Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Rev Relig Res. Author manuscript; available in PMC 2010 Jun 17.
Published in final edited form as:
Rev Relig Res. 2009 Jan 1; 51(2): 181–200.
PMCID: PMC2886990

Close Companion Friends in Church and Health in Late Life


This study has two principal aims. The first goal is to empirically evaluate new measures of close companion friendships that arise in church. The second goal is to embed these measures in a conceptual model that seeks to assess the relationship between close companion friends at church and health. Based on data from a nationwide sample of older people, the findings reveal that the newly devised measures are psychometrically sound. In addition, the results provide empirical support for the following linkages that are contained in our conceptual model: older people who have a close companion friend at church are more likely to feel they belong in their congregation; old adults who believe they belong in their congregation are more likely to feel grateful to God; and older individuals who feel grateful to God tend to rate their health more favorably.


An impressive number of studies suggest that people who are more deeply involved in religion tend to have better physical and mental health than individuals who are not involved in religion (Koenig, McCullough and Larson 2001). Now researchers must explain why this may be so. Religion is a complex multidimensional phenomenon (Fetzer/National Institute on Aging Working Group 1999) and as a result, it is likely to affect health in a number of ways. A growing body of research supports this view. Some studies suggest that deleterious effects of stress may be reduced for people who rely on religious coping responses (Pargament 1997). Other studies indicate that individuals who forgive others for the things they have done tend to have better physical and mental health than people who are not willing to forgive (McCullough, Pargament and Thoresen 2000). Prayer appears to be an important factor, as well (Levin 2004). Although these studies have made a valuable contribution to the literature, the purpose of the current study is to strike off in a different direction by delving into an area that has received much less attention: close companion friendships that arise among fellow church members.

Research on church-based friendships is justified because scholars have been arguing for over 100 years that social relationships lie at the very heart of religion. Evidence of this may be found, for example, in the work of Edward Alsworth Ross (1896), who was an early president of the American Sociological Association. In fact, Ross went as far as to define religion solely in terms of its social roots. More specifically, he proposed that religion is “… the conviction of an ideal bond between members of society and the feelings that arise in consequence of that conviction” (Ross 1896:434, emphasis in the original) (see also Durkheim 1915/1965). But sociologists were by no means the only scholars to recognize the strong social underpinnings of religion. James Mark Baldwin, an early president of the American Psychological Association, maintained that ”… the fact is constantly recognized that religion is a social phenomena. No man is religious by himself, nor does he choose his god, nor devise his offering, nor enjoy his blessings alone” (Baldwin 1902:325).

Although the justification for studying the social foundation of religion was established over 100 years ago, it is hard to translate the insights of the classic social theorists into specific hypotheses that can be evaluated empirically. The problem arises because there are many different types of social relationships in church and it is not clear which one(s) may be associated with health (see Krause 2008, for a detailed discussion of church-based social relationships). So far, the majority of the studies on church-based social ties have been concerned with social support that is provided by fellow church members. Social support is assistance that is given specifically to help someone cope with the deleterious effects of stressful events, such as the death of a loved one or financial difficulty (Antonucci 1985; Krause 2008). This work is important because research reveals that the social support provided by coreligionists tends to reduce the noxious impact of stress on health (e.g., Krause 2006a).

Two fundamental aspects of social support underscore the boundaries associated with this type of social relationship measure. First, assistance during difficult times may be provided by people who are and are not close companion friends. Second, even though people at church help each other when stressful events arise, the social relationships they develop and maintain with coreligionists involve much more than crisis management. Instead, a good deal of the time they spend with others takes place outside the context of the stress process. Focusing on companion friendships helps illustrate this point (Cocking and Kennett 1998). Rook (1987:1133) defines companionship as social relationships that involve “… shared leisure activities that are undertaken primarily for the intrinsic goal of enjoyment.” Rook (1987) shows that companion friends are important because they provide the opportunity for self-disclosure, the discussion of personal aspirations, and the sharing of private jokes and stories.

Other investigators have focused on different dimensions of social relationships, including social networks (e.g., Ellison and George 1994). Social networks are typically assessed by determining how often study participants have been in contact with the people they know. However, it is not possible to tell from these measures whether contact has been established and maintained with close companion friends, or whether contact has been made with more casual acquaintances.

So far, there appears to be only one study in the literature that empirically evaluates close companion friendships in church (McFadden, Knepple and Armstrong 2003). The purpose of this study was to see if the degree of emotional closeness is greater among friends at church than among friendships that are maintained in the wider secular world. These investigators found little difference in emotional closeness between friendships in these two social settings. Although this is an important issue to examine, this study was not designed to see if having a close friend in church is associated with better health. Even so, further work in this area is justified because researchers have been studying the influence of close confidants in secular settings for some time. For example, Lowenthal and Clayton (1968) report that having a close confidant is associated with greater psychological well-being among community-dwelling older adults. More recently, Hays et al. (1998) found that having a confidant reduced the level of depressive symptomatology in late life.

The goal of the current study is to introduce a new measure of church-based companion friendships and empirically evaluate a conceptual model that proposes one way in which this type of social relationship may affect health. This is accomplished in the four sections that follow. The key characteristics of close companion friends are identified in the first section. In the process, an effort is made to show why companion friendships that arise in church may be especially efficacious. Then, in section two, a latent variable model is introduced which specifies one way in which close friends at church may influence physical health. Feelings of belonging in a congregation and gratitude toward God figure prominently in this conceptual scheme. Following this, the study sample and measures are presented in section three. Then the results are reviewed and discussed in the fourth section.

