Article

Trust in God, Forgiveness by God, and Death Anxiety

OMEGA—Journal of Death and Dying 2015, Vol. 72(1) 20–41 ! The Author(s) 2015 Reprints and permissions: sagepub.com/journalsPermissions.nav DOI: 10.1177/0030222815574697 ome.sagepub.com

Neal Krause1

Abstract Research on religion and death anxiety has produced many contradictory findings. These conflicting findings arise, in part, from inadequacies in the measurement of religion as well as problems with the way the data have been analyzed. The purpose of the current study is to develop and empirically evaluate a conceptual model that contains the following core hypotheses: (a) People who go to church more often will receive more spiritual support from fellow church members (spiritual support is assistance provided by coreligionists for the explicit purpose of increasing the religious beliefs and practices of the recipient). (b) Individuals who receive more spiritual support will be more likely to trust God. (c) Those who trust God more deeply will be more likely to feel forgiven by Him. (d) People who feel forgiven by God will experience less death anxiety. Findings from a recent nationwide survey provide support for each hypothesis. Keywords religion, forgiveness, death anxiety

Research on the relationship between religion and death anxiety is over 50 years old (Feifel, 1959). Since then, a growing number of studies have been conducted to see if religion helps people cope with knowledge of their own inevitable demise. Ellis and Wahab (2013) recently reported that as of 2012, 84 studies

1

Department of Health Behavior and Health Education, School of Public Health, University of Michigan, Ann Arbor, MI, USA Corresponding Author: Neal Krause, Department of Health Behavior and Health Education, School of Public Health, University of Michigan, 1415 Washington Heights, Ann Arbor, MI 48109-2029, USA. Email: [email protected]

Krause

21

that assessed the relationship between some aspect of religion and death anxiety have appeared in the literature. Unfortunately, these investigators pointed out that the findings from this research are contradictory. More specifically, four different conclusions emerge from this body of work: (a) Some studies showed that involvement in religion lowers feelings of death anxiety. (b) Other studies indicated that various facets of religion are associated with greater death anxiety. (c) Yet other studies failed to find a statistically significant relationship between the two. (d) Some researchers reported a curvilinear relationship between religion and death anxiety whereby people with moderate levels of religious involvement are at greater risk. Although several theoretical perspectives have been developed to explain why religion may help reduce feelings of death anxiety, most researchers subscribe to Terror Management Theory (TMT; Greenberg, Pyszczynski, & Solomon, 1986). According to this view, people are deeply troubled by their own inevitable death, and as a result, they subscribe to culturally developed world views that have been devised to allay these concerns. Perhaps the most important world views are provided by religion (Soenke, Landau, & Greenberg, 2012). The basic tenets of many faith traditions help allay concerns about dying by providing assurance of life after death and many religions, such as the Christian faith, portray the afterlife in glowing terms. So if religion holds out the promise of a good life after death, it is not clear why some studies show that religion exacerbates a sense of death anxiety or why other studies show that religion fails to lower feelings of death anxiety. There are a number of reasons for these unanticipated findings, as Ellis and Wahab (2013) pointed out. Even so, the purpose of the current study is to address three issues that deal with the measurement and conceptualization of religion. The first problem with the literature on religion and death anxiety has to do with the fact that religion has been assessed in a number of studies with only one or two measures (Ellis & Wahab, 2013). For example, Lin (2003) focused solely on the relationship between a single item that assesses the degree of self-rated religiousness and fear of death. Unfortunately, researchers have known for some time that religion is a complex multidimensional phenomenon that can be measured in a number of ways (Fetzer Institute/National Institute on Aging Working Group, 1999). As a result, a broader array of religion measures are needed to more adequately depict the way in which involvement in religion may influence feelings of death anxiety. The second problem with the literature arises from the fact that, in many cases, the dimensions of religion that have been examined are more general or global in nature and do not speak directly to issues involving death. For example, some investigators focused solely on the frequency of church attendance (e.g., Morris & McAdie, 2009). However, attendance at worship services is a complex phenomenon in its own right that subsumes many different aspects of religion. For example, when people attend church services, they pray, they listen

22

OMEGA—Journal of Death and Dying 72(1)

to passages that appear in sacred texts, and they interact with their fellow church members. As a result, it is not clear which aspect(s) of church attendance are related to feelings of death anxiety. Other researchers relied on scales that assess intrinsic and extrinsic religiousness, which measure general feelings of religious commitment (e.g., Ardelt & Koenig, 2006). But once again, it is not clear how deeper commitment to one’s faith should assuage feelings of death anxiety. Moreover, faith traditions embrace a wide range of tenets and beliefs and people may feel more committed to some religious teachings than others. And the teachings they endorse more fully may not be related to death per se. So if deeper commitment (i.e., an intrinsic religious motivation) is associated with less death anxiety, the specific beliefs that individuals are committed to are not identified, and as a result, the theoretical process remains unclear. In contrast, other investigators have included measures in their work that assess beliefs about the nature of the afterlife, which obviously deal more directly with issues involving death (e.g., Lundh & Radon, 1998). However, people may believe in an afterlife without necessarily being involved in a formal faith tradition, and as a result, measures that assess belief in an afterlife may capture the influence of factors other than religion per se. The third problem with the literature on religion and death anxiety arises from the fact that the multiple dimensions that comprise religion are interrelated. Instead of comparing and contrasting the effects of different dimensions of religion on death anxiety in an effort to see which is most important, it makes more sense to build conceptual models that show how one dimension of religion leads to another dimension of religion and how both work in concert to influence feelings of death anxiety. For example, Krause and Bastida (2012) evaluated a conceptual model which shows that perceived contact with deceased loved ones fosters a stronger sense of religiously based feelings of connectedness with others, a greater sense of connectedness with others promotes stronger feelings of gratitude to God, and greater gratitude to God is, in turn, associated with less death anxiety. In the discussion that follows, a conceptual model is developed that relates four different dimensions of religion with death anxiety. Two main themes are captured in this conceptual scheme. First, the model of religion and death anxiety is socially based. This means that the reason why religion may help people deal more effectively with death anxiety arises from the social interaction that takes place in religious institutions. This orientation is consistent with what sociologists and psychologists have been arguing for over 100 years. For example, 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 along” (Baldwin, 1902, p. 325). The social aspects of religion are assessed with both formal measures of social contact (i.e., attendance at worship services) as well as informal interaction with

