Introduction

A substantial body of research has investigated the links between religious involvement and mental health, focusing attention on distress, depression, anxiety, and other affective outcomes (Koenig et al. 2001; Smith et al. 2003). Although the results are not univocal, the weight of the evidence suggests salutary effects of religiousness on mental health. Early studies centered on generic religious measures, especially (a) the frequency of religious practices, and (b) religious orientations or motivations. More recent work has refined the conceptualization and measurement of religiousness (Ellison and Levin 1998). Specifically, a number of studies have examined the influence of (a) functional aspects of religious involvement (e.g., coping practices, congregational support, other religious resources) (Ano and Vasconelles 2005; Hill and Pargament 2003; Krause 2008), and (b) spiritual struggles (e.g., strained relationships with God, negative interaction in church, chronic religious doubt) on mental health (Exline and Rose 2005; Pargament et al. 2005).

Although most studies of religion and mental health have used clinical or community-dwelling samples of adults, there also exists a literature on these issues among clergy (Hall 1997; Weaver et al. 2002). This is an important area for several reasons. First, the psychological functioning of clergy has direct bearing on their job performance and the quality of their personal and family life (Blanton 1992; Coate 1989; Hill et al. 2003). Moreover, there is perennial concern about clergy burnout (Kaldor and Bulpitt 2001; Turton and Francis 2007) because significant numbers of new pastors leave the clergy within the first few years following seminary (Hoge and Wenger 2005; Miner 2007a, b). Though there are many reasons for this, psychological strain is thought to be a contributing factor, exacerbating a shortage of congregational ministers in many Christian denominations (Hoge and Wenger 2005; Judd et al. 1970).

Much of the work in this area has built on models of job burnout (Maslach et al. 2001) that focus on the fit between the personality characteristics and the demands of congregational ministry (Francis and Rodger 1994; Francis et al. 2008; Miner 2007a, b). Although this is a valuable approach, we take a different tack in this study, capitalizing on developments in: (a) the conceptualization and measurement of health-relevant domains of religiousness (e.g., Idler et al. 2003), (b) growing interest in the religious and spiritual sources of clergy resilience as well as strain (e.g., Meek et al. 2003), and (c) the marriage of religion-health research with the broader life-stress (or stress-and-coping) paradigm (Ellison et al. 2001; Fabricatore et al. 2004).

Theoretical and empirical background

Religious resources

Religious belief systems and communities offer an array of resources that contribute to psychological well-being (for review, see Ellison and Levin 1998). For example, religious teachings and congregations may shape individual behaviors and lifestyles in ways that reduce the risk of various stressful events and conditions. Religion also promotes various psychological resources, such as feelings of self-esteem and mastery, as well as positive emotions and beliefs such as meaning, forgiveness, optimism, and gratitude. However, two of the most important and well-researched resources involve (a) specific religious coping styles and practices and (b) informal congregational support systems.

Pargament and his associates have identified several religious coping styles that are consistently associated with desirable mental and physical health (Pargament 1997; Pargament et al. 2000). In particular, individuals who use more active religious coping, especially those who form collaborative partnerships with God to deal with problems tend to fare especially well. By contrast, persons who engage in passive or deferential modes of religious coping, ceding control over (and responsibility for) coping to God, typically exhibit unfavorable psychological and physical health outcomes (Pargament et al. 1998). The benefits of collaborative religious coping practices have been found among college students (Pargament et al. 1994; Mickley et al. 1998), Protestant church members (Pargament et al. 1998), flood survivors (Smith et al. 2000), caregivers of Alzheimers patients (Shah et al. 2001), and others. Indeed, a meta-analysis of 49 studies confirmed a moderately strong association between such positive, collaborative forms of religious coping and mental well-being (Ano and Vasconcelles 2005).

Although a voluminous literature exists on social relationships and health (Cohen 2004), only recently have researchers examined the link between congregational support systems and mental and physical well-being (Krause 2008). Two facets of church-based social support appear to be particularly important for well-being: actual support, and anticipated support--i.e. support that is expected to be available if needed. This may not be surprising, since religious rhetoric and ideology emphasize solidarity and fellowship, and prosocial behavior is expected and esteemed (Ellison and George 1994). Research by Krause (2008) has linked congregational support--especially anticipated support--with psychological well-being, primarily among older adults. Among other things, Krause found informal support from church members had a greater effect on health and well-being among older adults than support from secular social relationships.