Identifying the Key Characteristics of Companion Friends

Unfortunately, a well-developed discussion of close companion friends at church is not available in the literature. Consequently, the discussion provided below begins by focusing on research that comes from a number of secular sources. Then, once the basic characteristics of secular companion friendships have been identified, an effort will be made to show the relatively unique ways in which these relationships may be manifest in church.

Close Companion Friends in Secular Settings

Regrettably, the literature on companion friends is in a state of disarray (Adams, Blieszner and De Vries 2000). The terms “friendship” and “companionship” are used interchangeably, there is no agreed-upon definition of either term, and researchers have not reached a consensus on the key characteristics and functions of companion friends. It is not possible to resolve these longstanding problems in the current study. Instead, we make an effort to contribute to the ongoing development of research in this field by focusing on several core characteristics of companion friends that cut across the discussions that have appeared so far (Ueno and Adams 2006), and that make the most sense from a theoretical point of view. Four interrelated functions of companion friends are examined briefly below.

First, companion friends are characterized by a high degree of self-disclosure (Patterson, Bettini and Nussbaum 1993). This means, for example, that close companion friends tell each other things they would not usually share with other people (Rook 1987). Clearly, this kind of intimate self-disclosure cannot take place without a high degree of openness, honesty, truthfulness, and trust.

Second, as Cocking and Kennett (1998) argue, companion friends share things they value highly. For example, they share plans, hopes, dreams, ambitions, and interests. And as the degree of intimacy grows between companion friends, they share private jokes and private stories that arise from their mutual experiences (Rook 1987). But there is some disagreement in the literature about the nature of the values and interests that are shared by companion friends. Some investigators, like Rook (1987), maintain that companion friends share common interests. However, other researchers, such as Cocking and Kennett (1998), convincingly argue that the interests that are shared by companion friends do not necessarily have to be the same. Instead, what matters most is that each person expresses interest in whatever their friend values highly.

Third, companion friends strive to emulate what they admire in their counterpart. As Sherman (1993) argues, “Each is inspired to develop himself more completely as he sees admirable qualities, not fully realized in himself, manifest in another whom he esteems … Character friends, as extended yet different selves, are eminently suited as models to be emulated” (Sherman 1993:105-106).

The final characteristic of companion friends is related to the previous one. It has to do with the way in which self-development and personal growth arises from emulating a close companion friend. Instead of merely being an aloof role model, companion friends actively encourage and spur the other on to greater self-awareness and greater personal growth. Simply put, companion friends invest time and effort to help each other attain that which they value in each other.

Church-Based Companion Friends

There is reason to believe that close companion friendships that arise in religious settings may be more efficacious than companion friendships that are found in the wider secular world. Recall that close companion friendships are characterized by a high degree of openness, honesty, and trust. Unfortunately, researchers have known for some time that people are often defensive; they deliberately withhold information from each other; and they may even manipulate situations in an effort to create a more favorable impression of themselves (Goffman 1959). However, the social milieu of the church is built upon fundamental religious teachings that may help overcome this type of defensive behavior. Support for this assertion may be found in the work of Thomas Oden (1969), a noted theologian. He maintains that, “It is only amid a community of genuine acceptance where unconditional forgiveness is mediated that persons find the freedom to put down their self-righteous defenses and freely enter into a responsible covenant with their neighbors” (Oden 1969:117).

Writing over a century ago, Ross illuminated how the shared values and experiences that arise in religious settings foster friendships that are especially close. He argued that, “To have the same gods, to be watched, loved, and protected by the same deities, to be destined to join the same unforseen company at death … these created fellowship” (Ross 1896:437). In fact, Ross (1896:441) went on to argue that friendships that form at church are so close that they become like family ties: “… conviction of our fundamental identity in nature and destiny … (is) … the modern counterpart of the old blood bond ….

Linking Close Companion Friendships with Health

Mendes de Leon (2005) recently observed that the mechanisms that link secular friendships with health are poorly understood. The same is certainly true with respect to church-based companion friends. The model that is provided in Figure 1 was developed to address this gap in the knowledge base. It should be emphasized that the relationships among the constructs depicted in Figure 1 were estimated after the effects of age, sex, education, and marital status were controlled statistically. Moreover, in order to make the model easier to read, the elements of the measurement model (i.e., the observed variables and measurement error terms) are not depicted in Figure 1. The core theoretical thrust in the model depicted in Figure 1 is captured in the following linkages: (1) people who go to church more often are more likely to have a companion friend in their congregation than individuals who do not go to church as often; (2) people who have a close companion friend in the place where they worship are more likely to feel they belong in their congregation; (3) individuals who feel they belong in their congregation are more likely to feel grateful to God; and (4) people who feel grateful to God are more likely to have better health. The theoretical rationale that supports these relationships is discussed below.

Figure 1
A Conceptual Model of Close Companion Friendships and Health

Church Attendance and Companion Friendships

As the discussion provided above reveals, companion friendships are characterized by a high degree of closeness, openness, and trust. However, this type of relationship does not develop quickly. Instead, companion friendships are cultivated slowly over the course of frequent social contact. But merely coming into contact with others may not be sufficient. Instead, a natural and conducive social environment is required. Engaging in shared activities, such as participating in worship services and the social activities that surround them (e.g., coffee hour, singing in the choir, serving as a greeter), “… allow participants to make discreet assessments of each other's suitability as potential friends, avoiding awkwardness and self-consciousness that can accompany an explicit focus on the extent of mutual attraction” (Rook 1991:105). Simply put, before close relationships can arise, people must have a high degree of contact with others in a natural setting that fosters the development of close ties in an unobtrusive manner. The frequency of church attendance serves as a useful marker of the type of setting in which the necessary social contact may take place.

