Religion and Public Conflict in the Post-COVID Era: The Case of Protestant Churches in South Korea

Abstract

Governments have attempted to contain the COVID-19 outbreak with a variety of regulations, including social distancing, facemask mandates, or limits on gatherings. South Korea was concerned by the “supercluster” case of a sectarian religious organization in February 2020. Since then, some Protestant churches have periodically caused cluster infections showing antagonism against health authorities. First, we traced all 2020 cluster cases and identified their denominational characteristics. We then utilized the 2020 CISJD data and conducted a series of multivariate regressions to answer the research question, “What causes differences among denominations in attitudes toward public disease control and in-person service attendance?” Results indicated that Protestants affiliated with liberal churches were more likely to follow public disease control guidelines and less likely to attend in-person religious services during the COVID-19 pandemic as compared with individuals from other denominations. Protestants affiliated with moderate, conservative, and fundamentalist churches tended to share antagonism toward public disease control, while cherishing in-person community rituals. This research highlights social implications of public conflict in Korea, where many Protestant churches have emphasized the significance of traditional worship services, claiming the constitutional right of religious freedom, while the majority of citizens, religious and non-religious, disagree with such exclusive claims against public safety.

1. Introduction

The COVID-19 pandemic has had chaotic effects at the individual, institutional, and public levels. In 2020, COVID-19 severely impacted South Korea in mid-February after its initial outbreak in Wuhan, China. However, after instituting strong lockdown policies, South Korea has maintained a high-level safety compared with other countries. While many people in other countries have held the Korean CDC in high regard, Protestant churches and their leaders have caused internal conflicts.

Protestant churches have been identified as main sources of cluster cases among religious organizations. When Koreans first learned of the COVID-19 outbreak in Shin-cheon-ji (literally translated as “a new heaven and earth”), a sectarian Christian church, they acknowledged the dangers of church cluster cases due to church members’ exclusive convictions and intimate interactions. Since then, unlike other religious organizations, conservative or fundamentalist Protestant churches in South Korea have maintained a confrontational stance against public disease control policies (i.e., state rules and regulations regarding social distancing, facemask wearing, quarantine, etc.). It is evident that there have been noticeable conflicts between the church and the state as quarantine authorities amid the social confusion caused by the spread of church-based cluster infections in South Korea (Choe 2020; Kim 2021).

It makes sense that religious activities and rituals are increasing globally during the COVID-19 pandemic as the central government’s control has increased to prevent the spread (Baker et al. 2020; Bentzen 2020). As religion’s role at the individual and public levels has drawn global attention, many studies have examined correlations among COVID-19-related and religious constructs, as religion might have both positive and negative effects during this pandemic. For example, Schnabel and Schieman (2021) have shown that religion has helped to protect mental health but also endangered physical health in the U.S., especially among the highly devout and evangelical Christians.

The main argument of research focused on the positive effects of religion during the pandemic is that it helps to maintain or improve individuals’ mental health. In several countries, including Italy and Poland, religious rituals, practices, and overall religiosity have increased in the short term because they provide comfort and explanation of the suffering caused by COVID-19 (Boguszewski et al. 2020; Molteni et al. 2021). Schnabel and Schieman (2021) also argue that deeply devout individuals and evangelicals have experienced less suffering since COVID-19 reached the U.S. in March 2020. Additionally, many psychoanalytic studies have shown that religious beliefs have had positive effects on individuals’ mental health during the isolation caused by the COVID-19 crisis (Fardin 2020; Legare 2021; Lucchetti et al. 2020; Peteet 2020).

On the other hand, numerous studies have examined the relationship between religion and public conflict in COVID-19 era. While many religious leaders and congregations have complied with the quarantine authorities, it is evident that conflict between the church and the state is intense in the U.S., especially with conservative evangelicals disregarding the disease and the authorities’ policies and demanding in-person religious gatherings in the name of religious freedom (The Economist 2020; Frum 2020; Merritt 2020). Perry et al. (2020a, 2020b) have demonstrated that U.S. conservative Protestants’ “Christian nationalism” justified their reckless attitudes toward quarantine, defense of religious freedom, support of a laissez-faire economy. Further, research has reported that religious rituals (e.g., service attendance) may cause cluster cases: even in a secular state such as the Netherlands, there is a strong relationship between church attendance and the number of patients hospitalized with COVID-19 (Vermeer and Kregting 2020).

However, none of the previous studies have explored or explained how religious organizations differ in response to public pandemic control or quarantine authorities’ policies. This study examined cluster cases of COVID-19 within religious organizations occurring within one year of the initial outbreak based on cumulative data from the Korea Disease Control and Prevention Agency (KDCA). We conducted a series of statistical analyses (i.e., multivariate ordinary least squares (OLS) and logistic regressions) based on the survey data to address the following research questions: What causes differences among Protestant denominations regarding attitudes toward public pandemic control? What are the main factors influencing such attitudinal differences among Christians overall? Particularly, are there any differences in attitudes about worship attendance (as one of the major issues of public control policy)?

2. COVID-19 and Church Cluster Cases

After the initial outbreak of COVID-19 in January 2020, the KDCA was alarmed by the impact of cluster cases. “Cluster cases” are administratively defined as two or more infections from exposure to an identical pathogen at the same time and place (KDCA 2021). Because the cluster cases were major sources of infections, particularly in the early stages, the KDCA (2020) has emphasized prevention of such situations, implementing detailed social distancing and disinfection guidance for multi-use facilities (e.g., pubs, karaoke bars, libraries, etc.) and subdividing the plan into stages based on the pandemic’s current degree of intensity. As health authorities have digitally tracked and publicly announced the infection routes, social stigma from being infected by the virus has been prevalent in South Korea. Often, patients with confirmed cases have been more worried about the social stigma than their own health, fearing being blamed for outbreaks in their workplace or communities. (Gyeonggi Public Health Institute 2020; Kang et al. 2021; Jung and Jun 2020). However, regardless of the quarantine authorities’ efforts to prevent cluster cases and rampant social stigma, some Protestant churches have behaved carelessly in the situation; instead, they have been using COVID-19-related issues to deprecate the government or establish their own legitimacy.