Krause

23

fellow church members (i.e., spiritual support). The second main theme involves the specific religiously oriented beliefs that arise from this social contact and are proposed to reduce feelings of death anxiety. Trust in God and feeling forgiven by God figure prominently in this respect. This appears to be the first time that the relationships among spiritual support, trust in God, and forgiveness by God have been examined in studies on death anxiety.

Religious Involvement, Forgiveness by God, and Death Anxiety Figure 1 contains the conceptual model that was evaluated in this study. Two steps were taken to simplify the presentation of this conceptual scheme. First, the elements of the measurement model (i.e., the observed indicators and measurement error terms) are not depicted graphically even though a full measurement model was evaluated when this conceptual model was estimated empirically. Second, the relationships among the latent constructs in Figure 1 were estimated after the effects of age, sex, education, race, and marital status were controlled statistically (i.e., treated as exogenous variables). All of the paths among the latent constructs were estimated in the analyses presented below (i.e., a fully saturated model was evaluated). However, the theoretical process that this conceptual scheme was designed to capture is depicted in the following core hypotheses: (a) People who go to church more often receive more spiritual support from the individuals who worship there

Figure 1. A conceptual model of trust in God, forgiveness by God, and death anxiety.

24

OMEGA—Journal of Death and Dying 72(1)

(spiritual support is assistance that is provided by coreligionists for the explicit purpose of increasing the religious beliefs and practices of the recipient). (b) Individuals who receive more spiritual support from fellow church members will be more likely to trust God. (c) Those who trust God more deeply will be more likely to feel forgiven by Him. (d) People who feel they have been forgiven by God will experience less death anxiety. The theoretical rationale for each of these relationships is provided below.

Church Attendance and Spiritual Support The conceptual model that was developed for this study is inherently social in nature. This emphasis is justified because scholars have been arguing for over 100 years that the essence of religion may be found in the relationships that are formed among individuals who share a common faith (Baldwin, 1902; Graham & Haidt, 2010). Consistent with this view, the process that is depicted in Figure 1 begins with the relationship between two key aspects of congregational life: church attendance and spiritual support. There are three reasons why more frequent church attendance should be associated with receiving more spiritual support. The first reason is straightforward. In order for an individual to receive spiritual support, they must obviously come into contact with potential support providers. Attending worship services more frequently provides the necessary context for this contact to take place. Second, in order for a supportive relationship to arise, support providers and support recipients must have contact on a fairly regular basis. Once again, regular attendance at worship services provides the context for developing and maintaining supportive relationships. Third, as Roberts (2007) argued, faith becomes more alive and more meaningful when it is shared with others: The church is a society of people who have undertaken the struggle to love one another with a spiritual love. We teach one another, week in and week out . . . We cultivate ourselves and one another in the consciousness of a calling to perfect fellowship. (p. 91)

If fellow church members subscribe to these teachings, then more frequent contact during worship services with like-minded others should provide a greater opportunity to put these religious teachings into practice through the exchange of spiritual support.

Church Attendance, Spiritual Support, and Trust in God The nature of the afterlife is shrouded in mystery. In fact, it is not possible to empirically verify that there is an afterlife in the first place. As a result, people must take religious teachings on the nature and existence of the afterlife on faith.

Krause

25

Put another way, people must trust that the word of God about the afterlife that appears in sacred texts is valid and accurate. It is for this reason that trust of God is contained in Figure 1. An important issue that emerges at this juncture has to do with how this sense of trust in God is developed and maintained in the first place. If the core assumption in this study about the social basis of religion is accurate, then the genesis and maintenance of trust in God should be found in key aspects of congregational life. As shown in Figure 1, it is proposed that trust in God arises from the interplay between attendance at worship services and spiritual support. Messages involving the importance of trusting God are often embedded in a number of aspects of formal worship services, including sermons, hymns, and congregational prayers. It follows from this that more frequent attendance at worship services should be associated with greater trust in God. This view is consistent with the theory of religion that was developed by Stark and Finke (2000). These investigators maintained that confidence in religious principles increases to the extent that people participate in religious rituals, such as worship services. Another important source of trust in God may be found in the informal spiritual support that is exchanged among fellow church members. Evidence that religious teachings are maintained and reinforced through social processes is found in Berger’s (1967) classic treatise on religion. He maintained that religious precepts are socially constructed and socially maintained. Berger (1967) went on to point out that this informal reinforcement of religious teachings is not a one-time affair. Instead, he maintains that people continuously reinforce and reconstruct basic religious beliefs. This view is also consistent with another proposition that is found in Stark and Finke’s (2000) theory of religion. These investigators argued that a person’s confidence in religious teachings and principles increases to the extent that fellow church members express confidence in them. Based on these insights, it is predicted in Figure 1 that people who receive more spiritual support from fellow church members will have greater trust in God.