Spiritual struggles

Though contemporary research on religion and mental health has emphasized the salutary effects of religious involvement, a long tradition of theory holds that certain facets of religious belief and experience may have deleterious effects on mental well-being. In recent years, a small but growing literature on “spiritual struggle” has provided support for this view. According to Pargament et al. (2005), spiritual struggle refers to “efforts to conserve or transform a spirituality that has been threatened or harmed” (p. 247), and he and his colleagues have identified three types of spiritual struggle: (a) divine struggle, or troubled relationships with God; (b) interpersonal struggle, or negative interaction in religious settings; and (c) intrapsychic struggle, or chronic religious doubting. The incidence of spiritual struggle tends to be low within community-dwelling samples of US adults (Idler et al. 2003), but notably higher among clinical samples (Fitchett et al. 2004).

One important aspect of personal religious life is the cultivation of a direct, intimate relationship with a (perceived) divine other (e.g., God, Jesus). Not all relationships with God are healthy or constructive, however. Recent studies indicate that some individuals experience trouble or strain in their relationship with God, particularly during personal crises or other difficulties. These individuals may come to question the beneficence of God (How could a good, caring God allow such problematic events to occur?), or the scope of divine powers (Does God really have the power to control such events?), or even the existence of God (How can such things happen if there really is a God?). They may even wonder if God is punishing them, and they may become angry at God for allowing misfortune to befall them. Struggles in one’s relationship with the divine are associated with a range of mental health outcomes, such as elevated rates of anxiety (Ano and Vasconcelles 2005), and depression and suicidality (Exline et al. 1999; Exline et al. 2000).

Interpersonal spiritual struggles are evident within religious communities. Unpleasant encounters among congregants may involve a range of issues; they may pertain to theological or doctrinal differences, disputes over clergy leadership or the administration of church affairs, or political issues such as war or homosexuality (Krause et al. 2000). Some, however, may have minimal religious content, instead involving mundane or petty squabbles that happen to occur in church rather than in other social settings. Several studies have shown that negative interactions among church members have undesirable implications for personal well-being (Ellison et al., in press; Krause et al. 1998) that may outweigh the positive implications of church-based social support. Although researchers have documented the deleterious psychosocial consequences of negative interactions in secular settings for some time, only recently has research examined the effects of such noxious encounters within congregations.

Intrapsychic spiritual struggle may take the form of chronic religious doubting. The struggle to sustain faith and religious belief has long intrigued theologians, and it has elicited recent attention from social and behavioral scientists (Krause and Ellison 2009). Though some have argued that a degree of doubt is essential for the maturation of religious faith, chronic, unresolved doubts can be profoundly disconcerting for religious adherents. Doubts can deprive the individual of a coherent religious belief system and sense of coherence, which otherwise provide a means of interpreting and assigning meaning to daily affairs and personal crises alike. Since open questioning and skepticism are discouraged within many religious groups, individuals who experience such doubts may be reluctant to express them, heightening their sense of isolation and struggle. The findings of several studies are consistent with this line of argument, showing that religious doubts are associated with elevated levels of psychological distress and psychiatry symptoms, and inversely associated with life satisfaction and other indicators of well-being (Ellison 1991; Galek et al. 2007; Krause et al. 1999; Krause 2006; Krause and Wulff 2004).

Relevance for clergy mental health

Although a number of studies have focused on clergy mental health (Hall 1997; Lewis et al. 2007; Weaver et al. 2002), few have investigated the roles of spiritual struggles or the religious resources, such as congregational support. As religious specialists, clergy may be particularly skilled at constructive forms of religious coping (Meek et al. 2003), and may be well-positioned to elicit various kinds of support from members of their congregations (Hammond et al. 1978; Finke and Dougherty 2002). Krause et al. (1998), for instance, found that Presbyterian (PCUSA) clergy members reported somewhat higher levels of congregational support than church elders, who, in turn, reported greater congregational support than rank-and-file church members. A more recent study by Pargament et al. (2001) found that Presbyterian clergy engaged in positive or collaborative forms of religious coping more often than elders or rank-and-file members.

These same studies, however, also found negative influences on pastors’ psychological well-being. Pastors reported much higher levels of negative interaction within the church than elders or rank-and file members, the deleterious effects of which were significantly greater for pastors than elders or members (Krause et al. 1998). Pastors in the later study experienced strained relationships with God just as often as their counterparts in other church roles, and these divine struggles significantly contributed to psychological distress (Pargament et al. 2001).