According to a major cluster infection report from the Epidemiological Investigation and Analysis Task Force in KDCA (

Table 1. Epidemiological links of confirmed COVID-19 cases (20 January 2020–19 January 2021).

Confirmed Cases
Total Cases	Major Clusters						Sporadic or Imported Cases
	Sub-Total	Shin-Cheon-ji	Religious Facilities	Nursing Home Facilities	Workplaces	Others
73,115 (%)	33,223 (45.4)	5214 (7.1)	5791 (7.9)	4271 (5.8)	3817 (5.2)	14,131 (19.3)	39,892 (54.6)

Source: KDCA Press Release (31 December 2020) and one-year outbreak major cluster infection report as of 19 January 2021, in the Republic of Korea (Jang et al. 2021).

), the total confirmed cases of COVID-19 as of 19 January 2021 (a year after the initial outbreak) was 73,115. Specifically, 11,005 cases (15%) originated from religious organizations, among which 5214 came from Shin-cheon-ji, and the remainder (5791) originated from religious facilities consisting mostly of conservative Protestant churches.

Another analytical report from the Christian Institute for the Study of Justice and Development (CISJD 2021), merged with the KDCA’s cluster case data from 1 May 2020 to 24 February 2021, indicates that most cluster cases within religious facilities originated from Protestant churches (

Table 2. Cluster cases originating from religious facilities (1 May 2020–24 February 2021).

Religious Facility	Cluster Cases	Confirmed Cases
Protestant church	51	2953
Catholic church	2	19
Buddhist temple	0	0
Shin-cheon-ji (Christian sect)	1	4714
Total	54	7686

Source: COVID-19 cluster cases and Protestant churches in South Korea (CISJD 2021, p. 8).

). Among 54 total cluster cases from religious facilities, 52 cases originated from Protestant churches, 2 originated from Catholic churches, and none originated from Buddhist temples. Catholic churches and Buddhist temples are strongly following religious policies directed by centralized authorities; furthermore, Buddhists have no sabbatical regulations while emphasizing private heart-oriented devotions.

Regarding the denominational affiliation among the 51 cluster cases from Protestant churches, none originated from liberal churches, 10 cases originated from fundamentalist churches, 38 cases originated from conservative churches, 14 cases originated from moderate churches, and 17 cases originated from other or unconfirmed churches.

Additionally, among 10 major cluster cases throughout 2020, the first and fourth case originated from fundamentalist Protestant churches (

Table 3. Ten major COVID-19 cluster cases in 2020.

Case (Num.)	Sarang Jeil church (1173)	Aug. 15 Seoul rallies (650)	Nightclubs (277)	Woori Jeil church (221)	Richway (210)
	Guro-gu call center (170)	Coupang fulfillment center (152)	Gwangju door-to-door sales (150)	Newly planted churches group (119)	Private educational institutes (112)

Source: KDCA Press Release (31 December 2020): Updates on COVID-19 in the Republic of Korea.

).

Despite strengthened regulations, the conflict between conservative Protestant churches and quarantine authorities has been rising. Indeed, more than 50 cluster cases of COVID-19 have been connected to fundamentalist or conservative Protestant churches.

Table 4. Religious facilities linked to 50+ cases of COVID-19 in 2020.

Period		Date of Initial Outbreak	Church Name	Denomination	Cases
Stage 1 (20 January–17 February)	Small-scale imported cases	-	-	-	-
Stage 2 (18 February–5 May)	Large-scale cluster cases	18 February	Shin-cheon-ji	Christian Sect	5213
Stage 3 (6 May–11 August)	Sporadic small cluster cases	12 August	Sarang Jeil Church	Presbyterian (GAPCK: factions)	1173
		12 August	Woori Jeil Church	Presbyterian (GAPCK: factions)	221
		15 August	Yoido Full Gospel Church	Pentecostal	56
Stage 4 (12 August–12 November)	Numerous small-middle cluster cases	24 August	Gwanju Sungrim Baptist Church	The Korea Baptist Convention	56
Stage 5 (13 November–31 December)	Large cluster cases nationwide	20 December	Gwangjin Jungang Church	Presbyterian (GAPCK)	63
		10 December	Youngshin Church	Presbyterian (GAPCK)	138
		9 December	Sungsuk Church	Presbyterian (GAPCK)	213
		12 December	Naeum Church	KAICAM	139
		18 November	Hongdae New Church	Presbyterian (GAPCK)	146

Source: KDCA Press Releases (1 January 2020 ~ 31 December 2020), and Korean Statistical Information Service (KOSIS) daily updates on COVID-19.

shows religious organizations and their denominations responsible for more than 50 cluster cases in 2020, categorized according to five stages outlined by the KDCA. Only conservative and fundamentalist Protestant churches were sources of these cluster cases (not Buddhist, Catholic, or other religious institutions).

Protestant denominations in South Korea can be categorized into four groups: fundamentalist, conservative, moderate, and liberal, as shown in

Table 5. Types of Protestant denominations in South Korea.

Fundamentalist	Conservative	Moderate	Liberal
The General Assembly of the Presbyterian Church of Korea: factions	The General Assembly of the Presbyterian Church of Korea	The Presbyterian Church in Korea	The Presbyterian Church in the Republic of Korea
Presbyterian Church in Korea	The Korea Baptist Convention	The Korean Methodist Church	The Anglican Church of Korea
The Presbyterian Church in Korea: Kosin	Korea Evangelical Holiness Church
The Assemblies of God of Korea
The Presbyterian Church in Korea: Baekseak

(Lee 2002; Noh 1995; Yoon and Oh 2021). Among the religious groups responsible for over 50 cases, most are fundamentalist (factions of Presbyterian (GAPCK) and Pentecostal churches) or conservative (Baptist) denominations.