Companion Friendships and Belonging

One important benefit of having a close companion friend at church arises from the fact that this type of relationship makes a person feel that he or she belongs in a congregation. Support for this notion may be found in a study by Winseman (2005). His research reveals that 84 percent of study participants who report having a close friend in church also feel they belong in their congregation. The construct of belonging has a long history in the social and behavioral sciences, but not enough has been said about it within the context of religion. Over half a century ago, Maslow (1954) identified belonging as one of the most basic human needs. Given the central role of belonging in life, it is not surprising to find that some investigators argue that one of the most important functions of religion is to help people find a sense of belonging (Baumeister 1991).

In order to see how close companion friends may promote a sense of belonging in a congregation, it is important to explore the essential nature and meaning of this key construct. As Carrier (1965) points out, belonging involves much more than a stated religious preference (e.g., “I am a Catholic”). Instead, it is an attitude, a psychological reality that encompasses a set of positive emotions and cognitions that arise from playing a meaningful role in a group. More specifically, Carrier argues that when a person feels that he or she belongs in a congregation, “The member sees himself taking part in his group; he identifies himself with it, he participates in it, he receives his motivation from it; in a word, he is in a state of disposition of interaction with the group which understands, inspires, and welcomes him” (Carrier 1965:58). Carrier's choice of words is important. By pointing out that people understand, inspire, and welcome each other, Carrier is, in essence, touching on some of the key characteristics of companion friends.

The insights provided by Carrier (1965) provide a useful point of departure for seeing how a close companion friend helps foster a sense of belonging in a congregation. But given the underdeveloped state of this literature, it is necessary to more clearly specify how companion friends perform this important function. Two factors figure prominently in this respect.

The first involves understanding the social milieu of the church. Before people can feel that they belong in a congregation, they must be able to accurately grasp this milieu and see their place in it. However, the social milieu of a congregation consists of a complex web of social relationships that requires a certain amount of skill to identify, evaluate, and understand. Having a close companion friend at church provides a safe haven in which the complex social milieu of a congregation can be candidly assessed without undue fear of censure. When they are viewed in this way, church-based companion friends act as a portal or gateway to the wider congregational milieu. Greater understanding of this important social integration function may be found by turning to the insights provided by Berger and Pullberg (1965). These investigators maintain that “… the human enterprise of producing a world is not comprehensible as an individual project. Rather, it is a social process: men together engage in constructing a world, which then becomes their common dwelling” (Berger and Pullberg 1965:201, emphasis in the original). We extend the observations of Berger and Pullberg by arguing that not just anyone helps construct and define social reality; instead, this challenging task is best accomplished within the context of a close companion friendship.

Second, some of the activities that take place in church can be highly emotional in nature. For example, funerals may evoke deep emotions that are difficult to share with church members who are known only casually. Being able to work through these emotions with a close and trusted other helps an individual appreciate the true value in the ceremonies that create them, thereby binding the individual closer to the faith and the institution that promotes them. But this process of bonding with the congregation may be found in more joyous occasions, as well. As Wuthnow (1999) points out, events such as christenings and weddings take on added significance when they are shared with a close other.

Belonging and Feelings of Gratitude Toward God

A recent study by Krause and Wulff (2005) provides some evidence that people who feel they belong in their congregation are more satisfied with their health than individuals who do not feel as tightly integrated into the place where they worship. However, these investigators did not empirically evaluate the intervening constructs that link a sense of belonging with health. As shown in Figure 1, turning to the notion of gratitude toward God may provide valuable insight into this issue.

The participants in the current study are all older adults. As Rosow (1976) pointed out some time ago, aging is a “role-less role.” This means that as people enter late life they tend to lose a number of important roles. For example, they retire, many become widowed, and their children reach adulthood and move out of the home. At the same time, society provides few new roles to fill the void that is created by these role exits. The essence of this problem was captured by Baltes and Smith (1999:158), who note that “… relatively speaking, old age is young; therefore neither biological nor cultural evolution has had sufficient opportunity to evolve a full and optimizing scaffolding … for the later phases of life.” Being members of a congregation and feeling as though they belong in it helps fill the social vacuum that people may encounter in late life. When older adults find a meaningful place in the church, a place where they feel they belong, they are likely to feel grateful. And because this sense of belonging arises within the church, older people are especially likely to feel grateful to God. Gratitude is an important construct to focus on because, as Emmons and Crumpler (2000) point out, it is cultivated by every major religion in the world. Following the definition proposed by Peterson and Seligman (2004), gratitude is defined as feelings of thankfulness toward a specific person or entity for the benefits that individual or entity has provided. They defined gratitude broadly so that it can be studied in a wide range of settings. But when it is cast within the context of religion, researchers typically think in terms of feelings of gratitude toward God (Krause, 2006b).

Gratitude toward God and Health

There are at least two ways in which gratitude may exert a beneficial effect on health. The first is more direct and may be found in the literature on positive psychology (Seligman and Csikszentmihalyi 2000). Gratitude is a positive emotion, and like many other positive emotions, it may have a direct and beneficial effect on the physiology of the body. For example, research reveals that interventions involving positive emotion refocusing techniques are associated with a significant reduction in blood pressure, improved functional capacity, reduced depression in patients with congestive heart failure, and a reduction in anxiety among individuals with HIV (McCraty and Childre 2004). But more important, research reviewed by McCraty and Childre (2004) reveals that developing a deeper sense of appreciation (an emotion that is closely akin to gratitude) has a significant influence on secretory IgA, which is a predominant antibody class in the body's defense against pathogens. Second, research reveals that a sense of gratitude may promote compassionate and altruistic behavior (McCullough and Tsang 2004). This is important because a small cluster of studies indicate that helping others tends to bolster a support provider's own health and may even help them to live longer (Krause 2006c).