According to identity theory, individuals construct personal identities based largely on the social roles they occupy (Burke 1991; Stryker 1987). These roles exist in hierarchies, with social roles in which a person has made greater temporal, financial, and psychological investments being more central to an individual’s identity than other roles. Clergy make significant investments in learning about their faith tradition, and their work life and occupational identity is oriented around their religious community (Finke and Dougherty 2002; Ingram 1981; Mueller and McDuff 2004; Perl and Chang 2000). As clergy, they play many different roles that involve many distinct activities and responsibilities, ranging from preaching and teaching, to planning and administering congregational affairs, to pastoral visitations and counseling (Francis et al. 2008; Kay 2000; Monahan 1999; Rowell 2000).

Identity theory further states that developments that challenge an individual’s occupation, performance, or competence in highly salient roles may be threatening to the sense of self, and can give rise to feelings of distress or other undesirable psychological states. Viewed from this perspective, one can see why spiritual struggles might be especially disturbing for clergy members (Krause et al. 1998). Given their training, personal faith, and professional dedication, clergy members are often esteemed as religious leaders and exemplars by others in their congregation, and they are likely to be highly invested in their faith and occupational role (Grosch and Olsen 2000; Nauta 1988). Thus, it may be especially stressful to face a troubled or strained relationship with God, to deal with criticism or other interpersonal challenges within the congregation, and to deal with prolonged, unresolved doubts about one’s faith. Such spiritual struggles raise the specter of failure in a highly important social role, and these strains might well spill over into other domains, such as marital or family life.

Stress, religion, and mental health: Three conceptual models

A number of recent studies have integrated insights from the social psychology of religion with the life-stress (or stress-and-coping) paradigm, which guides much of the contemporary work on social patterns and interpersonal variation in health and illness (Ellison et al. 2001; Pearlin 1989). This paradigm is based on at least three key assumptions. First, social stressors—e.g., negative life events or chronic conditions that tax the ability of the individual to respond—undermine health and well-being. Second, social and psychological resources assist the individual to resolve social stressors and deal with negative emotions arising from them. Third, individual variations in well-being can be partly explained in terms of differential exposure and differential vulnerability to social stressors. Religious involvement can provide personal resources that facilitate successful coping with stress, but can also create potential sources of stress that can impair psychological functioning (Ellison et al. 2001; Fabricatore et al. 2004).

Several relevant conceptual models of the relationships among stressors, religion, and well-being have been identified. In the first and simplest model, religious resources and spiritual struggles are unrelated to (or minimally correlated with) stressors. Whereas stressors and spiritual struggles undermine psychological functioning, religious resources are positively associated with well-being. This is termed the “main effects” model, because the hypothesized salutary effects of religious resources and the deleterious consequences of spiritual struggles are additive, rather than indirect or multiplicative.

Two other models propose a more complicated set of effects among these variables, which take different forms. For example, individuals confronting stressful events may turn to religious resources, engaging in more frequent collaborative religious coping and eliciting greater social support from church members. To the degree these religious resources are helpful, they may mask the “true” effect of stressors on well-being. Thus, when variations in religious coping and support are controlled, the net effect of the stressors will be increased, illustrating the extent of their deleterious impact on mental health. This is called the “suppressor” model.

Alternatively, stressors may make it more difficult to sustain collaborative partnerships with God, and may lead individuals to overtax and eventually exhaust their congregational support. Stressors may also increase levels of spiritual struggle, fostering a strain in one’s relationship with God, increase negative interaction with coreligionists, and triggering religious doubts. In this case, the net effects of stressors will be reduced when individual differences in religious resources and spiritual struggles are controlled, because one of the ways that stressors undermine well-being is indirect, e.g. by decreasing religious resources and increasing spiritual struggles. This is called the “indirect effects” or “mediator” model.

The final set of models imply yet a different set of relationships, in which the magnitude of the noxious effects of stressors varies according to levels of religious resources and spiritual struggles. Collectively, they are termed “contingent effects” or “interactive” models, of which there are two versions. Although stressful events and conditions tend to undermine well-being, this effect may be limited (or even nullified) for persons who practice collaborative religious coping and enjoy active congregational support networks. However, stressors may take a much greater toll on well-being for individuals who lack these valuable religious resources. In this “stress-buffering” model, religious resources mitigate the harmful psychosocial effects of stressors. On the other hand, the effects of stressors may be muted for persons who do not experience spiritual struggles. However, because such struggles can constitute major stressors in their own right, the harmful impact of stressful events may be markedly greater among persons who are experiencing strained relationships with God, negative interactions in church, or chronic doubts about their faith. This is termed a “stress-exacerbating” model.

