Searching for comfort in religion: insecurity and religious behaviour during the COVID-19 pandemic in Italy

The multifaceted impact that the COVID-19 pandemic is having on the global society is unprecedented. At the time of writing (early September 2020), the number of confirmed cases worldwide exceeds twenty-seven million and the number of deaths is about nine hundred thousand.¹ Moreover, according to June 2020 forecast of the World Economic Outlook, the real Gross Domestic Product (GDP) is expected to globally decrease by 4.9 per cent – the worst recession since the 1930s (International Monetary Fund 2020).

When facing such dramatic and life-threatening experiences, especially in the form of sudden shocks, people tend to use coping strategies to reduce their sense of insecurity. Religion may work as one of these strategies (Norris and Inglehart 2011). Was this the case during the COVID-19 pandemic? Was religiosity intended – and used – as a coping resource during the unfolding of the crisis?

In the months following the pandemic outbreak there has been a flourishing of reports on religiosity and COVID-19, most of them detecting a religious revival in several countries. On a global scale, Bentzen (2020) notes a 50% increase in Google search for topics related to prayer compared to the pre-Corona crisis period. On a smaller scale, one-quarter of U.S. adult population say their faith has become stronger because of the pandemic according to Pew Research Center survey data (Gecewicz 2020) while an Italian study found an increase in praying during the pandemic (Garelli 2020). There is also some contrasting evidence, e.g. in the Netherlands where the frequency of prayer seems to be unaffected by the COVID-19 crisis (Reeskens et al.2020).

Nonetheless, existent research did not allow providing direct evidence toward the individual relationship between existential insecurity, in terms of proximity to COVID-19 disease, and religious behaviour. By using original online survey data collected in Italy during the pandemic (ResPOnsE COVID-19 project), this article aims at addressing this issue. The main findings show a higher frequency of attendance at religious services (via web, radio and tv) and prayer among people who reported a contagion in their family. In addition, results are shown to hold especially for those who were religiously socialized, by suggesting that only people having already some religious tools can mobilize them when coping with difficult events. In the concluding section, we will provide some implications of these findings for the debate about secularization and religious change.

Among the several sociological theories aiming at identifying the culprits of the religious decline that is happening in western countries, the so-called existential insecurity theory proposed by Norris and Inglehart (2011) has recently attracted the interests of many scholars. According to their argumentation, the ‘feelings of vulnerability to physical, societal, and personal risks are a key factor driving religiosity’ (Norris and Inglehart 2011, 4) since religious beliefs, practice and ideologies can furnish people with predictable rules that are functional for coping with dangers and problems (Immerzeel and van Tubergen 2011). Conversely, when people live under conditions of greater security, they can tolerate more ambiguity and the need for rigid and predictable rules of conduct is less present (Norris and Inglehart 2011).

This body of knowledge is used by sociologists of religion to read religious change mainly as a consequence of the processes of modernization. According to this theoretical perspective, modernization has led to profound social changes ranging from the augmented rationality and economic security to the weakening of social ties (Ruiter and van Tubergen 2009). Among these changes, the increase of financial, social and existential security has resulted in a reduction of the need for religious reassurance (Höllinger and Muckenhuber 2019; Immerzeel and van Tubergen 2011; Molteni 2020; Norris and Inglehart 2011). The processes of secularization currently unfolding in the world can therefore be explained by augmented security that modernization brings with it.

When going deeper into the mechanisms behind this theoretical framework, the basic idea is that religious beliefs and behaviours are expected to have a positive role when experiencing insecurity. When confronted with life-threatening situations such as illness, war, death of a relative, but also unemployment or extreme poverty, religion can offer an effective way for dealing with such difficulties by providing psychological as well as social support. Accordingly, if the number of potential adverse situations decreases, the search for comfort in religiosity is expected to decrease as well. More precisely, the literature identifies two main dimensions of insecurity, the economic and the existential one, which refer to the exposure to financial and life-threatening risks respectively. Both of them are expected to enhance religiosity (Immerzeel and van Tubergen 2011).