The data for this study come from an ongoing nationwide survey of older Whites and older African Americans. The study population was defined as all household residents who were either Black or White, noninstitutionalized, English-speaking, and at least 66 years of age. Geographically, the study population was restricted to all eligible persons residing in the coterminous United States (i.e., residents of Alaska and Hawaii were excluded). Finally, the study population was restricted to currently practicing Christians, individuals who were Christian in the past but no longer practice any religion, and people who were not affiliated with any faith at any point in their lifetime. This study was designed to explore a range of issues involving religion. As a result, individuals who practice a faith other than Christianity were excluded because members of the research team felt it would be too difficult to devise a comprehensive battery of religion measures that would be suitable for individuals of all faiths.

The sampling frame consisted of all eligible persons contained in the beneficiary list maintained by the Centers for Medicare and Medicaid Services (CMS). A five-step process was used to draw the sample from the CMS Files. A detailed discussion of these steps is provided by Krause (2002).

The baseline survey took place in 2001. The data collection for all waves of interviews in this study was performed by Harris Interactive (New York). A total of 1,500 interviews were completed, face-to-face, in the homes of the study participants. Older African Americans were over-sampled so that sufficient statistical power would be available to assess race differences in religion. As a result, the Wave 1 sample consisted of 748 older Whites and 752 older Blacks. The overall response rate for the baseline survey was 62%.

The Wave 2 survey was conducted in 2004. A total of 1,024 of the original 1,500 study participants were re-interviewed successfully, 75 refused to participate, 112 could not be located, 70 were too ill to participate, 11 had moved to a nursing home, and 208 were deceased. Not counting those who had died or were placed in a nursing home, the re-interview rate for the Wave 2 survey was 80%.

A third wave of interviews was completed in 2007. A total of 969 older study participants were re-interviewed successfully, 33 refused to participate, 118 could not be located, 17 were too sick to take part in the interview, and an additional 155 older study participants had died. Not counting those who died, the re-interview rate was 75%.

The data that are used in the analyses for the current study come from the Wave 3 interviews because this was the first time information was gathered on church-based companion friendships as well as a sense of belonging in a congregation. After using listwise deletion to deal with item nonresponse, complete data were available for between 537 and 349 older study participants. The reduction in cases from 969 to 537 respondents was primarily due to the way information on church-based companion friendships was obtained. When this study was designed, the members of the research team felt that it did not make sense to administer questions on church-based companion friends to people who either never go to church or go to church no more than twice a year. Based on this reasoning, 333 older people were excluded from the analyses presented below. The reason for also analyzing a sample of 349 study participants will become evident after the measures of church-based companion friendships are discussed in the next section.

Preliminary analysis of the sample comprising 537 study participants revealed that their average age was 78.5 years (SD = 5.5 years) at Wave 3. Approximately 38% of these study participants were men, 45% were married at the time of the Wave 3 survey, and 49% were White. Finally, these study participants reported they had successfully completed an average of 12.2 years of schooling (SD = 3.3 years). These descriptive data, as well as the data used in the analyses presented below, have been weighted.


Table 1 contains the measures that were used in this study. The procedures used to code these items are provided in the footnotes of this table.

Table 1
Core Study Measures

Frequency of Church Attendance

The participants in this study were asked how often they attended religious services in the year prior to the Wave 2 survey. A high score represents more frequent church attendance. The mean level of church attendance is 7.4 (SD = 1.3).

Church-Based Companion Friendship

A two-part strategy was used to assess church-based companion friendships. First, as shown in Table 1, study participants were asked whether they had a very good friend or valued companion in their congregation. Approximately 64% of the older people in this study indicated they had a close friend in their church. Second, study participants who indicated they had a close friend in church were asked a battery of questions designed to assess the function and degree of involvement in this relationship. These questions were only asked of people who had a companion friend at church. This is why a smaller sample of 349 study participants is also used in the analyses presented below. Taken together, this two-part measurement strategy corresponds to measures that assess the structure (i.e., presence) and function of church-based companion friendships.

The measures that assess the nature and function of companion friendships were evaluated with a confirmatory analysis factor model (not shown here). A three-factor solution fit the data well. The three factors assess how often an older person shares things like plans and dreams with a companion friend, how often they feel they can be themselves around this close other, and how often a close companion friend contributes to their personal development. However, the findings from this factor model also revealed that the correlations among the three factors were quite high (e.g., .650). Viewed from a more substantive perspective, the high factor correlations suggest that companion friends who perform one function are also likely to perform the other functions, as well. Consequently, three composite measures were created by adding the items that assess each factor together. The composite measures were then treated as three observed indicators of the latent construct assessing the function of companion friendships (see Liang, Lawrence, Bennett, and Whitelaw, 1990, for a discussion of the use of composite measures in latent variable modeling). A high score denotes more effective companion friendship functioning. The mean of this composite measure is 26.4 (SD = 5.2).

Belonging in a Congregation

As shown in Table 1, belonging is measured with three indicators that assess whether older people feel that they belong in their congregation, whether being a member of their congregation is important to them, and whether they feel that they play a meaningful role in the life of their congregation. A high score on this short index reflects a stronger sense of belonging. The mean of this composite measure is 10.3 (SD = 1.7).

Gratitude toward God

Gratitude toward God is evaluated with three observed indicators. A high score on this construct means that study participants feel more grateful to God. The mean for this short scale is 11.4 (SD = 1.2).

Self-Rated Health

Self-rated health is assessed with three items that ask study participants to rate their overall health at the present time, compare their health to other people of their own age, and report how satisfied they are with their health. A high score stands for a more favorable health rating. The mean for this measure of health is 7.6 (SD = 1.7).

