Assessing the Prayer Lives of Older Whites, Older Blacks and Older Mexican Americans: A Descriptive Analysis

Abstract

The purpose of this study was to see whether differences emerge between older whites, older blacks, and older Mexican Americans in 12 measures of prayer. These measures assess four dimensions of prayer: The social context of prayer, interpersonal aspects of prayer, beliefs about how prayer operates, and the content or focus of prayers. Data from two nationwide surveys of older adults suggest that with respect to all four dimensions, the prayer lives of older whites appear be less developed than the prayer lives of older blacks and older Mexican Americans. In contrast, relatively few differences were found in the prayer lives of older African Americans and older Mexican Americans. The theoretical implications of these findings are discussed.

For over a thousand years, theologians and scholars have maintained that prayer lies at the very heart of religion. For example, Martin Luther argued that faith is A … prayer and nothing but prayer …@ (as reported by Heiler, 1932, p. xiii). Similarly, John Calvin (1536/2006) maintained that, AThe necessity and utility of this exercise of prayer no words can sufficiently express@ (p. 120). In addition, William James asserted that prayer is A … the very soul and essence of religion@ (James 1902/1997, p. 486). Describing prayer as Areligion in action,@ James believed that prayer is the arena in which the real work of religion is done (James, 1902/1997, p. 486). Like James, George Albert Coe was one of the early psychologists of religion. Writing in 1902, he argued that, APrayer is the heart of religion. When you have told what a man=s prayers are like, you have told what his religion is.@ (Coe, 1902, p. 329). Recent empirical research is consistent with this perspective because it suggests that prayer is the most common form of religious practice. More specifically, findings from a nationwide survey by Barna (2006) suggest that 84 percent of Americans indicate they prayed to God during the week before they were interviewed. In addition, 49 percent of the participants in a survey by Stark (2008) report they pray at least once a day. Although these finding are useful, they reflect the frequency of prayer among all Americans taken together. As a result, potentially important variations in prayer among different demographic groups may be overlooked. There is already some evidence that this may be the case. For example, some studies suggest that people who are presently older tend to pray more often than individuals who are currently younger (Barna, 2002). Moreover, research indicates that older African Americans tend to pray more frequently than older whites (Krause & Chatters, 2005).

The purpose of the current study is to contribute to the literature on prayer in two potentially important ways. First, virtually all the research on prayer has focused on African Americans and whites. An effort is made in the current study to expand the scope of inquiry by assessing prayer among members of three racial/ethnic groups: older whites, older blacks, and older Mexican Americans. It is important to bring older Mexican Americans into the picture because recent Census projections indicate that by the year 2050, older Hispanics will surpass older African Americans to become the second largest racial/ethnic group in the nation (Vincent & Velkoff, 2010). There appears to be only one study in the literature that examines race differences in prayer among multiple racial and ethnic groups (Fitchett et al., 2007). These investigators assessed prayer among middle-aged women who were white, black, Hispanic, Japanese, or Chinese. Unfortunately, only one dimension of prayer was evaluated in this study B the frequency of prayer. This limitation highlights the second way in which the current study may contribute to the literature.

The majority of investigators have focused solely on the frequency of private prayer (e.g., Taylor, Chatters, & Levin, 2004). Although this is clearly an important aspect of prayer, research reveals that prayer is a vast, multidimensional domain that should be measured in a number of ways (Laird, Snyder, Rapoff, & Green, 2004). For example, some investigators have studied the type of prayer an individual offers, such as petitionary prayer (Poloma & Gallup, 1991), whereas others have examined specific beliefs about how prayer operates, such as the belief that God answers prayers right away (Krause, 2004). The second way in which the current study may contribute to the literature arises from the fact that race/ethnic differences are evaluated in twelve measures of prayer.

Because no one has assessed race/ethnic differences in a large battery of prayer measures, the analyses that follow are descriptive in nature. A descriptive analysis is performed because it is important to describe a phenomenon like prayer before it can be explained. Since descriptive studies are not the most frequently utilized research design, it is important to briefly reflect on the role this type of study plays in the research process. As Friis and Sellers (2009) point out, research designs can be partitioned into two broad categories. The first involves descriptive studies. This type of research is designed to characterize the amount and distribution of a phenomenon, such as prayer, within a population. In contrast, the second type of research design, the analytic study, is developed to uncover the determinants (i.e., the Acauses@) of a phenomenon, such as the reasons why the frequency of prayer is either higher or lower in different population subgroups. Descriptive studies occupy a key place in the research process because the findings that are obtained from this type of research can ultimately be used to frame hypotheses that can be tested in analytic studies (Friedman 1994). It is important to emphasize that because descriptive analyses are provided in this study, specific hypotheses about race/ethnic differences in each dimension of prayer will not be proposed or tested. This practice is consistent with the views of Koepsell and Weiss (2003), who argue that, AThe hallmark of a descriptive study is that it is undertaken without a specific hypothesis@ (p. 95).

There are four closely-related reasons why descriptive research on prayer speaks directly to issues that lie at the core of psychology. First, a vast literature suggests that stress exerts a noxious effect on physical and mental health (Cartwright & Cooper, 2011). Consequently, psychologists have launched a wide range of programs to help people manage the undesirable events they encounter (Randall, 2010). Prayer is often included among these stress management techniques (Pargament, 1997). Moreover, there is some evidence that there are race/ethnic differences in the use of prayer to cope with pernicious stressors. For example, a recent study by Chatters and her colleagues indicates that African Americans are more likely than Non-Hispanic whites to rely on prayer when dealing with stressful situations in general (Chatters et al., 2004). Similarly, research reveals that African American women are more likely to use prayer as a coping strategy when they encounter spouse abuse (El-Khoury et al., 2004). Unfortunately, very little of this research has involved members of other racial or ethnic groups. This is especially true of Mexican Americans.