Trust in God and Forgiveness by God Being forgiven by God is the linchpin in Figure 1 because it bridges the dimensions of religion that have been discussed so far with feelings of death anxiety. However, before discussing why forgiveness by God may help reduce feelings of death anxiety, it is important to reflect on how people come to believe they are forgiven by God in the first place. As with many other aspects of religious life, there is no way to empirically confirm that people have been forgiven by God. Instead, people must seek assurance that they have been forgiven through indirect means, such as

26

OMEGA—Journal of Death and Dying 72(1)

teachings in sacred texts. Sacred texts, like the Christian Bible, typically convey three messages that are relevant here. First, sacred texts repeatedly indicate that God is aware of the fallible nature of all human beings. Second, sacred texts state that even though all people have shortcomings, God wants what is best for the individuals He created. Third, because God cares for all people, He is willing to forgive them for their transgressions. In fact, research by Lundberg (2010) reveals that every major faith tradition in the world endorses each of these precepts. However, because there is no empirical verification of these religious teachings, belief in them rests on trust in God. It is for this reason that it is proposed in Figure 1 that people who trust in God are more likely to feel they have been forgiven by Him. There do not appear to be any studies in the literature that empirically evaluate the relationship between trusting in God and feeling forgiven by God. One contribution of the current study arises from the fact that an attempt is made to see if the two are associated.

Forgiveness by God and Death Anxiety There are two reasons why forgiveness by God should be associated with lower feelings of death anxiety. The first is straightforward. If people believe they have not been forgiven by God, then they should be more likely to believe they will be punished by God for their transgressions in the next life. And if people believe that punishment awaits them after death, then they should experience more death anxiety. The second reason why feeling forgiven by God may lower feelings of death anxiety may be found by returning to TMT. Greenberg, Solomon, and Pyszczynski (1997) argued that self-esteem is a core tenet in TMT. According to this view, a strong sense of self-esteem arises and is maintained by displaying the attributes, behaviors, and achievements that are embedded in world views that have been developed to help people cope with their pending demise. By fulfilling the roles that are specified in these world views, people believe they are valued members of the universe who are worthy of the good things that the next life presumably provides. Greenberg et al. (1997) specifically mentioned that religion is an especially important source of these world views, but they fall short in their discussion of precisely how this important function is performed by the various faith traditions. This is especially true when the fallible aspects of human nature are taken into account. Because people inevitably fail to always fulfill the roles that are embedded in religious world views, it is not clear how they are still able to maintain the strong sense of self-worth that is needed to offset their fear of dying. If anything, their inevitable transgressions should serve to heighten death anxiety. Research by Barna (2002) provides a point of departure for showing why this may be so. He reports the results of six nationwide surveys which suggest that between 68% and 73% of all adults believe that God is the

Krause

27

all-knowing all-powerful creator of the universe. Kirkpatrick (2005) identified the psychological implications of these beliefs. He argued that if people view God as all-knowing, then He will be aware of their shortcomings. And if God is all-powerful, then He is fully capable of extracting revenge and punishing people for their transgressions. Kirpatrick (2005) went on to argue that these realizations are likely to produce considerable anxiety and fear. Although Kirkpatrick (2005) did not discuss self-esteem specifically, it is not hard to see why the consequences of believing in an all-knowing and all-powerful God are likely to diminish a person’s sense of self-esteem. A central premise in the current study is that feeling forgiven by God is the key to resolving the dilemma that is created by the need to live up to religious teachings and the failure to do so on every occasion. When people fail to live up to the tenets of their faith, they may experience feelings of guilt. This is important because research reveals that guilt is associated with diminished feelings of self-worth (Strelan, 2007). However, research further reveals that forgiveness can help bolster feelings of self-worth that have been eroded by guilt that arises from a transgression (Strelan, 2007). It follows from this that people who feel they have been forgiven by God are less likely to experience strong feelings of death anxiety because the forgiveness they have received helps restore and bolster feelings of self-worth that are needed to offset concerns about death.

Aging, Religion, and Death Anxiety As the discussion in the next section will reveal, the participants in the current study are age 50 and older. Consequently, it is important to discuss why it is important to focus on the relationship between religion and death anxiety in a sample that comprises middle-aged and older individuals. Insight into this issue may be found by turning to two patterns of findings that have emerged from the literature. Findings from the first group of studies suggest that the fear of death begins to taper off during midlife and does not decline further as people go through late life (e.g., Cicirelli, 2002). Initially, these data may be somewhat difficult to understand because older adults are closer to death than younger adults and older people are more likely to experience acute and chronic health conditions that may hasten their demise. Findings from a second group of studies indicate that people who are currently older tend to be more deeply involved in religion than individuals who are presently younger (see Krause, 2008, for a review of this research). Many researchers argue that these findings suggest that people tend to become more religious as they grow older. Taken together, these two bodies of research suggest that greater involvement in religion may promote less fear of death in late life. This is in fact what some studies indicate (e.g., Cicirelli, 2002). However, as Krause (2008) points out,

28

OMEGA—Journal of Death and Dying 72(1)

many researchers rely on crude measures of religious involvement, such as the frequency of church attendance. Consequently, it makes sense to reexamine the relationship between religion and death anxiety in samples of middle-aged and older adults with more sophisticated measures of religious involvement. One goal of the current study is to address this issue.

Methods Sample The data for this study come from an ongoing nationwide survey of Whites and African Americans. Altogether, five waves of interviews have been conducted. The study population for the baseline survey was defined as all household residents who self-identified as either Black or White, noninstitutionalized, Englishspeaking, 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). 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 it would be too difficult to devise a comprehensive battery of religion measures that would be suitable for individuals of all faiths. The three groups were not selected so that they could be compared and contrasted. Instead, the goal was to be able to clearly and cleanly assess the relationship between one faith tradition and health-related outcomes. The sampling frame consisted of all eligible persons contained in the beneficiary list maintained by the Centers for Medicare and Medicaid Services. A five-step process was used to draw the sample from the Centers for Medicare and Medicaid Services Files (see Krause, 2002a, for a detailed discussion of these steps). The baseline survey took place in 2001. The data collection for all waves of interviews was conducted by Harris Interactive (New York). A total of 1,500 interviews were completed, face-to-face, in the homes of the study participants. African Americans were oversampled so that sufficient statistical power would be available to assess racial cultural differences in religion. The overall response rate for the baseline survey was 62%. The Wave 2 survey was conducted in 2004. A total of 1,024 study participants were reinterviewed successfully. The reinterview rate was 80%. A third wave of interviews was completed in 2007. A total of 969 older study participants were reinterviewed successfully (reinterview rate was 75%). Wave 4 was completed in 2008. A total of 718 older study participants were reinterviewed successfully (reinterview rate ¼ 88%).