3. Fundamental Features of Conservative Protestant Landscape

Many scholars argue that fundamentalist churches claim exclusive salvation based on biblical literalism, disconnecting themselves from worldly values or norms (Jang 2021; Kim 2011; Lee 2002; McGuire 2008; Yoon and Oh 2021). Under the historical influences of American missionaries or Confucian legitimacy (Yoon and Oh 2021), Korean Protestant churches generally share fundamentalist features or exclusive religiosity, whereas liberal churches have been equipped with the ecumenical spirit of the World Council of Churches (Keum 2014). Jang (2019, 2021) argues that conservative Protestant churches have shared exclusive traits in the Korean church tradition and history, having experienced rapid modernization, the Korean War, and the influx of fundamentalism from the United States. Lee (2002) also elaborates on the conflicts stemming from fundamentalist elements of Korean Protestant churches—such as sexism, conflicts among denominations and social confusion in alignment with conservative (right-wing) political powers—focusing on Korea’s unique cultural tradition and the inflow of fundamentalist theology from American missionaries. Similarly, Ryu (2008) has traced the history of the unique evangelical nature of Korean Protestantism, arguing that the theological evangelism and ethical rigorism (e.g., professing Bible-centered doctrines and strictly honoring the Sabbath) among Korean Protestants is largely due to the arrival of American missionaries from evangelical denominations in the 20th century.

The fundamentalist features of Korean Protestant churches have grown dramatically during the COVID-19 pandemic. Our research contributes to theoretical explanation of fundamentalism among Protestants by explaining how public conflicts have triggered these traits (Almond et al. 2011; Ammerman 1987; Antoun 2008; Bruce 2008; Emerson and Hartman 2006; Lawrence 1990; Marsden 2006; Riesebrodt 2000), especially in comparison with liberal believers’ faith and attitudes toward public disease control (Perry et al. 2020a).

We attempted to analyze fundamentalist characteristics of Korean Protestants to understand their public actions against quarantine authorities (Almond et al. 2011; Ammerman 1987; Berger 2011; Iannaccone 1994; Lee 2002). Specifically, we examined three major elements of religiosity: (1) belief, (2) ritual, and (3) community (Durkheim 2008).

First, belief has cognitive and emotional dimensions. Conservative belief is based on biblical literalism, inerrancy, and deep faith in the sovereignty of God. As faith encompasses emotional, devotional, or integrational parts of belief and reflects the depth of one’s personal relationship with God, our study adopts belief for empirical analysis. Second, ritual represents the practical dimension of religiosity, on the individual or communal level. However, religious tradition has more strongly emphasized the importance of communal rituals, such as congregational worship and communal services. Third, a church as a community is characterized by denominational theology, tradition, and traits.

Korean conservative churches share following fundamental traits. Under the protection from “the sacred canopy,” guaranteed by scripture (e.g., Ps. 91:1–3; Eph. 6:10–12; Deut. 31:6, etc.; Bartkowski 1996; Berger 2011), some extremists refuse modern medicine in favor of divine healing (Glassman 2018), and they wage the good fight against enemies who try to hinder their religious traditions and devotion. By means of securing religious legitimacy, they have a strong sense of community, supporting their religious rights (Tamney 2002, p. 183). As conservative Protestant churches, previously regarded as a model religion in national development and modernization after the 1945 liberation in Korea, have gradually lost their influence and legitimacy in democratic regimes (1993~2008; Cho 2014; Kim 2013), they began actively affiliating with conservatives in the political sphere and campaigning for a return the glorious past. Not only to save nonbelievers needing redemption but also to maintain their religious legitimacy, active campaigning has become one of the most major missions to undertake and attending religious services has become a moral imperative for building an impermeable community. Fundamentalist traits of Korean Protestants based on exclusive legitimacy (over secular power and against quarantine authorities) could develop ingroup-biased attitudes of disregarding health authorities during the COVID-19 pandemic.

4. Public Pandemic Control and Religious Freedom

Conflicts have occurred in South Korea between authorities implementing public pandemic control (i.e., governmental regulations to contain this highly contagious disease) and religious groups insisting on their freedom. Despite the KDCA’s requests (i.e., regulations) to have virtual religious services and meetings, conservative Protestants have insisted on their public religious freedom. In an annual report from the Office of International Religious Freedom, the U.S. Department of State (2021) noted twofold attitudes of religious groups in Korea when the government restricted religious services to minimize the spread of COVID-19. Buddhists, Won Buddhists, and Catholics, which constitute major religions in South Korea, consistently have shown cooperative attitudes toward the government’s public policies; however, in the Protestant sphere, both cooperative and non-cooperative attitudes have emerged (Ko 2021). What influences these different attitudes towards public disease control? To answer this question, it is necessary to consider the contrasting theoretical stances of the religious market and public religion.

From the perspective of religious market theory, autonomous features of “individual churches” or “self-supporting pioneering churches” should be considered (Choi 2013; Yi 2014). First, it is difficult to impose centralized rules or systems not only for mainline or large denominations but also for minority or independent ones due to the “individual-centered nature” of Korean Protestantism (Kim 2013, p. 29; Noh 1995, p. 35). The administrative and financial centralization of Korean Protestant churches is far weaker than that of other religious organizations, such as Catholic churches (Ko 2021). Kim (2020a) argues that the financial burdens of most pastors of Protestant churches are very heavy because they must manage expenses in churches with their own resources, and COVID-19 has aggravated this situation. Additionally, the number of pastors in small churches who must work outside churches with second or third jobs has grown because the sizes of congregations have declined rapidly since the beginning of the pandemic (Kim 2020b). In such situations, fundamentalist or conservative leaders of individual Protestant churches might justify in-person worship services despite the risk of cluster infections.