Data

To examine these issues, we analyze data from a nationwide survey of active clergy--i.e., persons who are ordained as ministers of Word and Sacrament--in the Presbyterian Church USA. This sample was drawn in the fall of 1996, and data were collected via mail questionnaire in February, 1997. An initial sample of 2,530 ministers was drawn from the list of active clergy that is maintained by the national PCUSA office. Approximately 75% (n = 1,870) of these ministers responded to the initial survey. Of these clergy members, roughly 76% returned the questionnaire on Spirituality and Health, on which the present study is based. After listwise deletion of cases in which some variables had missing values, we are left with an effective sample of 1,272 clergy members for our analyses.

Measures

Dependent variables: Psychological well-being and distress

Clergy mental health was assessed using nine items from the SF-36 Health Survey (Ware and Sherbourne 1992). These questions ask about positive and negative affect during the 4 weeks prior to the survey; five of the items are keyed in a negative direction and reflect feelings of distress, while the remaining four items are positively scored and assess a positive affect. Specifically, each respondent was asked how often s/he: (a) was a very nervous person; (b) felt so down in the dumps that nothing could cheer [him/her] up; (c) felt downhearted and blue; (d) felt worn out; (e) felt tired; (f) was a happy person; (g) felt calm and peaceful; (h) had a lot of energy; and (i) was full of pep. For each item, response categories range from (1) none of the time to (6) all of the time. To calculate scales tapping distress (alpha = .80) and well-being (alpha = .85), we summed across the relevant items and then divided by the number of items that the respondent had answered.

Religious resources

Our analyses of clergy mental health consider three types of religious resources: positive religious coping practices, emotional support from church members, and anticipated support from church members. First, respondents were asked how much they relied on each of several religious coping strategies in the process of grappling with their most recent serious problem. Positive religious coping items included the following: (a) I tried to find a lesson from God in this problem; (b) I tried to give spiritual strength to other people; (c) I thought about how my life is part of a larger spiritual force; (d) I worked together with God as partners to get through this problem; and (e) I looked to God for strength, support, and guidance. These statements were developed by Pargament (1997; Pargament et al. 2000) to reflect a “collaborative” religious problem-solving approach. Response categories range from (1) not at all to (4) a great deal. Scores on the positive religious coping scale (alpha = .80) were produced by summing across the items and then dividing by the number of items to which the individual responded.

Information about emotional support was elicited by asking respondents to think about the relationships they have with others in their congregation. They were then asked how often during the preceding year those people had (a) made you feel loved and cared for and (b) listened to you talk about your private problems and concerns. Response categories range from (1) never to (4) very often; our measure of emotional support is based on the mean score on these two items (r = .51, p < .001). Anticipated support was measured by asking how much help people in the congregation would be willing to provide if needed (Idler et al. 2003). Specific questions were: (a) If you were ill, how much would the people in your congregation be willing to help out? (b) If you had a problem or were faced with a difficult situation, how much comfort would the people in your congregation be willing to give you? Response categories range from 1 (not at all) to 4 (a great deal). The measure of anticipated support is calculated by using the mean response on these two items (r = .73, p < .001).

Spiritual struggles

We examine three facets of spiritual struggle: divine, interpersonal, and intrapsychic. The first of these types of struggles involve strained relationships with God, and are gauged in terms of problematic religious coping experiences. Individuals were asked how much they relied on each of the following approaches in the process of dealing with their most recent major problem: (a) I felt that this problem was God’s way of punishing me for my sins or lack of spirituality, (b) I wondered whether God had abandoned me, (c) I questioned whether God really exists, and (d) I expressed anger at God for letting this problem happen. These items were designed by Pargament (1997; Pargament et al. 2000) to measure problematic styles of religious problem-solving that involve strained relations with God. Response categories range from (1) not at all to (4) a great deal. This measure, which we will term “conflicted religious coping,” was calculated by summing across the items and dividing by the number of items answered by the respondent (alpha = .51).