Krause

29

A fifth wave of interviews was completed in June, 2013. However, the sampling strategy for this round of interviews was complex. By the time the Wave 5 interviews were conducted, only 229 study participants were reinterviewed successfully. Many former study participants experienced significant health problems that were associated with their advanced age (M ¼ 83.2 years). Moreover, a number of former study participants died (N ¼ 611). So in order to have sufficient statistical power to conduct meaningful analyses, the following two-part sampling strategy was employed. First, we reinterviewed as many of the original study participants as possible (N ¼ 229). Second, this group was supplemented with a new sample of individuals who had not participated in the study previously (N ¼ 1,306). In the process of fielding the sample of new study participants, we lowered the age for eligibility from 66 to 50. This was done in order to make it possible to evaluate issues involving religious involvement in midlife, which is not the focus of the current study. The reinterview rate for prior study participants was 63%. The response rate for people in the new supplementary sample was 45%. There are clear differences between the old and new samples at Wave 5 (e.g., the old sample comprises people who are obviously older). But the upshot of the sampling plan was to derive a sample that is representative of U.S. adults age 50 and over when the old and new samples are combined. The sample of individuals who had not participated previously in the study was obtained in the following manner. Based on the data in the 2010 Census, 50 geographic areas (i.e., Census tracts) were selected to proportionally represent the population age 50 and over who were either White or African American. All households within each Census tract were enumerated. One eligible person per household was selected at random to participate in the study. Altogether, a total of 1,535 individuals participated in the Wave 5 interviews. The analyses presented below are based on the Wave 5 data only because this was the first time that questions on trust in God were administered. Recall that spiritual support is contained in the conceptual model for this study. When the questionnaire for this study was being developed, the members of the research team felt it did not make sense to ask questions about receiving spiritual support from fellow church members if study participants either never attend worship services or if they go to church only one or two times a year. Consequently, 381 participants with low rates of church attenders were excluded from the analyses presented below, resulting in a sample size of 1,154. The Full Information Maximum Likelihood (FIML) procedure was handle item nonresponse (Enders, 2010). Preliminary analysis revealed that the average age of the participants in the pooled study sample was 63.4 years (SD ¼ 11.7 years), 35.9% were men, 43.1% were married at the time of the Wave 5 survey, 61.1% self-identified as White, and the average level of educational attainment was 13.1 years of schooling (SD ¼ 2.3 years).

30

OMEGA—Journal of Death and Dying 72(1)

Measures Table 1 contains the indicators that were used to assess the core constructs in Figure 1. The procedures that were used to code these items are provided in the footnotes of this table. Church attendance. A single item was taken from research by the Fetzer Institute/ National Institute Working Group (1999) to measure how often study participants attend worship services. A high score denotes more frequent church attendance. This item is scored in the following manner (coding in parenthesis): never (1), less than once a year (2), about once or twice a year (3), several times a year (4), about once a month (5), 2 to 3 times a month (6), nearly every week (7), every week (8), and several times a week (9). The mean of this ordinal measure is 6.7 (SD ¼ 1.7).

Table 1. Core Study Measures. 1. Church attendance How often do you attend religious services? 2. Spiritual support A. Not counting Bible study groups, prayer groups, or church services, how often does someone in your congregation share their own religious experiences with you? B. How often does someone in your congregation help you find solutions to your problems in the Bible? C. How often do the examples set by others in your congregation help you lead a better religious life? D. How often does someone in your congregation help you to know God better? E. How often does someone in your congregation help you live according to your religious beliefs? 3. Trust in God A. I trust God completely. B. I believe that God will never let me down. C. I know that God will ultimately do what is best for me. D. I know God will never give me more than I can handle. E. I trust God because He has always been right there for me in the past. 4. Forgiveness by God A. I have done some things that even God may not forgive B. I believe God forgives me for the things I have done wrong. 5. Death anxiety A. I find it hard to face up to the fact that I will die. B. Thinking about death makes me feel uneasy. C. I do not feel prepared to face my own death. D. I am disturbed by the shortness of life.

Krause

31

Spiritual support. Five indicators that were developed by Krause (2002b) were used to assess spiritual support. A high score stands for study participants who received spiritual support from fellow church members more often. The spiritual support items are scored in the following manner: never (1), once in a while (2), fairly often (3), and very often (4). The mean spiritual support value is 13.0 (SD ¼ 4.2). Trust in God. Based on the procedures described by Krause (2002b), the following item development strategy was devised to create new measures of trust in God. First, 41 open-ended in-depth interviews were conducted in order to see how the typical individual thinks about things like trust in God. Following this, the main themes that emerged from these interviews were used to draft closed-ended item stems. Once a preliminary set of measures were in place, 40 cognitive interviews were conducted with a new sample of study participants. Cognitive interviews involve presenting study participants with newly developed closed-ended items that are followed by a series of open-ended probes. The probes were designed to determine if the item was understood in the intended manner and if there are better ways to phrase the question stems. Following this, the indicators were revised and evaluated with standard pilot testing procedures. The procedures that are described above were used to craft five items that assess trust in God. A high score stands for greater trust in God. The items that assess trust in God are scored in the following manner: strongly disagree (1), disagree (2), agree (3), and strongly agree (4). The mean is 18.1 (SD ¼ 2.6). Forgiveness by God. As shown in Table 1, two items that come from research by Krause (2002b) were used to assess whether study participants feel they have been forgiven by God. The first item was scored in the following manner: strongly agree (1), agree (2), uncertain (3), disagree (4), and strongly disagree (5). The second indicator was scored in the following manner: strongly disagree (1), disagree (2), uncertain (3), agree (4), and strongly agree (5). A high score denotes greater feelings of being forgiven by God. Summing the two indicators produced a brief composite with a mean of 8.5 and a standard deviation of 1.5. The bivariate correlation between the two measures of forgiveness by God is .242 (p < .001).