The relevance of religious coping mechanisms is extensively argued in the psychological literature, which highlights that people experiencing insecurity tend to feel stressed and to lose the ability of taking control of their lives (Pargament 1997). In such a situation, religion finds a fertile ground. When discussing about the ways in which religion may turn useful in adverse situations, two different sets of benefits are usually mentioned. On the one side, religion can furnish cognitive and psychological support by means of myths and parables (Stolz 2009), shared worldviews (Zapata 2018) and sense of certainty (Brandt and Henry 2012). On the other side, religion can provide and strengthen social ties and religious communities where people can find comfort and support (Ellison and George 1994; Lim and Putnam 2010). These communities can also provide concrete help in several forms, such as housing, financial assistance and food provision (Stolz 2009; Storm 2017). A parallel branch of literature, defined as Terror Management Theory (TMT), outlines the defensive role played by religion against the terror of death (Solomon et al.1991; Vail et al.2010). According to that perspective, the awareness of death increases the likelihood of acceptance of ‘culturally familiar supernatural agents’ (Norenzayan and Hansen 2006, 183) and the rejection of culturally unfamiliar ones. Compared to secular beliefs, such religious beliefs are intended as ‘particularly well suited to mitigate death anxiety because they are all encompassing, rely on concepts that are not easily disconfirmed, and promise literal immortality’ (Vail et al.2010, 1).

Sociology of religion tends to focus on the long-term trends of religiosity and reads the changing (in)security as a slow and continuous phenomenon which is considered as one of the main drivers of the process of secularization and religious decline. But what happens instead if this process of slowly increasing security is disrupted by an external shock?

Only few contributions adopted this approach and found the positive role of religion in dealing with unpredictable events. Sibley and Bulbulia (2012) studied the effects of February 22nd 2011 Christchurch earthquake in New Zealand, by showing, coherently with the main hypotheses behind existential insecurity theory, an increase in religious faith among those who directly experienced the negative effects of the earthquake (see also Bentzen 2019), while a decline among those who did not. Analogously, Zapata (2018) studied the impact of climate disasters in Canada in the period 1992–2012, by concluding that human losses have increased the intensity of religion, but just among individuals who were already religious. When looking at the impact of economic insecurity, Chen (2010) reported a causal relationship between economic distress caused by the Indonesia financial crisis and religious intensity, measured by the attendance at Koran study and Islamic school. Finally, by showing that September 11st 2001 attacks exerted only a modest and short-lived effect on the American young adults’ religiosity, Uecker (2008) argued that such insecurity shocks cannot be responsible for remarkable – and stable – religious revivals.

The COVID-19 pandemic condensed together all the dramatic aspects of the unpredictable events just mentioned. From the feeling of lack of control typical of an earthquake or a climate disaster to the serious impact that a long period of lockdown had – and is having – on the economic conditions of the population. Within this worldwide situation, what happened in Italy is somehow the quintessence of the negative aspects linked to a crisis of this magnitude. Italy has been the first western country to be hit by the pandemic (Remuzzi and Remuzzi 2020), and this confronted all the citizens with the complete sense of unpredictability that comes from a new and unknown threat.

Besides being a collective experience, the COVID-19 pandemic also exposed people to different levels of closeness to the disease which are expected to reflect in different levels of existential insecurity. Indeed, people who reported contagions in their family more likely suffered from the worst consequences of the crisis given the closeness to the virus and the impact it has on the loved ones. Therefore, they might have developed a higher sense of existential insecurity. In line with the expectations of the existential insecurity theory, the first working hypothesis then stands as follows:

Moving the discussion a little further, many scholars who have analysed the role of religion as coping strategy (Pargament 1997; Swidler 1986) agree that religious socialization is a necessary precondition for religion to be used effectively. In other words, religion may work for coping only if it is already part of ‘the individual's orienting system’ (Pargament 1997, 144) or cultural toolkits (Swidler 2001). As also shown by empirical evidence mentioned above, only individuals who were already religious tend to mobilize some sort of religious strategies when coping with adverse events. Moreover, religious cultural tools are intended to exert a major function especially during what Swidler (1986) calls ‘unsettled times’, and the COVID-19 pandemic represents much more than the perfect example of these times.

For the purposes of this article we focus on religious socialization as a way to ‘fill the toolbox’ with religious coping strategies. As a matter of fact, it is well accepted that families and early-life religious communities are the main socialization agencies who are responsible for transmitting the religious way of reading the world (Hoge et al.1982; Kelley and de Graaf 1997; Storm and Voas 2012). Therefore, people who are not raised religiously are less likely to have the interpretative framework which is necessary to resort to religion in case of unpleasant events (Bruce and Voas 2016). These people will probably refer to coping strategies other than religion in unpredictable situations such as the COVID-19 pandemic. Hence, it is likely that people who received some sort of religious socialization in their childhood more likely search for comfort in religion when confronted with the worst consequences of the crisis. The second hypothesis thus stands as follows:

