Overview
Advancements in digital technology and precautions to prevent COVID-19 have prompted several churches to expand their presence online through live streams or recordings of their worship services. During the pandemic, 75% of U.S. evangelical protestants (over 46 million people) reported having attended an online church service, according to a 2021 report by Grey Matter Research and Infinity Concepts. Almost half of those respondents think that online church is at least as good if not better than traditional in-person sermons.
However, many churches are experiencing declining revenues and a less engaged congregation (WSJ, 2021)—a hefty cost for online convenience. Furthermore, it’s unclear whether online services even provide an uplift to overall attendance, and if so, from whom. On the one hand, online services may attract more non-believers, introverts, or others who typically shy away from traditional in-the-pews services. On the other hand, online services may simply draw already engaged in-person churchgoers away to a less-engaging online medium.
However, many churches are experiencing declining revenues and a less engaged congregation (WSJ, 2021)—a hefty cost for online convenience. Furthermore, it’s unclear whether online services even provide an uplift to overall attendance, and if so, from whom. On the one hand, online services may attract more non-believers, introverts, or others who typically shy away from traditional in-the-pews services. On the other hand, online services may simply draw already engaged in-person churchgoers away to a less-engaging online medium.
The Experiment
We conducted a between-subjects experiment with 405 people on the online research platform Prolific to test whether people are more likely to attend online vs. in-person church services, and whether it differs for Christians, church-goers, and extroverts. We used a vignette involving a friend’s invitation to attend in-person or watch the church service online (randomly assigned), then measured each participant’s likeliness to attend or watch.
Participants were instructed as follows, viewing only one of the two texts below. Note that in addition to randomizing whether the invitation was to an online vs. in-person church service, we also randomized the expected time of the sermon (30 minutes vs. 1 hour) to test whether service duration matters.
Participants were instructed as follows, viewing only one of the two texts below. Note that in addition to randomizing whether the invitation was to an online vs. in-person church service, we also randomized the expected time of the sermon (30 minutes vs. 1 hour) to test whether service duration matters.
“Imagine that a friend of yours invites you to attend a church service together. The church is nearby, and the sermon is usually about [30 minutes / 1 hour].”
“Imagine that a friend of yours invites you to watch an online church service together. The video of the sermon is usually about [30 minutes / 1 hour].”
“Imagine that a friend of yours invites you to watch an online church service together. The video of the sermon is usually about [30 minutes / 1 hour].”
Following the text, participants were asked “How likely would you be to [attend / watch] this church service with your friend?” Participants rated their likeliness on a 1-7 scale (1 = Not at all, 7 = Very much).
To measure whether each participant was a Christian or not, we asked participants, “What is your religion?” with answer options Protestant, Roman Catholic, Mormon, Orthodox Greek, Orthodox Russian, Jewish, Muslim, Buddhist, Hindu, Atheist, Agnostic, Other, and prefer not to say. Participants were coded as Christian if they selected any of the first five choices or selected Other and specified Nondenominational Christian or a similar response. Christians comprised 42.7% of our sample, 173 respondents.
To measure participants’ actual history of church attendance, both online and in-person, we asked two survey questions adapted from Pew Research Center, “How often do you watch religious services online (excluding weddings and funerals)?”and “How often do you attend religious services in-person (excluding weddings and funerals)?” Answer options for both questions included More than once a week, Once a week, Once or twice a month, A few times a year, Seldom, and Never (1-6 scale).
Finally, to measure how extroverted each participant was, we used a five-item survey scale adapted from Goldberg’s (1999) Big Five Inventory. The scale items included “I start conversations,” “I talk to a lot of different people at parties,” and “I do not mind being the center of attention,” “I don’t like to draw attention to myself” (reverse-coded), and “I am quiet around strangers” (reverse-coded), all measured on a 1-5 scale. The average extroversion rating was 2.43.
To measure whether each participant was a Christian or not, we asked participants, “What is your religion?” with answer options Protestant, Roman Catholic, Mormon, Orthodox Greek, Orthodox Russian, Jewish, Muslim, Buddhist, Hindu, Atheist, Agnostic, Other, and prefer not to say. Participants were coded as Christian if they selected any of the first five choices or selected Other and specified Nondenominational Christian or a similar response. Christians comprised 42.7% of our sample, 173 respondents.
To measure participants’ actual history of church attendance, both online and in-person, we asked two survey questions adapted from Pew Research Center, “How often do you watch religious services online (excluding weddings and funerals)?”and “How often do you attend religious services in-person (excluding weddings and funerals)?” Answer options for both questions included More than once a week, Once a week, Once or twice a month, A few times a year, Seldom, and Never (1-6 scale).
Finally, to measure how extroverted each participant was, we used a five-item survey scale adapted from Goldberg’s (1999) Big Five Inventory. The scale items included “I start conversations,” “I talk to a lot of different people at parties,” and “I do not mind being the center of attention,” “I don’t like to draw attention to myself” (reverse-coded), and “I am quiet around strangers” (reverse-coded), all measured on a 1-5 scale. The average extroversion rating was 2.43.