Second, in “non-self-supporting churches,” which account for over half of all Protestant churches in Korea, the financial situation has deteriorated as church attendance has decreased during the COVID-19 crisis. The meaning of “non-self-supporting church” differs for each denomination, but generally it means a newly planted church that cannot support itself with few congregation members, located in a rural or suburban area (Noh 1995, p. 343). Leaders of such Protestant churches tend to cherish opportunities for in-person religious services and communal mealtimes, which provide a strong sense of belongingness for church members (Baek et al. 2020). These situations are highly likely to result in cluster infections. In fact, most incidents of cluster cases have originated from non-self-supporting small churches with under 100 congregation members with respect to the total 51 church-oriented cluster cases reported (

Table 6. Cluster cases from Protestant churches categorized by church size.

Number of Congregation Members	Less than 100	100–500	500–1000	Over 1000	Unconfirmed
Number of incidents	20	10	1	8	12
Number of confirmed cases	437	223	33	1906	354

Source: COVID-19 cluster cases and Protestant churches in South Korea (CISJD 2021, p. 12). Note: This report is based on KDCA’s epidemiological data collected from 1 May 2020 to 24 February 2021.

).

Fundamentalist or conservative churches emphasize “religious freedom,” often regarding public control or governmental regulations as unnecessary oppression (Klassen 2020). Fundamentalist or conservative churches tend to share strong faith and conviction through their rituals and services. In crisis, “strong churches” typically are more powerful, offering faith-oriented conviction stemming from strict understanding of the Bible in the midst of a highly uncertain world, as opposed to “lenient churches” based on liberal understanding of the Bible (Iannaccone 1994). However, during the COVID-19 crisis, fundamentalist or conservative churches’ struggles to maintain in-person services at the risk of causing public conflicts must be analyzed as complicated effects of religious confession and economic survival.

There have been numerous studies shedding light on the conflict between religious freedom and public control in many countries during the COVID-19 crisis (Androutsopoulos 2021; Haynes 2021; Istratii 2020; Madera 2020; Martínez-Torrón 2021; Mosquera 2021). Though many scholars have focused on how religious freedom is interpreted legally and politically in a given country, it should be interpreted in the contextual relationship between religion and the public. In fact, religious leaders have emphasized religious freedom, especially during the early stages of the outbreak, as a political means to avoid public control (Barlow 2021; Kim 2020c; Seo and Tizzard 2020). In contrast, advocates of “public religion” might have a different understanding of the issues of public pandemic control relative to proponents of religious market theory (Casanova 1994; Oh 2013). They argue for the necessity of church renewals in accordance with the “signs of the times” in the changing public sphere, as they understand that the purpose of religion is oriented toward the common good beyond group-oriented interests or concerns in the religious marketplace.

5. Methods

5.1. Data

This study utilized 2020 data from the Christian Institute for the Study of Justice and Development (CISJD). The CISJD has been conducting ‘Analysis of Protestants’ Perceptions on Major Social Issues’ yearly since 2018. Specifically, the 2020 CISJD survey focused on Korean Protestants’ attitudes on religious activities, ecology, security, public regulation, etc. The 2020 survey was designed by the CISJD and was conducted by JI&COM Research, an opinion poll research and analytics company specialized in Christian survey. In order to draw a nationally representative sample, JI&COM conducted proportional quota sampling, recruiting a panel of respondents who agreed to participate in surveys, and then it matched 1000 respondents proportionately closest to the National Census. Initially, JI&COM sent the survey questionnaire via e-mail to all collected panel members, and then selected multi-stage cluster samples—by region, gender, and age—of Protestants nationwide. The 2020 survey was conducted from 21 July to 29 July after the COVID-19 outbreak, and JI&COM collected a total of 1000 valid Protestants who completed all questions following ‘online’ survey mandates.

Table 7. Sociodemographic characteristics (N = 1000).

Variables		N	%	Variables		N	%
Sex	Male	446	44.6	Income (Won)	Lower than 1.5 million (mm.)	53	5.3
	Female	554	55.4		1.5~2 mm.	63	6.3
Age	20–29	174	17.4		2~3 mm.	139	13.9
	30–39	204	20.4		3~4 mm.	154	15.4
	40–49	245	24.5		4~5 mm.	199	19.9
	50–59	229	22.9		5~6 mm.	143	14.3
	60–69	148	14.8		6~7 mm.	88	8.8
Education	High school or less	167	16.1		7~8 mm.	67	6.7
	Attending college or higher	833	83.3		8~9 mm.	46	4.6
					9~10 mm.	16	1.6
					Higher than 10 mm.	32	3.2

shows sociodemographic features of collected samples.

5.2. Measures

5.2.1. Dependent Variables

• Anti-Public Control

We measured “anti-public control” as a dependent variable, using a summative scale based on four items with a five-point Likert scale (

Table 8. Item contents of anti-public control.

Factor	Items	Cronbach’s α
Anti-Public Control	(1) Do you consider the government’s recommendation to refrain from religious gatherings due to COVID-19 as a violation of religious freedom? (re-verse-coded) (2) What do you think of the media and public’s criticism against Protestant churches for causing cluster cases? (3) What is your opinion about practicing legal punishment for those not wearing a mask? (4) How much do you trust the present administration’s ability to handle the COVID-19 crisis appropriately?	0.637

). These items assessed attitudes toward public policies implemented by quarantine authorities. For item 1 and 3, respondents indicated their level of opinion on a scale ranging from 1 (“Strongly agree”) to 5 (“Strongly disagree”); item 2 includes 1 (“Never unfair”) to 5 (“Very unfair”); and item 4 from 1 (“Very trustworthy”) to 5 (“Do not trust at all”).

• In-Person Service Attendance before/during/after COVID-19

We also measured in-person service attendance at three timepoints: before, during, and after COVID-19 (

Table 9. Items for in-person service attendance.