To be effective, the study of the potential use of religion in a situation of crisis needs to be based on data collected during the unfolding of the crisis itself. The present article analyses data coming from the ResPOnsE COVID-19 – University of Milan rolling cross-section survey carried out in Italy from April 6th to July 8th (Vezzoni et al.2020).² The online data collection relies on respondents selected from an opt-in community of a private company (SWG spa) and the final sample reproduces population distributions of gender, age class, and macro-area of residence. In addition to a core questionnaire aimed at collecting information on behaviour compliance, perceptions of risk, social, economic and political attitudes and evaluations, and a set of socio-demographic variables, rotating modules on specific topics were included in the survey. For the analysis here presented, we refer to a module on religiosity consisting in 11 questions that have been administered in the period from April 20th to May 15th, shortly after the peak of the pandemic. Overall, the sample for this specific module comprises 4573 individuals.³

Religiosity is here intended as religious behaviour, and two separate measures are considered: the frequency of attendance at religious services via web, radio and tv – as churches have been closed and mass celebration completely forbidden during the lockdown period – and the frequency of prayer. This strategy allows considering both the more institutional and intimate forms of behaviour. The two measures refer to individual conduct during the week before the interview and, for the ease of interpretation, are recoded in three categories. Attendance at religious services is coded as ‘more than once a week’ (12.0% of valid cases), ‘once a week’ (21.0%), and ‘never’ (67.0%) while prayer is coded as ‘daily’ (26.2% of valid cases), ‘once/more than once a week’ (21.9%) and ‘never’ (51.9%).⁴

The main independent variable measures the presence of contagions in respondents’ family networks. The presence of such a variable is relevant because it allows partially overcoming one of the main limits that scholars testing existential insecurity theory with survey data face: the difficulty to capture individual exposure to real traumatic life events (Höllinger and Muckenhuber 2019). Such a variable is dichotomous with value ‘1’ (18.3% of valid cases) to indicate the presence of at least a member of the family being in quarantine, infected, hospitalized, or dead, and value ‘0’ (81.7%) otherwise.

To measure religious socialization, we consider the frequency of church attendance at 12 years old, here coded in three categories: ‘at least once a week’ (68.5% of valid cases), ‘at least on holy days’ (9.7%), ‘less often/never’ (21.7%).

To account for the unequal distribution of closeness to the contagion and religiosity, we include in the analysis the following categorical control variables: gender (2 categories, ‘Male, ‘Female’), age (6 categories, ‘18-24’, ‘25-34’, ‘35-44’, ‘45-54’, ‘55-64’, ‘65 and more’), education (3 categories, ‘At most lower secondary’, ‘Upper secondary’, ‘Tertiary or more’), area of residence (5 categories, ‘north-west’, ‘north-east’, ‘centre’, ‘south’, ‘islands’) and week of interview (from 1 to 5). This last variable is particularly relevant because the health situation has got better with time, and this likely resulted in a progressively diminished need for coping strategies.⁵

Since the two dependent variables are strictly ordinal, hypotheses are tested by means of two ordered logistic regressions models (McCullagh 1980) repeated for each of the dependent variables:

• MODEL 1 assesses the effect of having contagions in the family on individual religiosity by controlling for sex, age, education, area of residence and week of interview.

• MODEL 2 assesses the interaction effect between having contagions in the family and church attendance at 12 years old on individual religiosity. This allows testing the effect of the proximity to the disease by different levels of religious socialization, accounting for the full set of control variables.

Depending on the form of religious behaviour considered, data provide mixed empirical evidence to Hypothesis 2. For what concerns the frequency of attendance at religious services (Figure 2 – Panel A), the presence of contagions in family does not have any effect for the people not religiously socialized, namely those with no or very rare church attendance at 12. Conversely, the exposure to the disease leads to higher likelihood of attendance for those who were somehow religiously socialized. This holds especially for those who received weak religious socialization: their probability of attending religious services respectively once a week and more than once a week is 8.9 and 6.3 percentage points higher when contagions in family are reported (a positive difference in probabilities of respectively 3.3 and 4.4 percentage points was found among people who attended religious services weekly at 12 years old). The weaker effect we found for the most religiously socialized individuals could be explained by their level of attendance, far higher than the level of the less socialized ones already before the pandemic. The findings confirm that individuals who do not have already some sort of religious reservoir do not rely on religion when dealing with situations impacting their existential security.

When considering prayer (Figure 2 – Panel B), the proximity to the disease also leads to a statistically significant higher frequency (both daily and once/more than once a week) for individuals with both a weak and a strong religious background. This is coherent with the results reported for service attendance. Conversely, also those without religious resources show a higher – although not statistically significant – likelihood of praying when close to the disease. This last difference in findings between prayer and attendance may depend on the measurement of religious socialization, based on an indicator (church attendance at 12 years old) much closer to the latter indicator of religious practice during the pandemic. Moreover, prayer may be an individualized practice that you do not necessarily need to fill with canonical content, so it is more easily available also to those not well acquainted with the institutional religious forms.