Demographic Control Measures

As reported earlier, the relationships among the constructs depicted in Figure 1 were estimated after the effects of age, sex, marital status, and education were controlled statistically. Age is scored in years and education reflects the total number of years of schooling completed successfully by study participants. In contrast, sex (1 = men; 0 = women), race (1 = White; 0 = Black), and marital status (1 = presently married; 0 = otherwise) were scored in a binary format.


The findings from this study are presented below in two sections. The first section contains the estimates that were derived from estimating the model depicted in Figure 1. The central focus of this model involves the impact of having a close companion friend on belonging, gratitude toward God, and health. The findings in the second section were obtained after the measure of companion friendship functions was used in place of the single indicator that assesses the presence of a close friend in a congregation.

Presence of Companion Friends and Health

The model depicted in Figure 1 was estimated with a covariance matrix. Before turning to the substantive findings that were derived from estimating this model, it is important to first evaluate the fit of the model to the data as well as the psychometric properties of the observed indicators.

Fit of the model to the data

The findings (not shown in Table 2) reveal that the fit of the model to the data is good. More specifically, the Bentler-Bonett Normed Fit Index estimate of .964 is well above the minimum cut point of .900 (Bentler and Bonett 1980). Similarly, the standardized root mean square residual estimate (RMSEA) of .040 is below the recommended ceiling of .050 (Kelloway 1998). Finally, Bollen's (1989) Incremental Fit Index value of .982 is quite close to the ideal target value of 1.0 for this measure.1

Table 2
Measurement Model Parameter Estimates (N = 537)

Psychometric properties of the observed indicators

Table 2 contains the standardized factor loadings and standardized measurement error terms that were derived from estimating the model depicted in Figure 1. These coefficients are important because they provide preliminary information on the psychometric properties of the multiple-indicator study measures. Although there are no firm guidelines in the literature, Klein (2005) suggests that standardized factor loadings in excess of .600 have reasonable reliability. As shown in Table 2, the standardized factor loadings range from .632 to .888, suggesting that the measures used in this study have good psychometric properties.

Although the factor loadings and measurement error terms provide useful information about the reliability of each observed indicator, it would be helpful to know something about the reliability of the multiple item scales as a whole. Fortunately, these estimates can be computed with a formula provided by DeShon (1998). This formula uses the factor loadings and the measurement error terms provided in Table 2. These additional computations produced the following reliability estimates for the composite measures in Figure 1: belonging in a congregation (reliability = .872), feelings of gratitude toward God (reliability = .887), and self-rated health (reliability = .774).

Substantive findings

Table 3 contains estimates of the relationships among the latent constructs contained in Figure 1. It was somewhat surprising to find that older adults who attend church more often are not more likely to report having a close friend in the congregation than older people who do not go to church as frequently (Beta = .071; ns). This suggests that some factor other than mere social contact must be at work. Further insight into this issue may be found by turning to additional results that are provided in Table 3. These data reveal that older men are less likely than older women to report having a close companion friend in the place where they worship (Beta = -.103; p < .05). Moreover, the data further reveal that older Whites are less likely than older African Americans to report having a close friend in their congregation (Beta = -.126; p < .01). Because both older women and older Blacks have been shown to be more deeply involved in a wide range of church-based social relationships (Krause 2008), these data indicate that core social structural factors play a larger role than the sheer frequency of social contact in the development of church-based close companion friendships.

Table 3
The Presence of Companion Friends and Health (N = 537)

Returning to Table 3, the results provide support for all the remaining hypotheses that were developed for this study. More specifically, the data suggest that older people who report having a close companion friend in church are more likely to feel they belong in their congregation (Beta = .173; p < .001). The findings further reveal that older people who feel they belong in their congregations are especially likely to feel grateful to God (Beta = .475; p < .001). The size of this effect is quite large by social and behavioral science standards. Finally the data in Table 3 further indicate that older people who feel more grateful to God are more likely to rate their health in a favorable manner than older adults who feel less grateful to God (Beta = .269; p < .001).

The Functions of Companion Friendships in Church and Health

Table 4 provides the results of the analyses that were performed to see if focusing on the functions of companion friendships sheds additional light on how this type of church-based social relationship is associated with health. This model was also estimated with a covariance matrix. The fit of the model that was designed to evaluate this issue is adequate. More specifically, the Bentler-Bonett Normed Fit Index value of .931 (Bentler and Bonett 1980), the standardized root mean square residual estimate of .040, and Bollen's (1989) Incremental Fit Index value of .973 are all close to their respective recommended cut point scores.2 The standardized factor loadings (not shown here) ranged from .544 to .934. Although a standardized loading of .544 is below the cut point recommended by Klein (2005), the difference between this estimate and the recommended value is trivial. Therefore, when the standardized factor loadings are taken as a whole, the data suggest that the observed indicators that are used in this supplemental model have adequate psychometric properties. Additional computation (not shown here) suggests that the reliability estimate for the new measure of companion friendship functions is acceptable (.731).

Table 4
Companion Friendship Functions and Health (N = 318)

The data that are contained in Table 4 provide evidence of why it is important to study companion friendships that arise in the church. Specifically, the results reveal that older people who go to church more often are more likely to feel they can share private feelings and beliefs with their companion friend; they are more likely to feel they can be themselves around their close friend; and they are more likely to feel their companion friend contributes to their personal development (Beta = .137; p < .05). When coupled with the finding in the previous section, these results show that even though the frequency of church attendance does not determine if an older person develops a close companion friend, it does affect the ability of older people to extract the maximum benefits from this important type of church-based social relationship.

The data in Table 4 further suggest that older people who have developed a highly functional relationship with a close companion friend are more likely to feel as though they belong in their congregation (Beta = .181; p < .01). And consistent with the results that were provided in the previous section, the findings indicate that older people who feel they belong in their congregation also feel more grateful to God (Beta = .559; p < .001). Finally, the data reveal that older people who feel more grateful to God tend to rate their health more favorably (Beta = .197; p < .05).