Interpersonal struggle was measured via two items on negative interactions within the congregation (Idler et al. 2003; Krause et al. 1998). Clergy respondents were asked: Over the past year, how often have people in your congregation (a) ... made too many demands on you? And (b) ... been critical of you and the things you have done? Response categories range from 1 (never) to 4 (very often), and our measure of negative interaction is based on the mean response to these two items (r = .56, p < .001). Intrapsychic struggles, gauged in terms of religious doubts, are measured via items tapping the frequency with which (a) personal suffering, (b) conflicts of faith and science, and (c) the feeling that life really has no meaning have caused doubts about the respondent’s religious faith (Ellison 1991; Krause et al. 1999). Response categories range from 1 (never) to 3 (often), and our doubts measure is based on the mean response to these three items (alpha = .57).

Stressful life events

Respondents were asked whether each of the following types of problematic events had happened to them during the year preceding the survey: (a) death of spouse, child, parent, grandparent, or grandchild; (b) major financial loss that involved 20% or more of family income; (c) serious illness or accident involving spouse, child, parent, grandparent, or grandchild; (d) major disagreement or serious argument with spouse, child, parent, grandparent, or grandchild; or (e) major disagreement or serious argument with close friend. In addition, respondents were asked whether they experienced “any other major problem or challenge” during this period. Our measure of stressful life events is the sum of affirmative answers to these items.

Control variables

All multivariate models include statistical adjustments for the following covariates: age (measured in years); gender (1 = female, 0 = male); race (1 = nonwhite, 0 = white); and marital status (1 = married, 0 = unmarried).

Results

Descriptive statistics on all variables used in our analyses are presented in Table 1. Briefly, mean levels of psychological distress (2.158) are well below the midpoint of the scale, while average levels of psychological well-being (4.155) are above the midpoint of the scale. Clergy members also tend to exhibit high levels of anticipated support (3.612); levels of received emotional support (2.900) and collaborative religious coping practices (2.964) are above the midpoints of the respective scales. Although conflicted religious coping is uncommon in this sample, it appears that non-trivial numbers of clergy members may be experiencing intrapsychic struggle or chronic religious doubting. Further, average levels of negative interaction are only slightly below the scale midpoint (2.142), indicating that some clergy respondents face excessive demands and criticisms from church members on at least a semi-regular basis. Finally, the average PCUSA clergy respondent is 50.24 years of age and experienced more than one (1.367) major negative life event during the year prior to the survey. Of these respondents, 19.9% are female, 12.7% are unmarried, and 5.6% are from nonwhite racial/ethnic backgrounds.

Table 1 Clergy response to overall stress and religious variables (N = 1272)
Full size table

To examine the relationships between stressful life events, religious resources, spiritual struggles, and clergy mental health, we estimate a series of four ordinary least squares (OLS) regression models for each outcome. Model 1 includes sociodemographic control variables and stressful life events only. Model 2 adds our indicators of religious resources (positive religious coping, emotional support received, and anticipated support), while Model 3 replaces those variables with the indicators of spiritual struggle (conflicted religious coping, negative interaction, and religious doubting). The final model, Model 4, includes all of these variables simultaneously.

The multivariate models predicting clergy well-being are displayed in Table 2. Several findings are especially noteworthy. In Model 2, all religious resource variables bear a positive association with psychological well-being. Of these, collaborative religious coping has the strongest estimated net effect on well-being (b = .216, β = .173, p < .001). In substantive terms, that estimate indicates that for each one unit change in collaborative coping there is an associated average increase of 0.216 on our measure of well-being. In terms of standard deviations, one standard deviation change on collaborative coping indicates a 0.173 standard deviation change on well-being.

Table 2 Main effects of religious variables on clergy psychological well-being (N = 1272)
Full size table

Other religious resources, such as anticipated support from church members (b = .201, β = .142, p < .001) and emotional support received (b = .152, β = .125, p < .001) also bear a positive relationship with well-being. Taken together, the inclusion of religious resource variables in Model 2 accounts for roughly 10% of the variance in well-being, over and above the baseline model (Model 1). The addition of these religious resource variables does not alter the estimated net effect of stressful life events, which remains a potent negative predictor of clergy well-being (b = −.148, β = −.200, p < .001).

In Model 3, each of the spiritual struggle variables is significantly predictive of psychological well-being, although the estimated net effects and overall variance explained are notably weaker than those in Model 2. Comparing the standardized coefficients (β) we find that religious doubting is the strongest of the spiritual struggle variables in Model 3 (b = −.280, β = −.130, p < .001), followed by conflicted religious coping (b = −.367, β = −.118, p < .001) and negative interaction with church members (b = −.099, β = −.073, p < .01), respectively. Model 3 accounts for an additional 4.5% of the variance in clergy psychological well-being, above and beyond the baseline model. In contrast to the findings in Model 2, however, the inclusion of spiritual struggle variables in Model 3 results in a notable reduction of approximately 17% in the estimated net effect of stressful life events (b = −.121, β = −.164, p < .001) on well-being. This pattern is consistent with the mediator model outlined earlier, in which part of the influence of stressful life events on well-being may be due to the tendency for clergy members who have undergone recent traumas to experience elevated religious doubts or strained relationships with God.