Variables	Items
Before COVID-19	Did you attend church before the COVID-19 outbreak?
During COVID-19	In what way did you attend service last week (19 July 2020)?
After COVID-19	What is your opinion on attending a service after COVID-19?

). We recoded responses in a binary fashion: “0” for no attendance or virtual service and “1” for in-person service attendance.

For the item assessing service attendance before COVID-19, respondents could choose “No = 0” for virtual or no service attendance or “Yes = 1” for in-person service attendance.

For the item assessing service attendance during COVID-19, respondents could choose “I had an in-person service at the church I usually attend” = 1, “I had a virtual service broadcasted by the church I usually attend” = 2, “I had an in-person service at a church I do not usually attend” = 3, “I had a service via a Christian TV program” = 4, “I had a service via a Christian radio program” = 5, “I had a family service at home” = 6, “I had an in-person service at any church providing in-person services” = 7, or “I had no service” = 8. We recoded responses for virtual or no services (2, 4, 5, 6, 8) as 0 and for ‘in-person services’ (1, 3, 7) as 1.

For the item assessing service attendance after COVID-19, respondents could choose “We can have virtual or family service” = 0, “We must have an in-person service” = 1.

5.2.2. Independent Variables

The key independent variables were religiosity, status in church, and Protestant denominations (

Table 10. Descriptive statistics for key independent variables.

Measures	Description	N	Range	Mean or %	SD
Religiosity
Degree of faith	Level 1 to Level 4 (higher level indicates greater devotion)	1000	1–4	2.42	1.04
Religious life	In years	1000	0–68	27.67	14.46
Frequency of service attendance	Before the outbreak (in a year)	1000	0–365	86.43	89.11
Status in church		1000
Pastor	Head pastors, junior pastors	25	0–1	2.5%
Church officers	Elder, exhorter, ordained deacon	143	0–1	14.3%
Deacons	Deacon(ess), deputy deacon(ess)	270	0–1	27%
Laypeople	No position	562	0–1	56.2%
Protestant denominations		930
Liberal	The Presbyterian Church in the ROK, The Anglican Church of Korea	40	0–1	4.3%
Moderate	The PCK, The Korean Methodist Church	471	0–1	50.6%
Conservative	The General Assembly of the PCK, The Korea Baptist Convention, Korea Evangelical Holiness Church	224	0–1	24.1%
Fundamentalist	Factions of The General Assembly of PCK, PCK, The PCK: Kosin, Baekseok, The Assemblies of God of Korea	122	0–1	13.1%
Others	Independent denominations, etc.	73	0–1	7.8%

). As a proxy for religiosity, years of religious life, frequency of service attendance in a year, and degree of faith measure the intensity of religious behavior and how much individuals are devoted to their relationship with a deity (God or Jesus). We recoded the frequency of service attendance into a 365-day term (e.g., “once a month” to 12 days and “every week” to 52 days). For the degree of faith item, respondents indicated their general agreement with the following statements:

• Level 1: I believe in God, but I don’t know much about Jesus. Religion is still not a big part of my life.
• Level 2: I believe in Jesus, and I am doing many things to get to know Him.
• Level 3: I am close to Christ and rely on His guidance every day.
• Level 4: God is everything in my life, and I am enough with Him. All my work reveals Christ.

Status in church, classified based on the hierarchical structure within the church, differs in each denomination, but it generally captures embeddedness in or commitment to religious communities. Lastly, we classified Protestant denominations into five distinct categories based on Table 5: liberal (4.3%), moderate (50.6%), conservative (24.1%), fundamentalist (13.1%), and others (7.8%), with 70 missing values for those who did not know the affiliation of the church that they attend. “Others” indicate independent denominations not affiliated with mainline Protestantism.

5.2.3. Control Variables

Our analyses included sociodemographic controls (Table 7) for age (in years), sex (male = 1, female = 0), educational attainment (high school or lower = 1, attending/graduated college or higher = 0), and income level (lower than 1.5 mm. (million) = 1, higher than 10 mm. = 11).

5.3. Analytic Strategy

We used OLS regression analysis to examine the effects of religiosity, status in church, and denomination on public disease control attitudes. Then, we used logistic regression to analyze in-person service attendance at the three timepoints. For all dependent variables, we utilized two analytic models. Model 1 as a main effect model examined overall effects of religiosity, status in church, and denomination on public disease control attitudes. Model 2 included the interaction effect between degree of faith and denominational affiliation on the dependent variables.

6. Results

We first inspected the correlations among attitudes toward public disease control and in-person service attendance. As

Table 11. Correlations among major dependent variables (N = 1000).

	In-Person Service Attendance			Anti-Public Control
	A	B	C	D
A	1
B	0.298 ***	1
C	0.403 ***	0.373 ***	1
D	0.233 ***	0.303 ***	0.293 ***	1

Note: In-person service attendance (A: before COVID-19, B: during COVID-19, C: after COVID-19) and D: anti-public control. * p < 0.05, ** p < 0.01, *** p < 0.001.

indicates, there were strong, significant correlations among these attitudes and all three phases of service attendance (all p < 0.001). Next, we performed a one-way analysis of variance (ANOVA) comparing the five denominations (

Table 12. One-way ANOVA on dependent variables by denominations.