Second, a number of researchers have discussed the use of prayer in formal counseling sessions. For example, Beach and his colleagues explore a number of ways in which prayer can be integrated into counseling sessions that are designed to help people cope with marital difficulty (Beach et al., 2008). However, these researchers do not explore the ways in which the various types and dimensions of prayer may operate in this context.

Third, there is rapidly growing interest in the effects of racial and ethnic discrimination on health (Williams & Mohammed, 2009). Moreover, a number of researchers have argued that prayer may be an especially effective way of helping minority group members cope more effectively with these racial problems (Hunter & Lewis, 2010). Even so, there do not appear to be any studies that have empirically evaluated the extent to which prayer offsets the deleterious effects of racism on health and well-being.

Fourth, a considerable number of studies suggest that individuals with strong social support systems tend to enjoy better physical and mental health than people who have not developed close relationships with others (Cohen, 2004). This research is noteworthy because several investigators maintain that when people pray for one another, they are in essence exchanging a religiously-oriented type of social support (Ladd & McIntosh, 2008). Once again, it is disappointing to find that there has been very little empirical research that focuses on this dimension of prayer. Moreover, there appears to be only two studies that examine race differences in the frequency of being prayed for by others (Gillum & Griffith, 2010; Krause & Chatters, 2005), and both of these studies focus solely on African Americans and whites.

As this brief discussion reveals, conducting a descriptive study on race/ethnic differences in a large battery of prayer items holds out the promise of providing deeper insight into a range of issues that concern psychologists. However, before turning to these analyses, it is important to address another important issue.

Although descriptive analyses do not involve tests of specific hypotheses, it is still important to present contextual information that promotes a deeper appreciation for the social setting in which members of different racial and ethnic groups pray. This information is provided below by briefly exploring the historical factors that have shaped the way involvement in religion has emerged in the African American and Mexican American communities.

Historical Experiences and Religious Involvement among Older African Americans

Historical influences on the development of the church in the black community were discussed some time ago by Nelsen and Nelsen (1975). As these investigators show, the church was at the center of the African American community since its inception. Due to centuries of prejudice and discrimination, black people turned to the church for spiritual, social, and material sustenance primarily because it was the only institution in their community that was built, funded, and wholly owned by black people. As a result, the church became a conduit for the delivery of social services, and the first schools for black children were located in them as well. In fact, it is not surprising to find that many of the great political leaders in the black community have strong ties to the church, and many have been members of the clergy (e.g., Martin Luther King, Jr.).

Perhaps no one wrote more on the early history of the church in the black community than W. E. B. Du Bois. The core role that the church played in the lives of African Americans is evident in his work. Writing in 1887, he concluded that, AThe Negro church … provides social intercourse, it provides amusement of various kinds, it serves as a newspaper and intelligence bureau, it supplants the theater, it directs the picnic and excursion, it furnishes the music, it introduces the stranger to the community, it serves as a lyceum, library, and lecture bureau B it is, in fine, the central organ of organized life of the American Negro@ (Du Bois 1887/2000, p. 21). Although the observations of Du Bois were made over a century ago, there is ample evidence that they are still valid today (Mattis & Jaggers, 2001).

Since the church occupies such a central place in the black community, one might expect to find that older blacks adhere strongly to the religious practices that this institution endorses. To the extent that this is true, then perhaps the prayer lives of older African Americans are more fully developed than the prayer lives of older whites.

Historical Experiences and Religious Involvement among Mexican Americans

Like older blacks, older Mexican Americans have also struggled with pernicious historical circumstances. There are two ways in which these difficult historical experiences have shaped the world views of Mexican Americans. First, the conquest of Mexico by the Spanish caused a great deal of pain and suffering for the indigenous people who were living there (Leon 2004). Carrasco (1990) documents the shocking extent of this problem. He reports that in 1500 there were 25 million indigenous people living in Mexico, but due to factors such as disease and slavery, this population was reduced to 1 million by 1600. Given these data, it is not surprising to find that Leon (2004, p. 198) refers to this period of colonization as the AMexican diaspora.@

Second, the deleterious consequences of colonization were exacerbated by a number of subsequent historical events including the Mexican American War of 1848, the Mexican Revolution of 1910, and the great labor shortages during World War I. Each of these events rekindled earlier conditions of subordination and diaspora that were encountered during the Spanish colonization. The vestiges of these historical events are evident in the way contemporaneous scholars describe Mexican American culture. For example, Rodriguez (1994, p. 69), argues that: AWhat makes the experience of Mexican Americans unique compared to other ethnic populations that migrated to this country is their psychohistorical experience and their subsequent subjugation B all taking place in what the indigenous peoples considered to be their own land.@ Rodriguez goes on to point out that physical colonization of Mexican Americans has been accompanied by psychological colonization that fosters A… feelings of inferiority, lack of self-worth, hostility, apathy, (and) apparent indifference … .@

So like blacks, many Mexican Americans turned to their faith for support. But unlike African Americans, Mexican Americans have maintained a somewhat ambivalent view of the church. The wide majority of Mexican Americans (about 73%) affiliate with the Catholic Church (Hernandez et al. (2007). However, Mexican American Catholics were not able to exercise the same degree of control over their church as blacks. For example, as Fernandez (2007) points out, Catholic missionaries did not ordain Mexican American men into the clergy. Consequently, until the early twentieth century, most of the clergy who ministered to Mexican Americans were foreign born. Fernandez (2007) argues that this has created a A… lingering feeling of Mexican Americans toward the official Church is that it is an Anglo institution@ (p. 44). In fact, it is for this reason that evangelic Protestant denominations are growing at a rapid rate in the Mexican American community (Vila, 2005).