Model 4, the full model, estimates the net effects of all religious resource and spiritual struggle variables on well-being. Findings are broadly consistent with those reported in Models 2 and 3. Positive religious coping is the most potent of the religious predictors (b = .211, β = .169, p < .001), followed by emotional support received from church members (b = .150, β = .123, p < .001), and religious doubts (b = −.248, β = −.115, p < .001). The other theoretically relevant variables remain significant--albeit weaker--predictors of well-being in the full model. Interestingly, the estimated net effect of anticipated support from church members is reduced by around 25% when spiritual struggles are controlled in Model 4, as compared with Model 2. This pattern suggests that clergy members who are experiencing spiritual struggles tend to harbor greater skepticism about the potential reliability of support from church members. Taken together, the final model “explains” slightly more than 20% of the variance in clergy well-being, with the additive effects of religious variables combining to account for around 12.5% of this variance.

Next we turn to the models predicting psychological distress, which are displayed in Table 3. Compared with the baseline model (Model 1), the inclusion of religious resource variables in Model 2 adds roughly 4% to the explained variance in distress. Two of the three religious resource indicators are significantly linked with distress: anticipated support from church members (b = −.190, β = −.167, p < .001) and collaborative religious coping (b = −.096, β = −.096, p < .001). Emotional support received from church members is entirely unrelated to feelings of distress among the clergy. Controlling for religious resource variables yields no meaningful change in the estimated net effects of our stressful life events measure in Model 2.

Table 3 Main effects of religious variables on clergy psychological distress (N = 1272)
Full size table

Model 3 reveals that each of the spiritual struggle variables is positively associated with distress; of these predictors, conflicted religious coping bears the strongest association with distress (b = .508, β = .205, p < .001), followed by negative interactions with church members (b = .174, β = .160, p < .001) and religious doubts (b = .242, β = .141, p < .001). Taken together, the linear combination of spiritual struggle variables approximately doubles the predictive power of the baseline model; indeed, Model 3 accounts for 22% of the variance in clergy distress. Moreover, controls for these facets of spiritual struggle also reduce the estimated net effect of stressful life events, as compared with Model 1, by roughly 26% (to b = .108, β = .182, p < .001). This pattern is consistent with findings reported in Table 2, suggesting that a portion of the deleterious emotional effects of stressful events may be due to elevated levels of spiritual struggle among those who have experienced them. Findings in the full model, Model 4, are very similar to those presented in Models 2 and 3. The main exception again involves the estimated net effects of anticipated support on clergy distress, which are reduced by nearly one-third (to b = −.130, β = −.114, p < .001) when spiritual struggle variables are added to the model. The full model explains nearly one-quarter (24%) of the variance in psychological distress.

The models to this point have tested the offsetting and mediating effects models of the relationships between stress, religious and spiritual factors, and clergy mental health. As a final step, we added cross-product interaction terms (i.e., Religious Resources × Overall Stress, Spiritual Struggles × Overall Stress) to the full models in Tables 2 and 3. Prior to the calculation of interaction terms, variables were zero-centered to minimize collinearity between raw and product terms (Aiken and West 1991). Significant results are summarized in Table 4. Briefly, the stress-buffering and stress-exacerbating models outlined earlier find only limited support here. With regard to religious resource variables, only one of six possible interactions is statistically significant. In that equation, the deleterious effects of stressful life events on clergy well-being are reduced among those persons who also report collaborative religious coping styles (b = .065, p < .05). However, collaborative religious coping does not moderate the effects of stressors on feelings of distress, nor do other religious resource variables buffer the effects of stressors on clergy psychological functioning.

Table 4 Interactive effects of religious variables and stress on clergy psychological well-being and distress (N = 1272)
Full size table

We also find only one interactive relationship involving stressful events and spiritual struggles in predicting mental health outcomes. Negative interactions with church members appear to exacerbate the harmful consequences of stressors on the psychological distress of clergy members (b = .073, p < .001). However, none of the other spiritual struggle variables exacerbate the effects of stressful events on distress or well-being.