There is a common folk wisdom saying that ‘there are no atheists in foxholes’ meaning that everybody will refer to some sort of higher power when confronted with terrible and life-threatening events. This contribution has aimed at analysing the COVID-19 pandemic in Italy as one of these events.

In this respect, it offers first-hand original evidence on the short-term consequences of the COVID-19 pandemic on individual religiosity. Overall, people who suffered from the most severe effects of the crisis reported higher religious behaviour during the pandemic. More in detail, such a result mainly holds for people who received some religious socialization during their childhood. While this emerges very clearly for the more institutional form of religious behaviour, that is, attendance at religious services, the results are slightly less clear for individual prayer.

In wider terms, our contribution has specific implications for the debate about religious change. Findings show that a religious revival cannot be excluded when dramatic events happen. However, such a phenomenon is expected to have a negligible impact on the overall long-term trend of religious change and will be likely absorbed with the reduction of individuals’ existential insecurity. The use of religion as coping strategy has been shown to be relevant primarily for those who are already religiously socialized. Given that the amount of religiously socialized people will diminish together with the general religious decline (Storm and Voas 2012; Voas and Chaves 2016), the same is expected for those individuals who look at religion when experiencing existential insecurity. Thus, our results suggest that there are some atheists in the foxholes and also that their number is increasing over time.

Nonetheless, our analysis does not come without limitations. For what concerns the analytical strategy, we must be clear in saying that our findings say nothing about the effective psychological benefits of religious behaviour. In fact, an analysis of this kind should be based on ad-hoc psychological scales and this is far beyond the scope of this article. Also, panel data that measure religious behaviour before and during the pandemic would have allowed capturing individual religious change that cannot be analysed with the data here employed. Furthermore, religiosity and religious socialization are here measured by only two forms of religious behaviour: prayer and attendance to religious services. Therefore, our conclusions cannot be extended to the several other forms of religiosity which were not analysed here. Lastly, to timely respond to the crisis, the ReSPonSe COVID-19 survey is based on a non-probabilistic sample (see Vezzoni et al.2020 for more details). For what concerns our analysis, the presence of quotas by area, gender and age, together with the massive use of control variables, permits to strongly minimize the potential selection bias introduced by the non-probabilistic design. Also, the ResPOnsE COVID-19 sample turns out to be higher educated than the general population. If so, given the negative association between education and religiosity, our results could be conservative in terms of the estimation of the prevalence of both mass attendance and prayer.

Finally, we should highlight that these conclusions are limited to the Italian case. However, Italy presents some peculiarities that make it a very appropriate scenario for testing the existential insecurity theory. Firstly, it has been the first western country hit by the pandemic and, thus, Italians have probably suffered from the strongest sense of existential insecurity because of a virus previously unknown. Secondly, as religiosity in Italy is still widespread despite the ongoing process of secularization (Vezzoni and Biolcati-Rinaldi 2015) and, moreover, most people received religious socialization, this results in a huge basin for religious strategies. Thirdly, Christian religion tends to overperform in responding to events like a pandemic because of its strong emphasis on caring and charities compared to other doctrines (Stark 1996). Lastly, we cannot forget the role of the pope, who has been one of the most influent actors during the pandemic crisis.

To conclude, this article provides first evidence toward the use of religion as a coping strategy for those people who were personally close to the disease in a time of pandemic. Further research is welcome to analyse also the impact of economic insecurity on religiosity during and after COVID-19 emergency. Moreover, new research in the next years is required to analyse the long-term impact of COVID-19 on religiosity.

No potential conflict of interest was reported by the authors.

Cristiano Vezzoni is associate professor of political sociology at the University of Milan, Department of Social and Political Sciences.

Paolo Segatti is former full professor of political sociology at the University of Milan, Department of Social and Political Sciences.

Andrea Pedrazzani is assistant professor of political sociology at the University of Milan, Department of Social and Political Sciences.

Marco Maraffi is full professor of sociology at the University of Milan, Department of Social and Political Sciences.

Simona Guglielmi is assistant professor of sociology at the University of Milan, Department of Social and Political Sciences.

Giulia Maria Dotti Sani is assistant professor of sociology at the University of Milan, Department of Social and Political Sciences.

Antonio M. Chiesi is full professor of sociology at the University of Milan, Department of Social and Political Sciences.

Ferruccio Biolcati is associate professor of sociology at the University of Milan, Department of Social and Political Sciences.

Riccardo Ladini is post-doctoral fellow in sociology and political sociology at the University of Milan, Department of Social and Political Sciences.

Francesco Molteni is post-doctoral fellow and adjunct professor of sociology at the University of Milan, Department of Social and Political Sciences.