Death Anxiety A number of researchers assess death anxiety with complex, multidimensional scales, such as the Mutidimensional Fear of Death Scale (Hoelter, 1979), which captures distinct ways in which death anxiety is manifest (e.g., fear of the dying process and fear of being destroyed). In contrast to this comprehensive approach, four indicators were used in the current study to measure death anxiety. These items were taken from scales that are in the literature (see

32

OMEGA—Journal of Death and Dying 72(1)

Neimeyer, 1994, for a discussion of these measures). The indicators in this unidimensional scale assess more general or global fear of death as a whole. These items are coded so that a high score reflects more anxiety about dying. The death anxiety items are scored in the following manner: strongly disagree (1), disagree (2), agree (3), and strongly agree (4). The mean of the brief composite that was developed by summing the scores on these items is 8.6 (SD ¼ 2.7). Demographic control variables. Recall that the relationships among the core study measures were assessed after the effects of age, sex, education, marital status, and race were controlled statistically for. Age and education were coded continuously in years. In contrast, sex (1 ¼ men; 0 ¼ women), race (1 ¼ White; 0 ¼ Black), and marital status (1¼ presently married; 0 ¼ otherwise) were scored in a binary format.

Results The findings from this study are presented below in three sections. Some technical issues involving the estimation of the study model are discussed in the first section. Following this, data on the psychometric properties of the multiple item constructs are presented in the second section. The final section contains the substantive findings that were derived from estimating the study model.

Model Estimation Issues The model depicted in Figure 1 was evaluated with the maximum likelihood estimator in Version 8.80 of the LISREL statistical software program (du Toit & du Toit, 2001). Use of this estimator is based on the assumption that the observed indicators have a multivariate normal distribution. Preliminary tests (not shown here) revealed that this assumption had been violated in the current study. Although there are a number of ways to deal with departures from multivariate normality, the straightforward approach that is discussed by du Toit and du Toit (2001) was followed here. These investigators report that departures from multivariate normality can be handled by converting raw scores of the observed indicators to normal scores prior to estimating the model (du Toit & du Toit, 2001, p. 143). Based on these insights, the analyses presented below were performed with observed indicators that were normalized. Because the FIML procedure was used to deal with item nonresponse, the LISREL software program provides only two goodness-of-fit measures. The first is the full information maximum likelihood chi-square value (365.542 with 170 degrees of freedom, p < .000). Unfortunately, this statistic substantially underestimates the fit of the model to the data when samples are large, such as the sample in the current study (see Fornell & Larcker, 1981; Joreskog, 1969). Better insight into the fit of the model to the data is provided by the second

Krause

33

goodness-of-fit measure—the root mean square error of approximation. The root mean square error of approximation value for the model in Figure 1 is .032. Although there is some debate about the best cut point for determining an adequate fit of the model to the data, most investigators would agree that values below .05 indicate a good fit to the data (Hu & Bentler, 1999).

Psychometric Properties of the Multiple Item Measures The factor loadings and measurement error terms that were derived from estimating the study model are provided in Table 2. These coefficients provide

Table 2. Measurement Error Parameter Estimates for Multiple Item Study Measures (N ¼ 1,154). Construct 1. Spiritual support A. Others share religious experiencesc B. Find solutions to problems in Bible C. Lead a better religious life D. Help you know God better E. Live according to religious beliefs 2. Trust in God A. Trust God completely B. God never lets me down C. God will do what is best D. Never give more than I can handle E. He has always been there 3. Forgiveness by God A. Done things God would not forgive B. Feel forgiven by God 4. Death anxiety A. Hard to face up to death B. Death makes me uneasy C. Not prepared for death D. Disturbed by shortness of life a

Factor loadinga

Measurement errorb

.798 .787 .780 .846 .829

.363 .381 .392 .285 .313

.847 .908 .875 .823 .876

.283 .176 .234 .322 .232

.430 .703

.815 .505

.834 .858 .844 .773

.305 .265 .288 .403

Factor loadings are from the completely standardized solution. The first-listed item for each latent construct was fixed at 1.0 in the unstandardized solution. b Measurement error terms are from the completely standardized solution. All factor loadings and measurement error terms are significant at the .001 level. c Item content is paraphrased for the purpose of identification. See Table 1 for the complete text of each indicator.