Study Limitations

Although the findings from this study may have contributed to the literature, researchers would be well advised to pay careful attention to the limitations in the work that has been done. Two shortcomings are discussed briefly below.

First, the data that were analyzed in this study are cross-sectional. As a result, the temporal ordering among the constructs in our conceptual model was based on theoretical considerations alone. Consequently, it is possible to reverse the causal ordering among some of the linkages we proposed. For example, we proposed that people who feel grateful to God subsequently tend to enjoy better health. However, it is possible to argue that people who have better health are more likely to feel grateful to God. The physiological evidence provided by McCraty and Childre (2004) suggests that our specification is valid, but this as well as other causal assumptions that are embedded in study model can only be fully evaluated in studies that utilize true experimental designs.

When the theoretical rationale for this study was developed, we argued that companion friendships in church represent a different type of social relationship than church-based social support. We felt this distinction was justified because social support involves assistance that is provided for the specific purpose of alleviating the deleterious effects of stress, whereas companion friendships perform other functions. However, this may not be a valid distinction because the person who is a companion friend may also be the individual whom a person turns to for social support during stressful times. To the extent that this is true, social support and companionship may instead reflect different functions of the same broader social relationship. The only way to get to the bottom of this issue is to conduct a detailed church-based social network analysis in which the names of the individuals who provide social support are matched against the names of the people who are close companion friends.


Ralph Waldo Emerson (1841/1983:348), an ordained minister, argued that “… a friend may well be reckoned the masterpiece of nature.” When viewed at the broadest level, the goal of the current study was to provide one reason why this bold assertion may be true. The data reveal that having a close companion friend at church helps integrate older people more tightly into the place where they worship. Put another way, the findings suggest that older adults who have a companion friend at church are more likely to feel they belong in their congregation. The results further suggest that older individuals who feel they belong in their congregation also tend to feel more grateful to God. This finding is important because the data reveal that older adults who feel closer to God are also more likely to rate their health in a favorable way. Simply put, having a close friend at church may be a masterpiece of nature because this type of social relationship may be good for the health of older people. To the best of our knowledge, this is the first time that the relationship between close companion friendships at church and health has been empirically evaluated.

But research on church-based companion friends is in its infancy. One goal of the current study was to promote more research in this area by introducing and empirically evaluating new measures that were designed to assess this potentially important type of church-based social relationship. Evidence was provided to show that these new measures have adequate reliability. And the fact that these new measures are associated with constructs such as belonging in a congregation and gratitude toward God provides some evidence of their construct validity (Nunnally and Bernstein 1994).

Now that acceptable measures are in place, it is time to explore other ways in which close companion friends at church may influence health and well-being. There are a number of intriguing possibilities. For example, a growing number of studies suggest that people who are more involved in religion tend to have better health behaviors than individuals who are less involved in religion (e.g., Hill, Burdette, Ellison and Musick 2006). Included among the beneficial health behaviors that are associated with religious involvement are the avoidance of tobacco, the moderate use of alcohol, engaging in strenuous exercise, and the utilization of preventive health services (e.g., physical examinations). But it is not entirely clear how involvement in religion promotes these beneficial effects. One potentially important explanatory factor may be found in a recent study by Ellison and his colleagues (Forthcoming). These investigators report that people are more likely to adopt good health behaviors if they are encouraged to do so by the people in their congregation. But the people who are best suited for this task were not identified clearly in this study. The decision to engage in certain health behaviors, such as the consumption of alcohol, is a potentially sensitive issue. And as research by Tucker and Mueller (2000) reveals, people are likely to resent efforts to influence their health behaviors if they believe that others are trying to control or manipulate them. Perhaps the sense of closeness, openness, and trust that is found among companion friends at church provides a context in which health behaviors can be influenced without invoking the negative reactions identified by Tucker and Mueller.

Research on close companion friends in the church may be extended in other ways, as well. Two areas for further research may be found in a study by Wenger and Jerrome (1999) that was conducted in a secular setting. First, these investigators report that older people may change close confidants over time. We need to know if this also happens in the church, and if it does, the health-related consequences of doing so should be assessed. More specifically, research is needed to see if the process of switching close companion friends disrupts the benefits that potentially arise from having this type of social relationship. Second, Wenger and Jerrome (1999) report that as people reach advanced old age they tend to loose confidants due to factors such as death. Perhaps the loss of an especially close social relationship has particularly pronounced effects in late life. Research by Palinkas, Wingard, and Barrett-Connor (1990) highlights yet another way that research on companion friends may be enhanced. Their study shows that older women typically have more well-developed social relationships than older men. Although this study was conducted in the community, it raises the possibility that there may also be potentially important gender differences in the presence and functioning of close companion friends in the church.

A sense of belonging in a congregation emerged in the current study as a key construct linking close companion friends with health. Greater attention should be given to the intervening role that is played by belonging. Viewed in a general way, the data suggest that companion friendships influence health because they draw a person into a closer and more intimate relationship with the individuals in their congregation. However, a study by Lin, Ye, and Ensel (1999) that was conducted in the general population suggests there may be not one, but three increasingly intimate layers of social connectedness: belonging, bonding, and binding. Research is needed to evaluate the relationship between close companion friendship and each of these levels of intimacy. Doing so should provide valuable insight into the way in which companion friendships shape the nature of the social relationships that are formed in congregations.

Down through the ages numerous scholars, philosophers, and sages have extolled the virtues of friendship. Yet, it is surprising to find that contemporary researchers have yet to fully evaluate the functions and benefits of close friendships that arise in the church. This oversight is perplexing because so much has been made of the social basis of religion (e.g., Simmel 1898/1997). We hope the findings from our work encourage other investigators to take a closer look at the friendships that people form in the places where they worship. If centuries of reflection are valid, then exploring close friendships that arise in church should provide much valuable insight.


