stronger factor when statistically controlling for the effect of the other?

Shown again in Table 16.1, the logistic regression analysis reveals that the strongest predictor of in-person worship is, as we might expect, typical worship patterns. Those who say that they attended worship services weekly or more prior to the COVID-19 pandemic are 26 percent more likely to be attending in-person worship services than those who typically attended a few times a year. In contrast, typical attendance patterns were a much weaker factor in predicting in-person worship in late March, where frequent attenders were only 7 percentage points more likely to be worshipping in person compared to infrequent attenders.

There is also now an interesting contrast in terms of age. In late March, there was a linear relationship between age and in-person worship patterns—the older you were, the less likely you were to be attending in person. This was also one of the strongest predictors of in-person worship decisions. By late October, though, this pattern had shifted. Table 16.1 shows a curvilinear relationship—those most likely to be worshipping in person were middle-aged Gen Xers, with younger and older individuals both less likely to be attending in person. Indeed, in October the Gen X cohort was 9 percent more likely to be attending in-person services than all other age cohorts when controlling for the various other factors.

Table 16.1 also shows that one consistent factor of in-person worship patterns in the early weeks of the pandemic compared to seven months later was perceived religious persecution from the Democratic Party. In late March, it was one of the three strongest predictors of in-person worship (along with age and encouragement from congregation leaders to attend in person), and it remained so in late October. Those with high levels of agreement that the Democratic Party is actively persecuting Christianity in America were 21 percentage points more likely to be worshipping in person than those with low levels of agreement. It is interesting to note that the third strongest factor was also political in nature—those who strongly agree that “hysteria over the pandemic is politically motivated” were 14 percentage points more likely to be worshipping in person than those who strongly disagree. A third factor was also political in nature—those whose clergy had preached recently about religious freedom were 8 percentage points more likely to be worshipping in person than those whose clergy had not.

This again suggests that contextual factors, and political factors specifically, exerted the strongest influence over a person’s decision to attend in-person worship services by the beginning of November 2020 (perhaps not surprising in the closing days of a contentious presidential election). Adding the various factors together, someone whose clergy were preaching about the importance of religious freedom, who believed that pandemic hysteria was politically motivated, and who perceived active religious persecution from Democrats were 43 percentage points more likely to be worshipping in person than those who did not report those things. These factors exerted a stronger effect than a person’s preexisting religious behaviors and views (normal worship patterns and biblical literalism), which accounted for a 31 percent change in the likelihood of attending in-person worship services.

These findings are also important within the wider context of the relationship between worship patterns and the spread of COVID-19 in 2020. Figure 16.1 shows rates of in-person worship attendance between the different surveys discussed in this chapter, adding additional surveys fielded by Pew Research Center (2020a), Barna (Kinnaman 2020), and the University of Chicago School of Divinity in partnership with AP-NORC (AP-NORC 2020), comparing responses of Americans who attend religious services at least monthly. The general pattern is that in-person attendance for regular worshippers dropped to nearly one-quarter by the end of March and then to low single digits by mid-April. This had climbed to high single digits by the beginning of May, about one-third by the summer, and about half by the fall. For comparison, Figure 16.1 also shows that in-person religious attendance patterns increased steadily just as COVID rates were also exponentially rising throughout the fall of 2020.

Figure 16.1 In-person Worship Patterns among Regular Attenders Compared to Average Daily COVID-19 Rates throughout 2020. (Source: Data from March and October 2020 surveys [Chapter 1], PRRI, Pew Research Center, Barna, and the University of Chicago School of Divinity in partnership with AP-NORC.)

Conclusion

Collectively, this analysis provides strong evidence of the central role that religious and political leaders play in influencing the religious behavior of Americans, even in the middle of a global pandemic. When religious leaders preach about the importance of religious freedom in a national environment where such arguments were frequently invoked to oppose pandemic-related restrictions on in-person religious services, worshippers in the pews take note and respond accordingly. This also shows that when political or religious leaders politicize religion for partisan purposes by promoting baseless conspiracy theories (such as that a Democratic president would ban the Bible or that coronavirus “hysteria” is politically motivated), many people took note and made special efforts to continue to attend religious services in person in the middle of a pandemic.² Had more political and religious leaders instead chosen to support public health officials and emphasized the nonpolitical nature of public health emergencies, in-person worship in the United States would likely have been significantly reduced, thus also significantly reducing the spread of the virus and resulting fatalities.