34

OMEGA—Journal of Death and Dying 72(1)

information about the reliability of the multiple item study measures. However, researchers have yet to reach a consensus on the cut point for determining whether the magnitude of a factor loading is acceptable. Widaman (2012) suggests that items with standardized factor loadings in excess of .600 tend to have good reliability. The data in Table 2 reveal that only one coefficient was below .700. Even so, this coefficient (.430), which is associated with being forgiven by God, was retained. Eliminating this item would have left only one observed indicator to assess forgiveness by God. When latent constructs are assessed with only one item, researchers must assume that indicator contains no measurement error. This assumption is frequently untenable in psychosocial research. This issue aside, the overall magnitude of the standardized factor loadings suggests that the reliability of the multiple-item constructs is acceptable. Although the factor loadings and measurement error terms that are associated with the observed indicators provide useful information about the reliability of each item, it would be helpful to know something about the reliability for the scales as a whole. It is possible to compute these reliability estimates with a formula provided by DeShon (1998). This procedure is based on the factor loadings and measurement error terms. Applying the procedures described by DeShon (1998) to these data in Table 2 yields the following reliability estimates for the multiple item constructs in Figure 1: spiritual support (.904), trust in God (.938), and death anxiety (.895). Fornell and Larcker (1981) provide an alternative way of gauging the psychometric properties of study measures. They recommend researchers compute the average variance extracted (AVE) in a multiple item measure. The AVE reflects the amount of variance in a set of measures that is captured by the latent construct in relation to the amount of variance that is due to measurement error. They recommend that AVE values exceed .50. The following AVE values were observed for the multiple item study measures: spiritual support (.653), trust in God (.746), and death anxiety (.685). The only measure to fall below the cutoff of .50 was forgiveness by God (.340). However, Fornell and Larcker (1981) point out that the AVE is a conservative measure and that a researcher may conclude that a construct is adequate even though the AVE estimate is below .50.

Substantive Findings Table 3 contains the estimates of the relationships among the latent constructs in the study model. Taken as a whole, these findings tend to support the theoretical rationale that was developed earlier. The results reveal that people who go to church more often report receiving more spiritual support from their fellow church members (b ¼ .277; p < .001). Moreover, the data indicate that individuals who go to church more often (b ¼ .128; p < .001) and people who receive more spiritual support (b ¼ .322; p < .001) are more likely to have greater trust

Krause

35

Table 3. Trust in God, Forgiveness by God, and Death Anxiety (N ¼ 1,154). Dependent variables Independent Age Sex Education Race Marital Status

Church attendance

Spiritual support

Trust in God

.143***a (.021)b .020 (.071) .011 (.008) .057 (.195) .136*** (.461)

.146*** (.009) .028 (.044) .048 (.016) .201*** (.307) .003 (.005) .277*** (.123)

.056* (.002) .027 (.027) .050 (.010) .169*** (.164) .018 (.017) .128*** (.036) .322*** (.205)

.006 (.001) .069 (.077) .104*** (.024) .075*** (.082) .100** (.108) .084* (.027) .087 (.062) .519*** (.585)

.037

.137

.204

.347

Church Attendance Spiritual Support Trust in God

Forgiveness by God

Forgiveness by God Multiple R2

Death anxiety .120*** (.007) .064* (.090) .027 (.008) .074* (.102) .006 (.009) .046 (.018) .028 (.025) .015 (.021) .325*** (.407) .152

a

Standardized regression coefficient. Metric (unstandardized) regression coefficient. *p < .05. ** p < .01. *** p < .001. b

in God. The findings further suggest that study participants who have more trust in God are more likely to believe that God has forgiven them for the transgressions they have committed (b ¼ .519; p < .001). The findings further reveal that people who feel they have been forgiven by God tend to have lower levels of death anxiety (b ¼ .325; p < .001). It is possible to further refine the theoretical insights in this study by turning to the indirect and total effects that operate through the study model. A simple example will help clarify the meaning of these terms. According to the rationale that was provided earlier, trust in God is associated with feeling forgiven by God and forgiveness by God is, in turn, associated with lower death anxiety. If this turns out to be the case, then the findings would reveal that trust in

36

OMEGA—Journal of Death and Dying 72(1)

God exerts an indirect effect on death anxiety that operates through feeling forgiven by God. When this indirect effect is added to the direct effect of trust in God on death anxiety, the resulting total effect provides a more comprehensive view of the relationship between trust in God and death anxiety. Breaking down the relationship between study measures into direct, indirect, and total effects is often called the “decomposition of effects” in the literature (Alwin, 1988). Based on this logic, the relationship between three pairs of measures will be decomposed in order to sharpen the theoretical perspective that has been developed in this study. The first decomposition involves the example that was provided above (i.e., the relationship between trust in God and death anxiety). The data in Table 3 reveal that the direct effect of trust in God on death anxiety is not statistically significant (b ¼ .015; ns). However, further analysis (not shown in Table 3) reveals that the indirect effect of trust in God on death anxiety that operates through feeling forgiving by God is statistically significant (b ¼ .168; p < .001). Summing the direct and indirect effects shows that the total effect of trust in God on death anxiety is also statistically significant (.105 + .168 ¼ .183; p < .001; not shown in Table 3). Viewed in more substantive terms, this decomposition of effects suggests that, within the confines of the study model, trust in God is associated with lower levels of death anxiety primarily because people who place their trust in God are more likely to feel they have been forgiven by Him. The second decomposition has to do with the relationship between church attendance and death anxiety. Recall that findings from other studies suggest that more frequent church attendance is associated with less death anxiety (Morris & McAdie, 2009). Initially, the findings in Table 3 appear to be at odds with these findings because the data suggest that the direct effect of church attendance on death anxiety is not statistically significant (b ¼ .046; ns). However, a different picture emerges from the data when the indirect and total effects that operate through the study model are taken into account. More specifically, these additional analyses (not shown in Table 3) indicate that the indirect effect of church attendance on death anxiety that operates through the other variables in the study model is statistically significant (b ¼ .083; p < .001). Summing the direct and indirect effects produces a total effect that is more in line with the findings that have been reported by other investigators (.046 + .083 ¼ .129; p < .001). This decomposition of effects underscores importance of one of the criticisms of the literature that was raised earlier in this study. Earlier, studies that focus solely on the relationship between church attendance and death anxiety were faulted because they fail to show why the two are related. By tracing the indirect effects of church attendance that operate through spiritual support, trust in God, and forgiveness by God, the findings from the current study provide more insight into why church attendance might allay feelings of death anxiety.