Results
An independent samples t-test revealed no significant difference between the in-person (avg. = 2.83) and online (avg. = 3.10) church invitations (p = 0.179). As the graph below illustrates, although the invite to watch the sermon online was rated slightly higher in terms of likely attendance, this difference was small and not outside of our margin of error. There was also no significant difference in likely attendance when the expected time duration of the sermon was 30 minutes (avg. = 3.05) or 1-hour (avg. = 2.88), (p = 0.388), nor was there a significant interaction with the main results for online vs. in-person (p = 0.314).
We did, however, find that Christians were significantly more likely to attend either church service relative to non-Christians (p < 0.001). Furthermore, we found a marginally significant effect on Christians’ likeliness to attend the sermon relative to non-Christians (p = 0.079). Whereas non-Christians were no more or less likely to attend the sermon in-person or watch it online (difference = 0.11, p = 0.555), Christians were significantly more likely to watch the online sermon (difference = 0.70, p = 0.020).
This Christian-specific effect toward online services could be problematic if driven by those who currently attend church in-person. To test for this, we ran similar interaction analyses between our main results for online vs. in-person sermon invites and (a) respondents’ actual history of online church attendance and (b) respondents’ actual history of in-person church attendance.
For online church-goers, we found no significant difference in the main results (p = 0.480). However, traditionally in-person church-goers were marginally more likely to want to attend the online sermon than the in-person sermon (p = 0.066). For every 1-point increase in historical church attendance in-person, we’d expect a person’s likeliness to watch the online sermon to increase by 0.22.
For example, we’d expect a person who never attends church in-person to rate their likeliness to attend the online sermon only 0.12 points higher than the in-person sermon; however, a person who attends more than once a week would be 1.21 points more likely to want to attend the online sermon (1-7 scale). Although the results are barely within the margin of error and should be replicated, this does suggest that online services have the potential to cannibalize more engaged in-person church attendance.
Lastly, we found a significant relationship between how extroverted a person is and likeliness to attend either church service (p = 0.024). For each additional point on our 1-5 extroversion scale, a participant was 0.25 points more likely to attend the church service. However, there was no significant interaction between the main results for online vs. in-person sermon invite and extroversion (p = 0.588). More details regarding our methodology and statistical analysis can be found here.
Conclusion
The proliferation of online church services may come with a trade-off, namely the cannibalization of in-person church-goers. The results of this study suggest that as more churches provide the option of an online sermon, digital attendance may be coming less from nonbelievers or the unchurched and more from current in-person church attendees.
However, a few caveats are in order. First the results are only marginally significant and should be replicated before using them to inform major decisions. Second, the effect sizes are not large and the experimental conditions were only hypothetical. It’s difficult to say whether actual behavior would mirror these results. Last but not least, we can’t say whether the in-person church-goers who’d switch to online were even engaged to begin with, or if they were, that they’d be any less engaged online. If so, there might still be ways to reengage online congregants, perhaps through online volunteering opportunities, discussion forums, or special digital events. What may be needed is more research—both the rigorous academic kind, and practical trial and error—to reengage the digital congregation.
However, a few caveats are in order. First the results are only marginally significant and should be replicated before using them to inform major decisions. Second, the effect sizes are not large and the experimental conditions were only hypothetical. It’s difficult to say whether actual behavior would mirror these results. Last but not least, we can’t say whether the in-person church-goers who’d switch to online were even engaged to begin with, or if they were, that they’d be any less engaged online. If so, there might still be ways to reengage online congregants, perhaps through online volunteering opportunities, discussion forums, or special digital events. What may be needed is more research—both the rigorous academic kind, and practical trial and error—to reengage the digital congregation.
References
Glader, P. and Semakula, J. October 21, 2021. Are Internet Services as Good as Church? Wall Street Journal. https://www.wsj.com/articles/internet-services-as-good-as-church-livestream-in-person-11634848177
Goldberg, L. R. 1999. A broad-bandwith, public-domain, personality inventory measuring the lower-level facets of several five-factor models. In I. Mervielde, I. Deary, F. De Fruyt, & F. Ostendorf (Eds.), Personality psychology in Europe, vol. 7: 7–28. Tilburg, The Netherlands: Tilburg University Press.
Grey Matter Research and Infinity Concepts. 2021. The Ripple Effect: Congregations, COVID, and the Future of Church Life. https://www.infinityconcepts.com/the-ripple-effect/
Goldberg, L. R. 1999. A broad-bandwith, public-domain, personality inventory measuring the lower-level facets of several five-factor models. In I. Mervielde, I. Deary, F. De Fruyt, & F. Ostendorf (Eds.), Personality psychology in Europe, vol. 7: 7–28. Tilburg, The Netherlands: Tilburg University Press.
Grey Matter Research and Infinity Concepts. 2021. The Ripple Effect: Congregations, COVID, and the Future of Church Life. https://www.infinityconcepts.com/the-ripple-effect/