Factors		Protestant Denominations	N	Mean	SD	F	p
In-person service attendance	Before COVID-19	Liberal	40	0.6250 B***	0.49029	8.484	0.000 ***
		Moderate	471	0.8450	0.36228
		Conservative	224	0.9286 A	0.25812
		Fundamentalist	122	0.9016	0.29903
		Others	73	0.9178	0.27656
	During COVID-19	Liberal	40	0.1250 B***	0.33493	7.628	0.000 ***
		Moderate	471	0.3036	0.46030
		Conservative	224	0.4643 A	0.49984
		Fundamentalist	122	0.4344	0.49773
		Others	73	0.3699	0.48611
	After COVID-19	Liberal	35	0.1714 B**	0.38239	3.622	0.006 **
		Moderate	421	0.2328	0.42311
		Conservative	209	0.3397 A	0.47475
		Fundamentalist	114	0.3596	0.48202
		Others	64	0.2500	0.43644
Anti-public control		Liberal	40	−0.3173 B*	0.91219	5.929	0.000 ***
		Moderate	471	−0.0971	0.97768
		Conservative	224	0.1972	1.04305
		Fundamentalist	122	0.1469	0.98334
		Others	73	0.2400 A	1.03155

Note: Regarding post-hoc analysis (Bonferroni test), “A” represents the highest mean score, and “B” represents the lowest score. Asterisks adjacent to B indicate the significance of A–B. * p < 0.05, ** p < 0.01, *** p < 0.001.

). Results indicated that the mean scores of liberal denomination affiliates were the lowest on all dependent variables, conservative denomination affiliates’ scores for in-person service attendance were the highest, and those affiliated with other denominations exhibited the strongest anti-public control attitudes. There were significant differences among denominations on attitudes toward both public pandemic control and in-person service attendance.

In the multivariate OLS models presented in

Table 13. OLS regression models predicting Protestants’ anti-public control attitudes during the COVID-19 crisis.

		Anti-Public Control
		Model 1			Model 2
		b	S.E.	t	b	S.E.	t
(Constant value)		−0.376	0.205	−1.831	−0.747	0.403	−1.853
Religiosity	Degree of faith	0.233 ***	0.035	6.617	––	––	––
	Religious life (years)	−0.001	0.002	−0.458	−0.001	0.002	−0.412
	Frequency of service attendance (before the outbreak)	0.001 ***	0.000	3.642	0.001 ***	0.000	3.645
Status in church	Pastor	0.429 *	0.210	2.046	0.444 *	0.210	2.109
	Church officer	0.214 *	0.107	1.990	0.219 *	0.108	2.033
	Deacon	0.153	0.080	1.920	0.155	0.080	1.952
Denomination	Liberal	−0.267	0.189	−1.415	0.033	0.505	0.065
	Moderate	−0.146	0.121	−1.203	0.300	0.389	0.770
	Conservative	0.040	0.128	0.312	0.391	0.415	0.941
	Fundamentalist	0.018	0.140	0.129	0.285	0.439	0.648
	Liberal × Degree of faith	––	––	––	0.265	0.152	1.748
	Moderate × Degree of faith	––	––	––	0.198 ***	0.047	4.210
	Conservative × Degree of faith	––	––	––	0.238 ***	0.065	3.678
	Fundamentalist × Degree of faith	––	––	––	0.271 **	0.085	3.202
	Others × Degree of faith	––	––	––	0.354 **	0.119	2.966
Sociodemographics	Age	−0.005	0.003	−1.623	−0.005	0.003	−1.604
	Sex-Male	0.003	0.063	0.040	0.000	0.063	−0.006
	Education-High school or lower	−0.066	0.086	−0.768	−0.071	0.087	−0.815
	Income	−0.013	0.014	−0.963	−0.013	0.014	−0.943
F(p)		11.061 ***			8.691 ***
adj. R2		0.132			0.13
N		1000			1000

Note: Among the dummy variables, the reference groups were sex: “female”, education: “university or higher”, status: “laypeople”, and denomination: “others”, * p < 0.05, ** p < 0.01, *** p < 0.001.

, degree of faith (p < 0.001) and frequency of service attendance (p < 0.001) were significant predictors of anti-public control attitudes. As one’s degree of faith and frequency of religious attendance increase, one is significantly more likely to hold negative attitudes toward public health control.

Second, holding pastor and church officer status (compared with being a layperson) also was positively associated with anti-public control attitudes in both Models 1 and 2 (p < 0.05). We can explain this finding noting that people who are more deeply and directly engaged in managing churches regard such public restrictions as threats to maintaining churches in the competitive Protestant sphere.

Finally, we examined the relationship between denomination and anti-public control attitudes. The direct effect of denominational affiliation on anti-public control attitudes in Model 1 was not statistically significant; however, when considering interaction effects (faith * denomination) in Model 2, the association was indeed significant. The faith of liberal protestants did not affect anti-public control attitudes, whereas the faith of Protestants belonging to moderate (b = 0.198, p < 0.001), conservative (b = 0.238, p < 0.001), fundamentalist (b = 0.271, p < 0.01), or other denominations (b = 0.354, p < 0.01) was significantly positively associated with anti-public control attitudes.

In the multivariate logistic regression model presented in

Table 14. Logistic regression models predicting Protestants’ attitudes toward in-person service attendance.