Taken as a whole, this brief review suggests that there may be a potentially more complex use of prayer among older Mexican Americans. More specifically, older Mexican Americans may be more likely to engage in prayer than older whites when prayer is undertaken in private. But in contrast, older Mexican Americans may be less likely than older whites to engage in prayer that takes place in formal prayer groups that are supported by the church.

As the discussion that is provided up to this point reveals, prayer is an essential part of a religious life. Moreover, prayer is linked to a number of constructs that are of interest to many psychologists. Even so, significant gaps exist in the literature. An effort is made in the current study to address these limitations by assessing prayer in a wider range of racial and ethnic groups with a more comprehensive battery of prayer measures. The data as well as the measures that were used to accomplish this task are presented in the following section.

Methods

Samples

The data for this study come from two nationwide surveys. Older blacks and older whites participated in the first survey, whereas the second survey focused only on older Mexican Americans. For convenience, the first survey is referred to below as the RAH Survey (i.e., Religion, Aging, and Health Survey) and the second is called the RAHMA (i.e., Religion, Aging, and Health B Mexican American Survey).

RAH survey data

The population for the RAH Survey was defined as all household residents who were either black or white, noninstitutionalized, English-speaking, and 66 years of age or older. 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). In addition, the study population was restricted to individuals who were currently practicing Christians, people who were Christian in the past but no longer practice any religion, and individuals who were never affiliated with any faith at any point in their lifetime. People who practice a religion other than Christianity (e.g., Jews or Muslims) were excluded because it would be difficult to devise a set of religious support measures suitable for persons of all faiths.

The sampling frame consisted of all eligible persons contained in the files that are maintained by the Center for Medicare and Medicaid Services (CMS). It should be emphasized that people are included in this list even if they are not receiving Social Security benefits. A five-step process was used to draw the sample from the CMS data base (see Krause, 2002a) for a detailed discussion of these procedures).

Interviewing for the first wave of the RAH Survey began in March 2001, and concluded in August 2001. The data collection was performed by Harris Interactive (New York). Four waves of interviews have been conducted so far. However, because the goal of the current study is to describe church-based social relationships, only the data from the first wave of interviews are assessed below. A total of 1,500 Wave 1 interviews were completed face-to-face in the homes of the older study participants. Older blacks were over-sampled so sufficient statistical power would be available to examine race differences in religion. Consequently, the sample consisted of 748 older whites and 752 older blacks. The overall response rate for the study was 62%.

RAHMA survey data

The population for the RAHMA Survey was defined as all Mexican Americans aged 66 and over who were retired (i.e., not working for pay), not institutionalized, and who speak either English or Spanish. The sampling frame consisted of all eligible study participants who resided in the following five-state area: Texas, Colorado, New Mexico, Arizona, and California. The sampling strategy that was used for the widely-cited Hispanic Established Population for Epidemiological Study (HEPESE) was adopted for the current study (see Markides, 2003 for a detailed discussion of the steps that were followed). All interviews were conducted by Harris Interactive (New York). The interviews were administered face-to-face in the homes of the older study participants. All interviewers were bilingual and study participants were offered the choice of being interviewed in either English or Spanish. The wide majority of interviews (84%) were conducted entirely in Spanish. Interviewing took place in 2009-2010. A total of 1,005 interviews were completed successfully. The response rate was 52%.

Table 1 contains some basic descriptive data on the older blacks, older whites, and older Mexican Americans who participated in these surveys. Perhaps the most notable difference across the three groups involves the larger proportion of Mexican Americans who affiliate with the Catholic Church and the lower level of educational attainment among Mexican Americans.

Table 1.

Preliminary Descriptive Data

	Whitea	Blackb	Mexican Americanc
Aged	75.4 (SD = 6.8)	74.9 (SD = 6.5)	74.0 (SD = 6.6)
Percent male	.41	.36	.43
Percent currently married	.59	.59	.57
Educatione	12.5 (SD = 3.2)	10.0 (SD = 3.4)	6.8 (SD = 4.0)
Percent Catholic	.33	.07	.79

^aDue to item non-response, the number of cases varies from 726 to 754.

^bDue to item non-response, the number of cases varies from 712 to 734.

^cDue to item non-response, the number of cases varies from 890 to 1005.

^dThe measure of age is scored continuously in years.

^eThe measure of education reflects the total number of years of schooling that were completed successfully by study participants.

Measures

The measures of prayer that are used in the analyses presented below were developed as part of a larger effort to create a comprehensive set of indicators to measure religion in late life. A detailed discussion of the item-development program in the RAH Survey is provided by Krause (2002b). A shorter item-development strategy was followed in the RAHMA Survey. Part of the item-development work for the two surveys involved conducting a series of individual in-depth interviews to capture the dimensions of religion that are most salient for older people. A total of 12 indicators were developed from these qualitative data to assess the prayer life of older people. Some items were taken from the literature, whereas others were developed from scratch. These measures were designed to capture four key dimensions of prayer. These dimensions, as well as the items that were used to measure them, are listed in Table 2. Information on the coding of these items is provided in the footnotes of this table.