Krause

37

The third decomposition of effects has to do with the relationship between spiritual support and death anxiety. The data in Table 3 initially seem to suggest that spiritual support is not associated with feelings of death anxiety (b ¼ .028; ns). However, this finding appears to be inconsistent with social constructionist views of religion (e.g., Berger, 1967), which maintain that religious beliefs arise from and are reinforced by interaction with like-minded others. An examination of the indirect and total effects of spiritual support on death anxiety tends to bring the findings from the current study in line with the social constructionist perspective. Additional analyses (not shown in Table 3) indicate that the indirect effect of spiritual support on death anxiety that operates through trust in God and forgiveness by God is statistically significant (b ¼ .087; p < .001). When the direct and indirect effects are summed, the resulting total effect (b ¼ .115; p < .001; not shown in Table 3) suggests that individuals who receive more spiritual support from their fellow church members tend to experience less death anxiety.

Supplementary Analysis The AVE estimates that were provided earlier are also useful because they can be used in conjunction with the estimate of the amount of variance that is explained in the relationship between two constructs to provide an estimate of discriminant validity (Fornell & Larcker, 1981, p. 46). As the measure of trust in God is new, it would be helpful to provide an estimate of the discriminant validity of this index. This will be accomplished by applying the formula given by Fornell and Larcker (1981) to the relationship between the frequency of church attendance and trust in God. The correlation between the two latent constructs is .226 (not shown in Table 3). Fornell and Larcker (1981) suggest that if the AVE value is greater than the squared correlation between the two measures the requirements for discriminant validity have been satisfied. Recall that the AVE for the trust in God measure is .746. This is substantially larger than the squared value of the correlation between church attendance and trust in God. This analysis suggests that there is some evidence for the discriminant validity of the new trust in God measure. Because the death anxiety measure that is used in this study is brief and has not appeared elsewhere in its current form, it is also helpful to provide evidence of the discriminant validity of this construct as well. This will be accomplished by focuses on the relationship between forgiveness by God and death anxiety. The squared correlation between the two latent constructs is .037 (not shown in Table 3). The AVE value for the death anxiety measure that was reported earlier is .685, which is once again larger than the amount of variance that is explained in the relationship between forgiveness by God and death anxiety. Consequently, evidence for the discriminant validity of the death anxiety measure is provided by these analyses.

38

OMEGA—Journal of Death and Dying 72(1)

Discussion Many studies that examine the relationship between religion and death anxiety focus on one or two measures of religion and typically evaluate only the direct effects of these measures on the fear of death. The purpose of the current study was to pursue a different strategy that involved creating a conceptual model to assess the interplay among multiple dimensions of religion and death anxiety. Data from a recent nationwide survey provide support for the following core relationships in this conceptual scheme: Individuals who attend worship services more often tend to receive more spiritual support from fellow church members, people who get more spiritual support tend to have greater trust in God, those who trust God more are more likely to feel they have been forgiven by God, and people who feel God has forgiven them tend to experience less death anxiety. Further analysis of the indirect and total effects that operate through the study model helped to sharpen the theoretical conclusions that can be drawn from this model. For example, the data suggest that greater trust in God is associated with less death anxiety, but this relationship is due solely to fact that people who trust in God are more likely to feel they have been forgiven by God and this forgiveness is, in turn, associated with less death anxiety. Even though the findings from the current study provide some potentially useful insights into the relationship between religious involvement and death anxiety, a considerable amount of work remains to be done. Forgiveness by God was assessed with only two indicators in the current study. However, the process leading to the belief that one has been forgiven by God is likely to involve more than is assessed by these items. For example, a study by Krause and Hayward (2013) that was conducted with data provided by older Mexican Americans provides some insight into the finer nuances of this process. These investigators found that performing acts of contrition in order to earn God’s forgiveness is associated with lower death anxiety among members of this ethnic group. Acts of contrition involve things like asking for God’s forgiveness, promising God that the transgression will not be repeated, and providing restitution for a transgression whenever possible. Research is needed to find out if these findings can be generalized beyond older Mexican Americans. Another possibility is found in a study by McConnell and Dixon (2012). These researchers argue that people who feel they have been forgiven by God are more likely to forgive themselves. Perhaps people who are more likely to forgive themselves are, in turn, less likely to experience death anxiety. In addition to pursuing new issues in research on death anxiety, researchers should also address the limitations in the current study. It is important to briefly discuss two shortcomings here. First, the data for this study are cross-sectional, and as a result, the temporal ordering among the constructs in the study model was based on theoretical considerations alone. Consequently, other investigators might propose other legitimate causal orderings among the study constructs. For example, the model in Figure 1 specifies that people with greater trust in God are

Krause

39

more likely to be forgiven by God, but it could just as easily be argued that people who feel they have been forgiven by God are subsequently more likely to trust Him. This as well as other causal assumptions in the study model must be more rigorously assessed with data that have been gathered at more than one point in time. Second, researchers should use more sophisticated, multidimensional measures to assess death anxiety rather than the four brief global measures of death anxiety that were used in the current study. As research on death anxiety has revealed, many people are deeply concerned about their eventual demise (Neimeyer, 1994). Religion is uniquely positioned to ease concerns associated with death because it deals directly with issues like forgiveness and because it holds out the promise of life after death for those who remain faithful to the teachings of the church. Yet research on religion and death anxiety is not fully developed. Hopefully, the strategy that was implemented in the current study will encourage other researchers to delve more deeply into the complex processes that link religion with feelings of death anxiety. Declaration of Conflicting Interests The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was supported by grants from the John Templeton Foundation and the National Institute on Aging (RO1 AG014749).