On a final note, these analyses reveal few effects of control variables on clergy mental health. As noted above, the number of stressful life events is strongly related to well-being and distress. In the only consistent finding involving sociodemographic predictors, on average older clergy members enjoy higher levels of well-being and lower levels of distress than their younger counterparts. In some (but not all) models, men report greater well-being than women, although no gender differences in distress were found. There are no meaningful effects of race or marital status on either outcome.

Discussion

This study has examined the complex relationships between stressful events, religious variables, and two dimensions of clergy mental health: psychological well-being and distress. Our work has contributed to the body of work on clergy mental health by focusing on both religious resources (e.g., collaborative coping, congregational support) and spiritual struggles (e.g., conflicted coping, chronic doubts) as they bear upon clergy functioning. Moreover, our study has explored three conceptual models of the links between stressors, religion, and mental health: (a) main effects; (b) indirect effects; and (c) interactive effects. Several findings are particularly noteworthy.

Overall, our results indicate strong support for the main effects model, in which stressors, resource, or struggle variables have independent relationships with the outcomes. Indeed, in all models reported here, stressful life events appear to take a toll on clergy mental health. In 5 of 6 instances, religious resource variables are associated with salutary psychological outcomes, partly offsetting the deleterious impact of stressors. In all 6 instances, spiritual struggle variables are linked with undesirable outcomes, i.e., adding to the harmful influence of stressors on clergy mental health. Taken together, our measures of religious resources and spiritual struggles account for 12.5% of the variance in well-being outcomes and 9% of the variance in distress outcomes, over and above the estimated effects of background predictors. Though these R2 values are modest in absolute terms, they are substantial in the context of data gathered on a large, heterogeneous nationwide sample such as the one employed in this study. Although broadly similar patterns emerge for both mental health outcomes, religious resources tend to predict well-being better than distress, while spiritual struggles generally predict distress better than well-being. It is unclear whether this results from shared method variance or other artifactual reasons, but that issue warrants additional study in the future.

There is also limited support for the indirect effects model. To be sure, we find no evidence that religious resources mediate the links between stressful events and distress, or that they mask the deleterious effects of stressors on well-being. However, our results do show that spiritual struggles partly mediate the link between stressors and psychological functioning. This raises the intriguing and important possibility that these stressors may undermine clergy mental health partly by fostering or deepening the sense of spiritual struggle. In other words, pastors who experience negative events, especially multiple stressors within a given year, are more inclined than others to encounter spiritual difficulties, coming to doubt core tenets of their faith or to feel abandoned by, and angry toward, God in the wake of these traumatic events. Indeed, this may be one way in which stressful events influence mental health outcomes. On the other hand, there is only weak support for the interactive approach; only 1 of 6 possible stress-buffering effects and only 1 of 6 possible stress-exacerbating effects emerged as significant in these data.

These findings have several implications. First, with few exceptions, religious resources and spiritual struggles are independent predictors of distress and well-being. Each of these variables is germane to clergy mental health. Active PCUSA clergy benefit from collaborative problem-solving relationships with God, and they also receive significant emotional support from members of their congregations. In addition, these pastors typically have optimistic views of their church-based social networks, anticipating that members of their congregations could be counted on to offer emotional and material assistance if ever called upon. The latter finding is particularly interesting since most research has tended to view congregants more as sources of stress than sources of support (e.g., Han and Lee 2004; Lee 1999; Morris and Blanton 1998; Mueller and McDuff 2004). Indeed, there is only one previous study of which we are aware that analyzed the effects of social support from congregants, and that study found no significant association between congregational support and clergy’s psychological well-being (Lee 2007).

For the most part, these religious resources operate similarly in shaping psychological functioning for ministers regardless of the degree of their recent exposure to negative events. These patterns resonate with research on elders and rank-and-file laypersons within the PCUSA (Krause et al. 1998; Pargament et al. 2001), as well as various other samples, and with qualitative studies of clergy resilience in other denominations (Meek et al. 2003).

Our work also contributes to a growing body of evidence concerning the association between three types of spiritual struggle (divine, interpersonal, and intrapsychic) and mental health. Few previous studies have focused on such struggles among clergy members (for a partial exception, see Meisenhelder and Marcum 2004). Troubled relationships with God (e.g., feelings of divine abandonment, anger toward God, etc.) are likely to be very detrimental to clergy in that Meek et al. (2003) found that a close connection with God was a major element in clergy’s ability to cope with stress. Moreover, according to Guthrie and Stickley (2008), clergy make “no clear distinction between spiritual and mental distress” (p. 395).