Table 2.

Prayer Measures

1. The Social Context of Prayer

A. How often do you pray by yourself?a

B. How often do you participate in prayer groups that are not part of regular worship    services or Bible study groups?b

2. Interpersonal Aspects of Prayer

A. When you are by yourself, how often do you pray for other people?c

B. How often do you think other people pray for you?d

3. Beliefs About How Prayer Operates

A. How often are your prayers answered?c

B. Learning to wait for God=s answer to my prayer is an important part of my faith.e

C. When I pray, God does not always give me what I ask for because only He knows what is    best.e

4. Substantive Content of Prayerc

A. When you=re by yourself, how often do you pray that God=s will be done?

B. When you=re by yourself, how often do you offer prayers of thanksgiving?

C. When you’re by yourself, how often do you pray for guidance?

D. When you<re by yourself, how often do you pray for your own health?

E. When you=re by yourself, how often do you pray for material things, like a job, money,    or a car?

^aThis variable is scored in the following manner (coding in parentheses): never (1), less than once a month (2), once a month (3), a few times a month (4), once a week (5), a few times a week (6), once a day (7), several times a day (8).

^bThis variable is scored in the following manner: 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), 2-3 times a month (6), nearly every week (7), every week (8), several times a week (9).

^cThese variables are scored in the following manner: never (1), once in a while (2), fairly often (3), very often (4). These questions were not administered to study participants who report they never pray when they are alone.

^dThis variable is scored in the following manner: never (1), once in a while (2), fairly often (3), very often (4).

^eThese variables are scored in the following manner: strongly disagree (1), disagree (2), agree (3), strongly agree (4). These questions were not administered to study participants who report they never pray when they are alone as well as respondents who do not believe that their prayers are answered. .

Social context of prayer

This dimension of prayer, which is measured with two indicators, assesses the social location in which prayer may take place. The first measures how often study participants pray when they are alone. In contrast, a second item was administered to find out how often study participants attend formal prayer groups at church that are not part of regular worship services. A high score on either indicator denotes more frequent prayer.

Interpersonal aspects of prayer

Research by Krause and Chatters (2005) suggests that older people frequently pray for others. Often, these prayers involve requests for the well-being of family members. In fact, research by Gallup and Lindsay (1999) indicates that 98 percent of Americans indicate they have prayed for family members. Krause and Chatters (2005) also report that many older adults believe that others pray for them often, as well. Taken together, the two prayer practices may be thought of as reflecting interpersonal aspects of prayer. Two items were developed to assess these prayer practices. The first asks study participants how often they pray for other people. This question was not administered to study participants who said they never pray when they are alone (N = 106 across all race/ethnic groups). The second measure assesses how often older study participants believe that people are praying for them. This item was administered to all study participants regardless of whether they pray themselves. A high score on either item represents more frequent prayer.

Beliefs about how prayer operates

Research by Krause (2004) suggests that when people pray, they expect certain outcomes B certain responses from God. Based on the qualitative item-development work that was discussed above, three measures were devised to capture beliefs about how prayer operates. These questions were not administered to study participants who report they never pray. The first asks study participants how often they feel their prayers have been answered. A high score on this item suggests that study participants believe their prayers are answered often.

The second item dealing with beliefs about prayers has to do with views on the timing of answers to prayers. This question was not administered to study participants who believe their prayers are never answered (N = 53 across all three race/ethnic groups). As Krause, Chatters, Meltzer, and Morgan (2000) report, some older people believe that when they pray, they typically receive an answer fairly quickly. In contrast, other older adults indicate that answers to prayers come when they are needed the most B not too early or too late. This view of prayer is based on the belief that God answers prayers when He feels it is best, even though an individual may not always initially understand His reason for doing so. Ultimately, however, the timing of the response to prayer is perceived to be in the best interest of the person who offered the prayer. A single item was developed by Krause (2004) to assess this belief about prayer. A high score suggests that older study participants strongly believe that learning to wait for God=s answer to their prayers is an important part of their faith.

Research by Krause (2004) further indicates that some older people believe that when they pray, they usually get exactly what they pray for. However, other study participants indicated that at times, they did not get precisely what they asked for. Instead, they found that God answered their prayers in a different way. However, these respondents went on to point out that, with time, they found the answer they received was what they really needed the most. A third item was written to capture this prayer belief. Once again, this question was not administered to older people who believe their prayers are never answered. A high score denotes older adults who strongly believe that God does not always give them what they ask for.

Substantive content of prayer

The fourth dimension of prayer has to do with substantive focus of prayers. Five items were developed to determine how often older people pray that God=s will be done, how often they offer prayers of thanksgiving, how often they pray for guidance, how often they pray for their own health, and how often they pray for material things. A high score on these indicators denotes that each specific type of prayer was offered more often. The questions that deal with the substantive content of prayer were not administered to study participants who report they never pray.

Data Analysis Strategy

The goal of the current study is to see whether differences emerge between older whites, older blacks, and older Mexican Americans in the four dimensions of prayer that were discussed above. This was accomplished by estimating a series of univariate analysis of covariance (ANCOVA) and multivariate analysis of covariance (MANCOVA) models. The choice between ANCOVA and MANCOVA models rested largely on skip patterns in the data. For example, questions on the content of prayers were only asked if study participants indicate that they pray in the first place. So, in order to retain the maximum number of cases for the analysis of the frequency of prayer measure, separate ANCOVA models were estimate for each of these variables. Because the RAH and RAHMA studies are not based on an experimental design, it is important to statistically control for well-known correlates of prayer as well as demographic differences across the three racial or ethnic groups.