This research was supported by a grant from the National Institute on Aging (RO1 AG014749) and a grant to Duke University from the John Templeton Foundation.


1Researchers often include chi-square goodness-of-fit information when reporting the findings from the analysis of latent variable models. The chi-square value for this model is 127.618 with 66 degrees of freedom (p < .001). However, chi-square estimates should be viewed with caution because they are quite sensitive to sample size. Further evidence that the fit of this model to the data is satisfactory is provided by the comparative fit index (CFI – see Bollen 1989) value of .982, which is close to the ideal recommended value of 1.0. Similarly, the standardized root mean square residual (SRMR) value (.026) is below the recommended ceiling of .050 (see Kelloway 1998).

2The chi-square goodness of fit estimate for this model is 154.276 with 96 degrees of freedom (p < .001). The CFI (Bollen 1989) estimate (.972) and the SRMR (Kelloway 1998) estimate (.041) both suggest that the fit of the model to the data is good.


  • Adams Rebecca G, Blieszner Rosemary, De Vries Brian. Definition of Friendship in the Third Age: Age, Gender, and Study Location Effects. Journal of Aging Studies. 2000;14:117–133.
  • Antonucci Toni c. Social Relationships: An Examination of Social Networks, Social Support, and Sense of Control. In: Birren James E, Schaie K Warner., editors. Handbook of the Psychology of Aging. Sand Diego, CA: Academic Press; 1985. pp. 427–451.
  • Baldwin James M. Fragments in Philosophy and Science of Being: Collected Essays and Addresses. New York: Charles Scribner's Sons; 1902.
  • Baltes Paul B, Jacqui Smith. Multilevel and Systemic Analysis of Old Age: Theoretical and Empirical Evidence for a Fourth Age. In: Bengtson Vern L, Schaie K Warner., editors. Handbook of Theories of Aging. New York: Springer; 1999. pp. 153–173.
  • Baumeister Roy F. Meanings of Life. New York: Guilford; 1991.
  • Bentler Peter M, Bonett Douglas G. Significance Tests and Goodness-of-Fit in the Analysis of Covariance Structures. Psychological Bulletin. 1980;88:588–600.
  • Berger Peter L, Pullberg Stanley. Reification and sociological critique of consciousness. History and Theory. 1965;4:196–211.
  • Bollen Kenneth A. Structural Equations with Latent Variables. New York: Wiley; 1989.
  • Carrier Herve. The Sociology of Religious Belonging. New York: Herder and Herder; 1965.
  • Cocking Dean, Kennett Jeanette. Friendship and the self. Ethics. 1998;108:502–527.
  • DeShon Richard P. A Cautionary Note on Measurement Error Correlations in Structural Equation Models. Psychological Methods. 1998;3:412–423.
  • Durkheim Emile. Elementary Forms of Religious Life. New York: Free Press; 1915/1965.
  • Ellison Christopher G, George Linda K. Religious Involvement, Social Ties, and Social Support in a Southeastern Community. Journal for the Scientific Study of Religion. 1994;33:26–61.
  • Ellison Christopher G, Lee Jinwoo, Benjamins Maureen R, Krause Neal, Ryan Nichole, Marcum Jack. Forthcoming. Congregational Support Networks, Health Beliefs, and Annual Medical Exams: Findings From a Nationwide Sample of Presbyterians. Review of Religious Research
  • Emerson Ralph W. In: Ralph Waldo Emerson: Essays & Lectures. Porte Joel., editor. New York: Library Classics of the United States; 1841/1983.
  • Emmons Robert A, Crumpler Cheryl A. Gratitude as a Human Strength: Appraising the Evidence. Journal of Social and Clinical Psychology. 2000;19:56–69.
  • Fetzer Institute/National Institute on Aging Working Group. Multidimensional Measurement of Religiousness/Spirituality for Use in Health Research. Kalamazoo, MI: John E. Fetzer Institute; 1999.
  • Erving Goffman. The Presentation of Self in EverydayLlife. New York: Anchor Books; 1959.
  • Hays Judith C, Landerman Lawrence R, George Linda K, Flint Elizabeth P, Koenig Harold G, Land Kenneth C, Blazer Dan G. Social Correlates of the Dimension of Depression in the Elderly. Journal of Gerontology: Psychological Sciences. 1998;53B:P31–P39. [PubMed]
  • Hill Terrance D, Burdette Amy M, Ellison Christopher G, Musick Marc A. Religious Attendance and Health Behaviors Among Texas Adults. Preventive Medicine. 2006;42:309–312. [PubMed]
  • Kelloway E Kevin. Using LISREL for Structural Equation Modeling. Thousand Oaks, CA: Sage; 1998.
  • Klein Rex B. Principles and Practice of Structural Equation Modeling. New York: Guilford; 2005.
  • Koenig Harold G, McCullough Michael E, Larson David B. Handbook of Religion and Health. New York: Oxford University Press; 2001.
  • Krause Neal. Church-Based Social Support and Health in Old Age: Exploring Variations by Race. Journal of Gerontology: Social Sciences. 2002;57B:S332–S347. [PubMed]
  • Krause Neal. Exploring the Stress-Buffering Effects of Church-Based Social Support and Secular Social Support on Health in Late Life. Journal of Gerontology: Social Sciences. 2006a;61B:S34–S43. [PubMed]
  • Krause Neal. Gratitude Toward God, Stress, and Health in Late Life. Research on Aging. 2006b;28:163–183.