References Alwin, D. (1988). Structural equation modeling in research on human development and aging. In K. W. Schaie, R. T. Campbell, W. Meredith & S. C. Rawlinngs (Eds.), Methodological issues in aging research (pp. 71–170). New York, NY: Springer. Ardelt, M., & Koenig, C. S. (2006). The role of religion for hospice patients and relatively healthy older adults. Research on Aging, 28, 184–215. Baldwin, J. M. (1902). Fragments in philosophy and science of being: Collected essays and addresses. New York, NY: Charles Scribner’s Sons. Barna, G. (2002). The state of the church 2002. Ventura, CA: Issachar Resources. Berger, P. L. (1967). The sacred canopy: Elements of a sociological theory. New York, NY: Doubleday. Cicirelli, V. G. (2002). Fear of death among older adults: Predictions from Terror Management Theory. Journal of Gerontology: Psychological Sciences, 57B, P358–P366. DeShon, R. P. (1998). A cautionary note on measurement error correlations in structural equation models. Psychological Methods, 3, 412–423. du Toit, M., & du Toit, S. (2001). Interactive LISREL: User’s guide. Lincolnwood, IL: Scientific Software International.

40

OMEGA—Journal of Death and Dying 72(1)

Ellis, L., & Wahab, E. A. (2013). Religiosity and fear of death: A theory-oriented review of the empirical literature. Review of Religious Research, 55, 149–189. Enders, C. C. (2010). Applied missing data analysis. New York, NY: Guilford. Fetzer Institute/National Institute on Aging Working Group. (1999). Multidimensional measurement of religiousness/spirituality for use in health research. Kalamazoo, MI: John E. Fetzer Institute. Feifel, H. (1959). The meaning of death. New York, NY: McGraw Hill. Fornell, C., & Larcker, D. F. (1981). Evaluating structural equation models with unobserved variables and measurement error. Journal of Marketing Research, 18, 39–50. Graham, J., & Haidt, J. (2010). Beyond beliefs: Religions bind individuals to moral communities. Personality and Social Psychology Review, 14, 140–150. Greenberg, J., Pyszczynski, T., & Solomon, S. (1986). The causes and consequences of a need for self-esteem: A terror management theory. In R. F. Baumeister (Ed.), Public self and private self (pp. 189–212). New York, NY: Springer-Verlag. Greenberg, J., Solomon, S., & Pyszczynski, T. (1997). Terror Management Theory of selfesteem and cultural worldviews: Empirical assessments and conceptual refinements. Advances in Experimental Social Psychology, 29, 61–139. Hoelter, J. W. (1979). Multidimensional treatment of fear of death. Journal of Consulting and Clinical Psychology, 47, 996–999. Hu, L., & Bentler, P. M. (1999). Cutoff points for fit indices in covariance structure analysis: Conventional criteria versus new alternatives. Psychological Methods, 1, 130–149. Joreskog, K. G. (1969). A general approach to confirmatory maximum likelihood factor analysis. Psychometrika, 34, 183–202. Kirkpatrick, L. A. (2005). Attachment, evolution, and the psychology of religion. New York, NY: Guilford. Krause, N. (2002a). Church-based social support and health in old age: Exploring variations by race. Journal of Gerontology: Social Sciences, 57B, S332–S347. Krause, N. (2002b). A comprehensive strategy for developing closed-ended survey items for use in studies of older adults. Journal of Gerontology: Social Sciences, 57B, S263–S274. Krause, N. (2008). Aging in the church. How social relationships affect health. West Conshohocken, PA: Templeton Foundation Press. Krause, N., & Bastida, E. (2012). Contact with the dead, religion, and death anxiety among older Mexican Americans. Death Studies, 36, 932–948. Krause, N., & Hayward, R. D. (2013). Acts of contrition, forgiveness by God, and death anxiety among older Mexican Americans. International Journal for the Psychology of Religion, 25, 57–73. Lin, A. H. (2003). Factors related to attitudes toward death among American and Chinese older adults. Omega: Journal of Death and Dying, 47, 3–23. Lundberg, G. D. (2010). Unifying truths of the world’s religions. New Fairfield, CT: Heavenlight Press. Lundh, L. G., & Radon, V. (1998). Death anxiety as a function of belief in an afterlife: A comparison between a questionnaire measure and Stroop measure of death anxiety. Personality and Individual Differences, 25, 487–494. McConnell, J. M., & Dixon, D. N. (2012). Perceived forgiveness from God and selfforgiveness. Journal of Psychology and Christianity, 31, 31–39.

Krause

41

Morris, G. J., & McAdie, T. (2009). Are personality, well-being, and death anxiety related to religious affiliation?. Mental Health, Religion & Culture, 12, 115–120. Neimeyer, R. A. (1994). Death anxiety handbook: Research instrumentation, and application. New York, NY: Taylor & Francis. Roberts, R. C. (2007). Spiritual emotions: A psychology of Christian virtues. Grand Rapids, MI: William B. Eerdmans. Soenke, M., Landau, M. J., & Greenberg, J. (2012). Sacred armor: Religion’s role as buffer against the anxieties of life and the fear of death. In K. I. Pargament (Ed.), APA handbook of psychology, religion, and spirituality (pp. 105–122). Washington, DC: American Psychological Association. Stark, R., & Finke, R. (2000). Acts of faith: Explaining the human side of religion. Berkeley, CA: University of California Press. Strelan, P. (2007). Who forgives others, themselves, and situations? The roles of narcissism, guilt, self-esteem, and agreeableness. Personality and Individual Differences, 42, 259–269. Widaman, K. F. (2012). Exploratory factor analysis and confirmatory factor analysis. In H. Cooper (Ed.), APA handbook of research methods in psychology (Vol. 3, pp. 361–389). Washington, DC: American Psychological Association.

Author Biography Neal Krause, PhD, is the Marshall H. Becker Collegiate Professor in the Department of Health Behavior and Health Education at the University of Michigan School of Public Health. His research focuses on religion and health. An emphasis is placed in this work on the health-related benefits of support that is provided by religious others.