The following covariates were used throughout the analyses presented below: Age (scored in years), sex (1 = men, 0 = women), marital status (1 = married; 0 = otherwise), and whether study participants affiliate with the Catholic Church (1 = Catholic; 0 = otherwise).¹ These covariates were selected for the following reasons. Research that involves people age 65 and over indicates that the frequency of prayer is higher among the old-old than the young-old (Braam et al., 2007). A considerable number of studies also reveal that women pray more often than men (e.g., Barna, 2006). Marital status was included in the list of covariates because research indicates that individuals who are currently married tend to pray more frequently than people who are not currently married (Roff et al., 2007). Finally, a covariate that contrasts Protestants and Catholics was included in the study models because research reveals that Protestants tend to pray more often when they are alone than Catholics (O’Connor, Hoge, & Alexander, 2002).

Results

The findings from the current study are presented below in four sections. These sections correspond to the four dimensions of prayer that were identified earlier. The findings involving each dimension of prayer are provided in Table 3.

Table 3.

Prayer Life of Older Whites, Older Blacks, and Older Mexican Americans

	Blacka	Whiteb	Mexican Americanc	Black versus White	Black versus Mexican American	White versus Mexican American
PANEL A. Social context of prayer
Private prayer	7.37d (7.41)e 1.08g	6.36 (6.36) 2.08	6.97 (6.95) 1.58	p < .001g	p < .001	p < .001
Prayer groups	2.83 (3.19) 2.79	2.15 (2.23) 2.32	2.26 (1.96) 2.20	p < .001	p < .001	N.S.
PANEL B. Interpersonal prayer
Pray for others	6.90 (7.02) 1.43	6.10 (6.13) 1.98	6.95 (6.85) 1.43	p < .001	N.S.	p < .001
Others pray for you	3.23 (3.29) .78	2.76 (2.79) .88	2.96 (2.90) .89	p < .001	p < .001	p < .001
PANEL C. How prayers operate
Prayers answered	3.06 (3.12) .81	2.69 (2.72) .83	3.25 (3.19) .82	p < .001	p < .001	p < .001
Wait for God=s answer	3.34 (3.39) .54	3.11 (3.12) .60	3.33 (3.29) .60	p < .001	N.S.	p < .001
God gives what is best	3.44 (3.46) .56	3.23 (3.24) .57	3.65 (3.64) .52	p < .001	p < .001	p < .001
PANEL D. Prayer content
God=s will	3.58 (3.64) .62	3.08 (3.10) .93	3.65 (3.59) .73	p < .001	N.S.	p < .001
Thanksgiving	3.60 (3.63) .64	3.16 (3.18) .85	3.78 (3.75) .54	p < .001	p < .001	p < .001
Guidance	3.50 (3.61) .65	3.08 (3.11) .91	3.43 (3.32) 1.00	p < .001	N.S.	p < .001
Own health	3.59 (3.63) .68	3.02 (3.04) .97	3.69 (3.66) .57	p < .001	N.S.	p < .001
Material things	1.77 (1.80) .97	1.51 (1.51) .78	2.03 (2.01) 1.11	p < .001	p < .001	p < .001

^aThe number of cases for older blacks varies from 509 to 690 depending upon missing data and skip patterns.

^bThe number of cases for older whites varies from 525 to 697 depending upon missing data and skip patterns.

^cThe number of cases for older Mexican Americans varies from 858 to 1003 depending upon missing data and skip patterns.

^dMean adjusted for age, sex, marital status, and affiliation with the Catholic church.

^eUnadjusted mean.

^fSignificance tests for adjusted means.

^gUnadjusted standard deviation.

Social Context of Prayer

A multivariate analysis of covariance (MANCOVA) model was used to test for race/ethnic differences in the indicators that assess the frequency of private prayer as well as the frequency of participation in prayer groups. A MANCOVA model evaluates race/ethnic differences in the two measures simultaneously, thereby taking the correlation between the two prayer indicators into account. Preliminary analysis revealed that this type of model is justified because the correlation between the two measures is .210 (p < .01). The findings from the MANCOVA model are presented in Table 3 (see Panel A). This table contains the unadjusted means and unadjusted standard deviations for each racial group as well as the means that have been adjusted for the covariates that were discussed above (i.e., age, sex, marital status, and affiliation with the Catholic Church).

The multivariate tests of significance (e.g., Pillai=s Trace = .060; F = 36.926; hypothesis df = 4; error df = 4768; p < .001 B not shown in Table 3) suggest that statistically significant race/ethnic differences are present in either one or both of the measures that assess the social context of prayer. This conclusion is confirmed by the univariate tests of significance, which indicate that statistically significant race/ethnic differences are present in both indicators. A series of simple contrasts was performed to determine where race/ethnic differences in the social context of prayer arise. The data in Table 3 (see Panel A) suggest that, compared to older whites (adjusted M = 6.364) and older Mexican Americans (adjusted M = 6.971), older blacks pray more frequently when they are alone (adjusted M = 7.373). The differences in both contrasts were significant at the .001 level. The results further indicate that older Mexican Americans pray more frequently than older whites when they are alone (p < .001).

The findings in Panel A (Table 3) further reveal that compared to older whites (adjusted M = 2.154) and older Mexican Americans (adjusted M = 2.255), older African Americans attend formal prayer groups more often (adjusted M = 2.832). The differences in these contrasts are significant at the .001 level. However, the results further indicate that the difference between older whites and older Mexican Americans is not significant at the .05 level.