Unresolved religious doubts may be especially challenging for clergy members, many of whom have devoted long periods of their lives to theological training, reflection, and spiritual leadership. For them, spiritual struggle may be non- (or counter-) normative, and may threaten fundamental aspects of personal identity. Further, they may raise important issues of competence and confidence; pastors may find it difficult to counsel members of their flock effectively if they are experiencing internal religious conflict and turmoil themselves. Moreover, while laypersons can bring religious concerns and questions to their pastors, clergy members may be reluctant to discuss their own spiritual issues with others, thus making these burdens even lonelier and more disturbing. Finally, a comparison of the univariate data presented in Table 1 with data from (a) companion samples of PCUSA elders and rank-and-file laypersons (Krause et al. 1998; Pargament et al. 2001) and (b) US adults in the 1998 NORC General Social Survey (Idler et al. 2003) reveals that PCUSA pastors report relatively high levels of spiritual struggle, particularly the frequency of negative interactions with church members. The prevalence and consequences of these three types of struggles among clergy and other religious professionals clearly warrant further investigation.

It is important to acknowledge several study limitations. First, this research is based on cross-sectional data, and it is not possible to establish the causal direction of associations from these models. Strong causal inferences would ideally require multiple (i.e., three or more) waves of longitudinal data collected over a significant period of time. Second, our sample of ministers is drawn from a single denomination, the Presbyterian Church (USA), which is distinctive in several ways: (a) it faces declines in overall membership; (b) it disproportionately comprises older members; (c) it faces internal political (and to a lesser extent theological) rifts that may affect the functioning of congregations and higher levels of denominational organization; and (d) it has experienced recurrent clergy shortages. Future research might profitably include clergy from a range of religious groups and theological traditions to overcome this limitation. Third, although much of the previous literature on clergy mental health has centered on associations between personality, job characteristics, and burnout, we do not have data on these issues. In the future, it would be helpful to include appropriate measures of these constructs along with the religious resources, spiritual struggles, stressors, and psychological outcomes examined here. This would allow investigators to gauge the degree of overlap (or orthogonality) between personality, religious resources, and spiritual struggles.

Our measures of stressors consist of major life events (e.g., bereavement, serious interpersonal conflicts, financial loss) that occurred within the year preceding the survey. However, there are sound reasons to anticipate that chronic stressors may take an even greater toll on clergy mental health. A substantial body of work has focused on chronic stressors that may be intrinsic to this occupation, particularly boundary-related strains such as the “stained glass fishbowl,” administrative demands, pressures on spouses and other family members (Coate 1989; Hill et al. 2003; Lee 1999; McMinn et al. 2005). The next logical step in this line of work is to explore the roles of religious resources and spiritual struggles in mediating, buffering, or exacerbating the effects of such chronic stressors on clergy mental health.

Although beyond the scope of this study, it would also be useful to investigate the availability and support provided by fellow clergy, either ordained members of the pastor’s own congregation or other ministers in the community. If accessible, support from other clergy may be especially beneficial. Studies in other areas show that support relationships may benefit from shared statuses and common experiences and problems (e.g., Suitor et al. 1995). This could make fellow clergy better sources of understanding and advice than laypersons or others who have not experienced strain or burnout from work in helping professions. There seem to be only two studies that have looked at peer support and mental health among clergy. One, which is a study of Dutch Reformed pastors (Evers and Tomic 2003), found that social support from peers was associated with lower levels of emotional exhaustion and depersonalization, and higher levels of personal accomplishment. The second is a study of three groups of Catholic clergy: parish (or secular) priests, priests belonging to a religious order, and monks (Virginia 1998). In that study, social support from peers was negatively correlated with depressive symptoms in all three groups, although the correlations were statistically significant only for parish priests and monastic clergy.

We believe that our study offers fresh evidence that both religious resources and spiritual struggles (e.g. collaborative religious coping, congregational support) shape clergy mental health. As researchers continue to investigate multiple facets of clergy functioning (e.g. troubled relationships with God, negative interactions, and chronic doubting), closer attention to these factors should enhance our understanding of strain, resilience, and well-being. Moreover, these findings may also be helpful for persons engaged in counseling and in designing interventions to assist ministers and their families. Further research along the lines sketched above can further illuminate the crucial issues of strain and psychological functioning among clergy members from diverse religious and social backgrounds.