  • Krause Neal. Church-Based Support and Mortality. Journal of Gerontology: Social Sciences. 2006c;61B:S140–S146. [PubMed]
  • Krause Neal. Aging in Church: How Social Relationships Affect Health. Conshohocken, PA: Templeton Foundation Press; 2008.
  • Krause Neal, Wulff Keith M. Church-Based Social Ties, a Sense of Belonging in a Congregation, and Physical Health Status. International Journal for the Psychology of Religion. 2005;15:73–93.
  • Levin Jeffrey S. Prayer, Love, and Transcendence: An Epidemiologic Perspective. In: Schaie K Warner, Krause Neal, Booth Alan., editors. Religious Influences on Health and Well-Being. New York: Springer; 2004. pp. 69–95.
  • Jersey Liang, Lawrence Rene H, Bennett J, Whitelaw Nancy A. Appropriateness of Composites in Structural Equation Modeling. Journal of Gerontology: Social Sciences. 1990;45:S52–S59. [PubMed]
  • Nan Lin, Ye Xiaolan, Ensel Walter M. Social Support and Depressed Mood: A Structural Analysis. Journal of Health and Social Behavior. 1999;40:344–359. [PubMed]
  • Lowenthal Marjorie F, Haven Clayton. Interaction and Adaptation: Intimacy as a Critical Variable. American Sociological Review. 1968;33:20–30. [PubMed]
  • Maslow Abraham H. Motivation and Personality. New York: Harper; 1954.
  • McCraty Rollin, Childre Doc. The Grateful Heart: The Psychophysiology of Appreciation. In: Emmons Robert A, McCullough Michael E., editors. The Psychology of Gratitude. New York: Oxford University Press; 2004. pp. 230–255.
  • McCullough ME, Pargament KI, Thoresen CE. Forgiveness: Theory, research, and practice. New York: Guilford; 2000.
  • McCullough Michael E, Tsang Jo-Ann A. Parent of Virtues? The Prosocial Contours of Gratitude. In: Emmons Robert A, McCullough Michael E., editors. The Psychology of Gratitude. New York: Oxford University Press; 2004. pp. 123–141.
  • McFadden Susan H, Knepple Amy M, Armstrong Jessica A. Length and Locus of Friendship Influence, Church Members' Sense of Social Support, and Comfort Sharing Emotions. Journal of Religious Gerontology. 2003;15:39–55.
  • Mendes de Leon Carlos F. Why do Friendships Matter for Survival? Journal of Epidemiology and Community Health. 2005;59:538–539. [PMC free article] [PubMed]
  • Nunnally Jum C, Bernstein Ira H. Psychometric Theory. 3rd. New York: McGraw-Hill; 1994.
  • Thomas Oden. The Structure of Awareness. Nashville, TN: Abingdon; 1969.
  • Palinkaas Lawrence A, Barrett-Connor Elizabeth. The Biocultural Context of Social Networks and Depression Among the Elderly. Social Science and Medicine. 1990;30:441–447. [PubMed]
  • Pargament Kenneth I. The Psychology of Religious Coping: Theory, Research, and Practice. New York: Guilford; 1997.
  • Patterson Brian R, Lorraine Bettini, Nussbaum Jon F. The Meaning of Friendship Across the Life Span: Two Studies. Communication Quarterly. 1993;41:145–160.
  • Peterson Christopher, Seligman Martin E. Character Strengths and Virtues: A Handbook and Classification. New York: Oxford University Press; 2004.
  • Rook Karen S. Social Support Versus Companionship: Effects of Life Stress, Loneliness, and Evaluations of Others. Journal of Personality and Social Psychology. 1987;52:1132–1147. [PubMed]
  • Rook Karen S. Facilitating Friendship Formation in Late Life: Puzzles and Challenges. American Journal of Community Psychology. 1991;19:103–110. [PubMed]
  • Rosow Irving. Status and Role Change Through the Life Span. In: Binstock Robert H, Shanas Ethel., editors. Handbook of aging and the social sciences. New York: Van Nostrand Reinhold; 1976. pp. 457–482.
  • Ross Edward A. Social control V. American Sociological Review. 1896;2:433–445.
  • Seligman Martin E, Csikszentmihalyi Mihaly. Positive Psychology: An Introduction. American Psychologist. 2000;55:5–14. [PubMed]
  • Sherman Nancy. Aristotle and the Shared Life. In: Badhwar Neera., editor. Friendship: A Philosophical Reader. Ithaca, NY: Cornell University Press; 1993. pp. 91–107.
  • Simmel Georg. A Contribution to the Sociology of religion. In: Helle Horst J., editor. Essays on Religion B Georg Simmel. New Haven, CT: Yale University Press; 1898/1997. pp. 101–120.
  • Tucker Joan S, Mueller Jennifer S. Spouse's Social Control of Health Behaviors: Use and Effectiveness of Specific Strategies. Personality and Social Psychology Bulletin. 2000;26:1120–1130.
  • Koji Ueno, Adams Rebecca G. Adult Friendship: A Decade Review. In: Noller Particia, Feeney Judith A., editors. Close relationships. New York: Psychology Press; 2006. pp. 151–169.
  • Clare Wenger G, Dorothy Jerrome. Change and Stability in confidant relationships: Findings from the Bangor Longitudinal Study of Ageing. Journal of Aging Studies. 1999;13:269–294.
  • Winseman Albert L. Friendship Feeds the Flock: People with Best Friends at Church More Spiritually Committed. 2005. Paper posted on the following website: http://poll.galup.com.
  • Wuthnow Robert. Growing Up Religious: Christians and Jews on Their Faith Journeys. Boston: Beacon Press; 1999.
PubReader format: click here to try


Save items

Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


  • PubMed
    PubMed citations for these articles

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...