The data in Table 3 are helpful because they also provide some insight into how often the two social settings in which prayers are offered vary within each racial/ethnic group. Looking down the columns in Panel A (Table 3) reveals that preferred social setting for prayer is the same for the members of each racial/ethnic group B all older people regardless of their race/ethnicity pray more often when they are alone. Moreover, in each instance, the difference between praying alone and participating in a pray group is striking.

Interpersonal Aspects of Prayer

A separate univariate analysis of covariance (ANCOVA) was performed on each of the items that assess interpersonal aspects of prayer. This was done for the following reason. Only older adults who say they pray were asked whether they pray for other people. But in contrast, all study participants were asked whether others pray for them. A decision was made to administer this question to everyone in the study because other people may pray for a study participant even though the respondent does not pray themselves. The results from both ANCOVA models are presented in Panel B of Table 3.

The findings from the first ANCOVA model suggest that statistically significant race/ethnic differences in the frequency of praying for others are present in the data (sum of squares = 310.251; df = 2; mean square = 155.126; F = 64.512; p < .001 B not shown in Table 3). A series of simple contrasts revealed that the older blacks (adjusted M = 6.901) as well as older Mexican Americans (adjusted M = 6.951) pray for others more often than older whites (adjusted M = 6.099). The difference between these means is significant at the .001. However, the results further indicate that the difference in the frequency of praying for others was not statistically significant at the .05 level for older blacks and older Mexican Americans.

The findings that emerged from the second ANCOVA model in Panel B (Table 3) also suggest that significant race/ethnic differences are present in the item that assesses how often study participants feel that others pray for them (sum of squares = 326.273; df = 2; mean square = 163.136; F = 226.926; p < .001). A series of simple contrasts reveals a clear rank ordering among the three races: Older blacks are the most likely to believe that others pray for them (adjusted M = 3.228), followed by older Mexican Americans (adjusted M = 2.956) and older whites (adjusted M = 2.757), respectively. The differences in the means of the three racial groups were all significant at the .001 level.

Preliminary analysis revealed that a large number of study participants reported that they did not know if others pray for them (N = 568 in the total pooled sample). Researchers typically view Adon=t know@ responses as missing data (i.e., item non-response). However, in this instance, a Adon=t know@ response may be more meaningful. There is only one way for people to know if social network members are praying for them B the significant other must tell them they are doing so. As discussed earlier, when a significant other tells an older person they are praying for them, the person offering the prayer is, in essence, providing social support. Viewed in this way, a Adon=t know@ response serves as a indicator of the amount of interaction and the degree of cohesion in a social network.

An additional set of analyses (not shown in Table 3) was performed to see if there are race/ethnic differences in the extent to which study participants report they Adon=t know@ if others pray for them. A simple binary variable was created that contrasted a Adon=t know@ response (scored 1) with responses that reveal how often significant others are praying for a study participant (scored 0). Then, using logistic regression, this binary outcome was regressed on the covariates that have been used throughout this study, as well as two dummy variables. The first dummy variable contrasts older blacks (scored 1) with older Mexican Americans (scored 0) and the second contrasts older whites (scored 1) with older Mexican Americans (scored 0). The findings reveal that compared to older Mexican Americans, older blacks are more likely to report they Adon=t know@ if others are praying for them (b = .949; p < .001; odds ratio = 2.584). Similarly, the data indicate that compared to older Mexican Americans, older whites are also more likely to indicate they Adon=t know@ if others are praying for them (b = .844; p < .001; odds ratio = 2.325). These analyses compare blacks with Mexican Americans as well as whites with Mexican Americans, but tests cannot be performed in the same model to see if older blacks differ significantly from older whites. The logistic regression analysis was performed a second time after older whites served as the reference category instead of older Mexican Americans. The data indicate that compared to older whites, older blacks are not more likely to provide a Adon=t know= response when asked if others pray for them (b = .071; n.s.; odds ratio = 1.074).

When coupled with the data in Table 3, the results that emerged from the analysis of Adon=t know@ responses indicate that even though older Mexican Americans report that others pray for them less often than older blacks, older Mexican Americans appear to be more aware than older blacks of the extent to which others are praying for them.

Beliefs about How Prayer Operates

When the survey for this study was administered, respondents were first asked if they believe their prayers are answered. Then, if they believe their prayers are answered, they were asked the two questions about the way they believe God answers prayers. Because of this skip pattern, an ANCOVA model was estimated to assess race/ethnic differences in the extent to which study participants feel their prayers are answered and a MANCOVA model was estimated to examine race/ethnic differences in the two items that assess beliefs about how God answers prayers. The results of both analyses are provided in Panel C (Table 3).

The data suggest that there are statistically significant race/ethnic differences in the extent to which older study participants feel their prayers are answered (sum of squares = 98.910; df = 2; mean square = 49.455; F = 75.180; p < .001). Once again, a set of simple contrasts was conducted to see where significant race/ethnic differences arise. The findings reveal that older Mexican Americans (adjusted M = 3.247) are more likely than either older blacks (adjusted M = 3.055) and older whites (adjusted M = 2.692) to believe their prayers are answered. The differences between both estimates are significant at the .001 level. The results further indicate that older blacks are more likely than older whites to feel that God answers their prayers (p < .001).