		In-Person Service Attendance
		Before COVID						During COVID						After COVID
		Model 1			Model 2			Model 1			Model 2			Model 1			Model 2
		b	S.E.	O.R.	b	S.E.	O.R.	b	S.E.	O.R.	b	S.E.	O.R.	b	S.E.	O.R.	b	S.E.	O.R.
(Constant value)		0.585	0.815	0.515	0.591	1.346	0.193	−2.357 ***	0.538	19.236	−1.676	1.137	2.173	−1.092	0.572	3.642	−0.211	1.161	0.033
Religiosity	Degree of faith	0.368 **	0.140	6.942	––	––	––	0.408 ***	0.092	19.822	––	––	––	0.568 ***	0.095	36.122	––	––	––
	Religious life (years)	−0.001	0.009	0.006	0.000	0.009	0.000	0.013	0.007	3.740	0.013 *	0.007	4.030	0.015 *	0.006	5.542	0.015 *	0.006	5.546
	Frequency of service attendance (before the outbreak)	0.033 ***	0.004	73.018	0.033 ***	0.004	71.899	0.007 ***	0.001	53.238	0.008 ***	0.001	54.532	0.008 ***	0.001	30.914	0.008 ***	0.001	30.594
Status in church	Pastor	0.037	1.201	0.001	0.047	1.211	0.002	1.020	0.568	3.225	1.038	0.570	3.316	−1.229	0.646	3.621	−1.193	0.651	3.360
	Church officer	0.814	0.503	2.619	0.807	0.508	2.520	1.336 ***	0.271	24.238	1.360 ***	0.274	24.723	0.100	0.290	0.119	0.099	0.290	0.117
	Deacon	1.508 ***	0.367	16.887	1.516 ***	0.369	16.908	1.097 ***	0.205	28.591	1.112 ***	0.206	29.022	0.473 *	0.213	4.942	0.471 *	0.213	4.892
Denomination	Liberal	−1.563 *	0.693	5.090	−2.480	1.657	2.240	−0.878	0.608	2.089	−1.275	1.782	0.512	−0.334	0.500	0.447	−1.316	1.385	0.903
	Moderate	−0.341	0.549	0.386	−0.300	1.281	0.055	0.133	0.305	0.191	−0.207	1.114	0.035	−0.227	0.361	0.397	−1.095	1.130	0.940
	Conservative	−0.146	0.607	0.058	−0.054	1.409	0.001	0.656 *	0.318	4.249	−0.338	1.171	0.083	0.340	0.390	0.762	−0.798	1.198	0.444
	Fundamentalist	−0.008	0.634	0.000	0.053	1.470	0.001	0.692 *	0.349	3.936	−0.973	1.275	0.583	−0.247	0.405	0.373	−0.958	1.236	0.601
	Liberal × Degree of faith	––	––	––	0.881	0.588	2.242	––	––	––	0.289	0.542	0.285	––	––	––	0.627	0.395	2.523
	Moderate × Degree of faith	––	––	––	0.339	0.183	3.447	––	––	––	0.272 *	0.126	4.646	––	––	––	0.569 ***	0.125	20.654
	Conservative × Degree of faith	––	––	––	0.319	0.306	1.086	––	––	––	0.521 **	0.169	9.505	––	––	––	0.690 **	0.204	11.486
	Fundamentalist × Degree of faith	––	––	––	0.333	0.365	0.832	––	––	––	0.768 **	0.242	10.027	––	––	––	0.503 *	0.223	5.089
	Others × Degree of faith	––	––	––	0.364	0.485	0.564	––	––	––	0.190	0.328	0.338	––	––	––	0.269	0.357	0.569
Sociodemo- graphics	Age	−0.032 **	0.011	8.816	−0.032 **	0.011	8.815	−0.035 ***	0.008	18.538	−0.036 ***	0.008	19.405	−0.009	0.008	1.462	−0.009	0.008	1.458
	Sex-Male	0.846 **	0.254	11.085	0.838 **	0.256	10.742	0.062	0.162	0.145	0.070	0.162	0.185	0.017	0.167	0.010	0.017	0.167	0.011
	Education—High school or lower	0.223	0.345	0.418	0.217	0.347	0.390	0.052	0.222	0.055	0.063	0.223	0.080	−0.581 *	0.226	6.603	−0.577	0.226	6.507
	Income	0.005	0.053	0.010	0.006	0.053	0.012	0.049	0.036	1.899	0.051	0.036	2.023	0.005	0.036	0.016	0.002	0.037	0.002
-2LL		466.527			465.581			962.824			958.202			899.914			898.74
N		930			930			930			930			890			890

Note: Among the dummy variables, the reference groups were sex: “female,” education: “university or higher,” status: “laypeople”, and denomination: “others.” * p < 0.05, ** p < 0.01, *** p < 0.001.

, as with the previous analysis, the degree of faith and frequency of service attendance significantly predicted in-person service attendance for all three phases (before, during, and after COVID-19) in both Models 1 and 2. This result indicates that deeper faith and more frequent past church attendance made individuals significantly more likely to attend in-person services in all phases.

However, in contrast to the previous analysis, deacons were more likely to attend in-person services before (b = 1.508, p < 0.001) and during COVID-19 (b = 1.097, p < 0.001), church officers only during COVID-19 (b = 1.336, p < 0.001), and pastors no more likely to attend services in any phase compared with laypeople. This result indicates that deacons who are actively involved in general church affairs are more likely to attend church services than are laypeople.

Finally, we examined the relationship between denominational affiliation and in-person service attendance. Before COVID-19, liberal Protestants attended in-person church services less frequently than did Protestants in other denomination (b = −1.563; p < 0.05). During COVID-19, conservative and fundamentalist Protestants attended in-person services more frequently than did those of other denominations (p < 0.05). After COVID-19, there were no significant differences among denominations with respect to service attendance. On the other hand, when accounting for interaction effects (faith * denomination) in Model 2, Protestants of moderate, conservative, and fundamentalist denominations were more likely to attend services during and after COVID-19 compared with those of liberal or other denominations. In sum, Protestants belonging to liberal denominations demonstrated the least concern about attending in-person religious services.

7. Discussion

This study began by recapitulating social problems stemming from the confrontational relationship between religion and health authorities in the public sphere. First, we traced the confirmed cluster cases of COVID-19 using the KCDA data. We found that most major cluster cases among all religious organizations in Korea came from Protestant churches, particularly the ones that are fundamentalist, conservative, or sectarian. What has caused some religious organizations to disobey governmental or health authorities’ regulations in the COVID era? To explore this question, we utilized CISJD data and statistically examined the influences of elements of religiosity and denominations on attitudes toward public disease control and in-person religious services.

First, the two-stage multivariate OLS regression on anti-public control attitudes yielded interesting findings (Table 13). In Model 1, the degree of faith and frequency of service attendance (dimensions of religiosity) as well as status in church were significantly positively associated with anti-public control attitudes. In Model 2, when considering interaction effects (faith * denomination), the faith of liberal Protestants did not significantly affect anti-public control attitudes, whereas the faith of Protestants belonging to moderate, conservative, fundamentalist, or other denominations was significantly positively associated with anti-public control attitudes.