The MANCOVA model was estimated in order to see if there are race/ethnic differences in the extent to which study participants believe it is important to wait for God to answer their prayers, and if they believe God does not always give them exactly what they ask for. Preliminary analysis suggests that it is appropriate to assess the two items simultaneously because they are significantly correlated (r = .170; p < .01). The multivariate tests of significance (e.g., Pillai=s Trace = .085; F = 48.859; df = 4; error df = 4380.0; p < .001 B not shown in Table 3) indicate that there are race/ethnic differences in either one or both of the items assessing beliefs about how prayer operates. Subsequent univariate tests reveal that race/ethnic differences arise in both measures. A set of simple contrasts was performed to see where the race/ethnic differences arise. The data suggest that older blacks (adjusted M = 3.342) are more likely than older whites (adjusted M = 3.109) to believe that God does not always answer a prayer right away. The difference between these estimates is significant at the .001 level. However, the findings further reveal that older blacks are not more likely than older Mexican Americans (adjusted M = 3.326) to believe it is important to wait for God=s answer to a prayer. A third simple contrast suggests that older Mexican Americans are more likely than older whites to believe it is important to wait for God=s answer to a prayer. The difference between these estimates is significant at the .001 level.

The data in Panel C (Table 3) also suggest that older Mexican Americans are more likely than older African Americans to believe that God answers prayers in the way He feels is best (adjusted mean for Mexican Americans = 3.653; adjusted mean for blacks = 3.444; p < .001). The results further indicate that older African Americans are, in turn, more likely than older whites to feel that God answers prayers in the way that He feels is best (adjusted mean for blacks = 3.444; adjusted mean for whites = 3.231; p < .001).

Substantive Content of Prayer

Recall that five items were administered to assess the content or type of prayers that the older people in this study offer. The average correlation among the five items is .385. This is important for two reasons. First, it shows that it is appropriate to analyze these indicators simultaneously in a single MANCOVA model. Second, the data suggest that the typical older person tends to offer prayers that involve more than one type of content.

The results from the MANCOVA analysis of prayer content are provided in Panel D (Table 3). The multivariate test of significance suggests that there are race/ethnic differences in one or more of the prayer content items (e.g., Pillai=s Trace = .193; F = 47.752; hypothesis df = 10.0; error df = 4474.0; p < .001). The univariate tests of significance reveal that race/ethnic differences emerge in each of the five prayer content measures.

A series of simple contrasts reveals that race/ethnic differences in the content of prayers roughly mirror the race/ethnic differences that emerged in the other prayer dimensions. More specifically, the findings reveal that older whites are less likely than either older blacks or older Mexican Americans to pray that God=s will be done, offer prayers of thanksgiving, pray for guidance, pray for their own health, or pray for material things. The differences between older whites and the members of the two minority groups are all significant at the .001 level.

The pattern of differences in prayer content is somewhat more complicated for older blacks and older Mexican Americans. The means for members of the two racial groups do not differ significantly for prayers that God=s will be done, prayers for guidance, and prayers for health. However, when differences emerge between older blacks and older Mexican Americans, the means for older Mexican Americans are consistently higher. More specifically, the data suggest that older Mexican Americans are more likely to offer prayers of thanksgiving and to ask for material things than older blacks. These differences are significant at the .001 level.

The data in Table 3 are also useful because they provide a sense of how often each type of prayer is offered within each racial group. Looking down the columns in Panel D (Table 3) reveals that the most frequent type of prayer is the same for the members of each racial group B prayers of thanksgiving. In contrast, the means for each racial group are consistently the lowest for the measure that assesses prayers for material things.

Conclusions

Research on prayer has come a long way since the early psychologists of religion first argued that prayer should be a major focal point of academic studies on religion. For example, researchers have learned that prayer is a complex construct in its own right that should be measured in a number of ways (Laird et al., 2004). In addition, a range of theoretical perspectives have been developed to explain how prayer might influence a wide array of outcomes, including physical health status (Breslin & Lewis, 2008). But it is especially important for the purposes of the current study to note that researchers have discovered that there may also be important demographic variations in the prayer lives of individuals in our society. More specifically, research indicates that the prayer lives of older African Americans appear to be more developed than the prayer lives of older whites (Krause & Chatters, 2005; Taylor, Chatters, & Levin, 2004). Although this work may have moved the field forward, there is still a lot to learn about prayer. The purpose of the current study was to contribute to the evolving literature on prayer by assessing whether differences emerge in 12 measures of prayer among older whites, older blacks, and older Mexican Americans. This research is unique because there do not appear to be any studies in the literature that empirically evaluate differences in an array of prayer measures across three racial/ethnic groups simultaneously.

Summarizing the findings from this study in a succinct manner is challenging because the scope of the analyses is broad. Even so, it is possible to discern three potentially important trends in the data. Consistent with the aims of descriptive analysis, these trends are identified below and some thoughts are offered about the theoretical processes that may be driving them. Hopefully, these theoretical speculations will serve as a point of departure for developing more precise hypotheses that can be evaluated empirically.

First, and perhaps most important, the level of every dimension of prayer is lower for older whites than for either older blacks or older Mexican Americans. Simply put, the prayer lives of older whites do not appear to be as well developed as the prayer lives of their minority group counterparts. The sheer consistency of these findings is noteworthy and suggests that some broad underlying theoretical process may be at work. Useful insight into the nature of this underlying process may be found in Schieman=s (2008) research on the deprivation-compensation hypothesis. According to this view A… socioeconomically disadvantaged individuals derive the most psychological benefits from religion@ (Schieman, 2008, p. 711). Research consistently shows that older blacks and older Mexican Americans are more socioeconomically challenged than older whites (Crimmons, Hayward, & Seeman, 2004). If these older minority group members suffer from greater socioeconomic problems, then it follows from the deprivation-compensation hypothesis that they should derive more benefits from religion than older whites. And if older minority group members obtain more benefits from religion, then the extent of their religious practices and the depth of their religious beliefs should be more pronounced than the religious beliefs and practices of older whites. Because prayer is thought to lie at the heart of religion (Coe, 1902), evidence of the deprivation-compensation hypothesis should be especially evident with respect to this particular religious practice. And this is what the data show.