Second, we conducted a series of multivariate logistic regressions on in-person service attendance (before, during, and after COVID-19) with two models (Table 14). In Model 1, the degree of faith and frequency of service attendance (before the outbreak) as well as status in church (e.g., deacon status) were significantly positively associated with in-person service attendance (before, during, and after COVID-19). The analysis also uncovered denominational effects. Before COVID-19, liberal Protestants reported attending church services significantly less frequently than Protestants of other denominations. During COVID-19, conservative and fundamentalist Protestants attended church services significantly more frequently than Protestants of other denominations. After COVID-19, there were no significant differences among denominations with respect to in-person service attendance. However, when considering interaction effects (faith * denomination) in Model 2, Protestants of moderate, conservative, and fundamentalist denominations reported attending church services significantly more often than Protestants of liberal or other denominations during and after COVID-19.

This study shed light on a congenial aspect of religious culture and denominational differences among Korean Protestants, which represent significant contributions to the sociology of religion. This study complements the literature on religion and COVID-19, especially studies focused on religion’s role in forming negative attitudes toward disinfection policies during the pandemic (Hill et al. 2020; Perry et al. 2020a). Regarding public disease control, although only some sectarian, fundamentalist, or conservative churches have been involved in cluster cases, antagonism against public control among Korean Protestants depends on religiosity (e.g., degree of faith and frequency of service attendance) and status in church, regardless of denominational affiliation. However, the degree of faith has a significant effect on such antipathy toward public regulations, except for those of liberal denominations. Thus, it implies a significant aspect of religious culture in Korean Protestantism: with stronger faith, Protestants outside liberal denominations are more tightly anchored to the constitutional right of religious freedom (Hurd 2015) and more likely to oppose governmental health authorities. Additionally, pastors and church officers, who are generally deeply engaged in managing organizations, showed stronger antagonism toward public quarantine measures than did laypeople, regardless of denominational affiliation, and deacons, who are actively involved with general church affairs, showed higher willingness to attend in-person services at all timepoints than did laypeople.

The following results also show the important elements of Korean Protestants’ religiosity. As Korean Protestants’ antagonism toward public regulations is directly related to governmental prohibition or limitation of in-person service attendance, the second part of the multivariate regressions explains overall denominational differences regarding in-person service attendance. During COVID-19, Protestants affiliated with conservative or fundamentalist denominations insisted on in-person worship. However, when faith entered the analysis, the inclination toward in-person worship during and after COVID-19 appeared equally among Protestants affiliated with moderate, conservative, and fundamentalist churches. Therefore, we have encountered new significant findings that Korean Protestants’ faith has a common ethos concerning religious freedom (e.g., anti-public control) and religious service attendance. In fact, all these concepts represent the “elementary forms of religious life” (Durkheim 2008): belief, ritual, and community. No matter how important these are, when Protestants reject or blame governmental health authorities in public, its character becomes fundamentalist exclusivism in the public sphere in the eyes of non-religious or other religious citizens.

Still, we must consider two denominations. First, liberal denominations show distinct differences from fundamentalist ones. Liberal protestants, enlightened by modern biblical hermeneutics and imbued by the ecumenical spirit of church renewal in accordance with the World Council of Churches (WCC) and the National Council of Churches in Korea (NCCK), are freed from an exclusively church-oriented theology and allowed to pursue civil and common values in the public sphere (Keum 2014). Actively engaged with and open to accepting socio-political discourses and social movements toward the enhancement of the public good, some liberal theologians and liberal Protestant organizations have taken stands for human rights of the marginalized and sexual minorities (Jung 2014; Masci and Lipka 2015; Ryu 2015) and have been involved in environmental movements with self-awareness of the church’s past role in the global environmental crisis (Chun 2010). Such characteristics of Korean liberal Protestantism regarding the public good are also well expressed in their main doctrines of public theology and Minjung Theology, the latter of which literally means theology of “the (marginalized) people” (Seo 1975, 1983; Song 1984). In line with such ecumenical doctrines, liberal Protestantism has been distinctly aware of its role in the public sphere, endeavoring to overcome church-centeredness and to engage with public issues. With willingness and openness toward advancing the public good and an ecumenical understanding of theological concepts compared with traditional devotion, they are more likely to comply with public disease control and not take confrontational stances against health authorities; ultimately, this means that they are responsible for fewer COVID-19 cluster cases (CISJD 2021, p. 9).

Second, the “other” denominations must be considered, although we cannot specify their churches merely by analyzing these surveys. When accounting for faith, Protestants affiliated with other denominations exhibited the greatest antagonism toward public control among all five denominations. However, their willingness to attend in-person worship services was not as strong as that of mainline Protestants affiliated with moderate, conservative, and fundamentalist churches. These findings imply that non-mainline Protestants tend to object to governmental control; however, their faith is not significantly associated with church service attendance.

This study has several limitations. First, the survey data were collected “online” and at only one timepoint in 2020, not in multiple waves; thus, they may not represent the long-term religious landscape, and the KDCA data may become less accurate as the epidemiological situation continues developing. Second, this study focused on the degree of faith rather than the cognitive level of doctrinal belief. For the latter part, Yoon and Oh (2021) have explained the exclusive characteristics of fundamentalist and conservative Protestant denominations in Korea. Third, we attempted to examine overall tendencies of Korean Protestants in their attitudes toward public disease control and in-person service attendance with limited survey data containing 1000 responses from Protestants only; thus, we cannot generalize our findings to other religious groups, such as Catholics or Buddhists. Fourth, in future surveys, the “others” denomination must be clarified by allowing respondents to fill in the parentheses. Lastly, in the future, we recommend qualitative studies, especially critical discourse analysis, on church cluster cases to explore the in-depth meaning of their intentionality and public actions.