Second, further support for the deprivation-compensation hypothesis is provided by the data on race/ethnic differences in the frequency of praying for material things. Because older blacks and older Mexican Americans are more economically challenged than older whites, the deprivation-compensation hypothesis would predict that rates of prayer for material things should be greater for these older minority group members than for older whites. The data are consistent with this view.

Third, differences also arise in the prayer lives of older blacks and older Mexican Americans. On the one hand, the data suggest that older blacks pray more frequently when they are alone and attend formal prayer groups more often than older Mexican Americans. But in contrast, Mexican Americans are more likely to offer prayers that have a wider range of content than older blacks. Although older blacks and older Mexican Americans are just as likely to pray that God=s will be done, pray for guidance, and pray for their own health, the data indicate that older Mexican Americans pray more often for material things and they are more likely to offer prayers of thanksgiving than older blacks.

The fact that older Mexican Americans are more likely than older blacks to pray for material things merits further comment. Because both older blacks and older Mexican Americans are economically challenged, the greater frequency of prayers for material things among older Mexican Americans must be driven by something other than just the deprivation-compensation hypothesis. Greater insight into this issue may be found by turning to a culturally unique aspect of religion in the Mexican American community. As Elizondo (2000) maintains, AThere is no greater and more persistent symbol of Mexican and Mexican American identity than devotion to Our Lady of Guadalupe@ (p. 128). Many Mexican Americans believe that in 1531, the mother of Jesus (i.e., Our Lady of Guadalupe) appeared to a Mexican peasant to assure him that she was well aware of the suffering of Mexican people and that she was there to help them. Several miracles were performed to establish her authenticity and to convince the Mexican people of her intentions. Since then, Our Lady of Guadalupe has been viewed as the unique protector and source of resilience for all people of Mexican heritage. In her detailed study of Our Lady of Guadalupe, Rodriquez (1994) reports that Mexican Americans are especially likely to petition her for material things: AOur Lady of Guadalupe takes a central role regarding the vital necessities of life B food, shelter, safety, and concern for family. She is petitioned for everything from health to the protection of a family owned business@ (p. 148). If the insights provided by Rodriguez (1994) are valid, then more frequent prayers for material things among older Mexican Americans cannot be explained solely by the deprivation-compensation hypothesis. Instead, a more complete understanding of prayers for material things among older Mexican Americans can be found by focusing on the intersection of culturally unique aspects of religion and the deprivation-compensation hypothesis.

It is important to keep a statistical issue in mind when reviewing the statistically significant differences in the data. The samples that are used in the analyses presented above are large and as a result, statistically significant findings are likely to emerge in instances where the differences between groups are not very large. One way to deal with this situation is to focus solely on group differences that are significant at the .001 level. However, this strategy is not helpful in the current study because all relationships that are statistically significant are significant at the .001 level. Another strategy is to focus on the instances in which the absolute difference between groups is greatest. Since the most consistent differences occur between older whites and older blacks, it might be useful to examine the magnitude of the differences between older people in these two racial groups. The three biggest differences that emerge in this instance involve the greater participation of older blacks in prayer groups (the score for older blacks is 24% larger than the score for older whites), prayers for one’s own health (16%), and prayers for material things (15%). These differences could be generated by two factors. First, greater participation in prayer groups could reflect the greater role that the church plays in the black community. Second, frequent prayers for health as well as more frequent prayers for material things may be driven by socioeconomic factors.

Although the findings from this study may have contributed to the literature, researchers are encouraged to keep the limitations of this work in mind as they review the results. Two shortcomings are discussed briefly below.

First, even though 12 measures of prayer were examined in this study, several other dimensions of prayer were not assessed in the survey. For example, prayers are an integral part of most formal worship services, but formal group prayers during church services were not measured in this study. Similarly, Poloma and Pendleton (1989) have devised measures of subjective prayer experiences (e.g., feeling the presence of God during prayer). No effort was made to include measures of these subjective prayer experiences in the current study. Consequently, researchers would benefit by expanding the scope of research on race/ethnic differences in prayer to include these, as well as other dimensions of prayer in their work.

Second, a central premise in this study is that prayer may vary across members of different social demographic groups, such as race and ethnicity. But race and ethnicity are not the only potential source of demographic variation in prayer. For example, there is some evidence that women have more well-developed prayer lives than men (Levin, Taylor, & Chatters, 1994). No effort was made to assess gender differences in prayer in the current study because simultaneously assessing variations in 12 measures of prayer by race/ethnicity and by gender across three racial groups would have been complex, cumbersome, and difficult to present in a succinct and coherent fashion. Even so, research is needed to explore variations in the prayer life of members of other demographic groups, including gender.

Although there clearly are shortcomings in the work that has been presented here, the issues that have been raised may encourage others to delve more deeply into this core religious practice. The call for further work is justified by Spilka (2005), who sadly observes that A… it is shocking to realize how much psychological speculation has been devoted to prayer over the last century and how little empirical work has been undertaken in this area@ (p. 372).