7. Religion and Family Formation

By Tim B. Heaton and Kristen L. Goodman

Tim B. Heaton and Kristen L. Goodman, “Religion and Family Formation,” in Religion, Mental Health, and the Latter-day Saints, ed. Daniel K. Judd (Provo, UT: Religious Studies Center, Brigham Young University, 1999), 107–128.

Religion and Family Formation​

Tim B. Heaton and Kristen L. Goodman

 

Tim B. Heaton was a professor of sociology at Brigham Young University and Kristen L. Goodman was a research supervisor at the Department of Correlation Research at the LDS Church Office Building when this was written. This article was originally published in Review of Religious Research 26:343–359; reprinted with permission.

 

Abstract​

This chapter considers the correlation between religion and family formation patterns. The family formation patterns of members of The Church of Jesus Christ of Latter-day Saints, with its strong family-centered theology, were compared with those of Roman Catholics, Protestants (divided into liberal and conservative groups), and people with no religious preference. It was found that people with no religious preference are less likely to marry, more likely to divorce, less likely to remarry after divorce, and tend to have smaller families than members of the varying religious groups.

When compared with the other religious groups, Latter-day Saints have larger families, the highest rates of marriage and fertility, and the lowest divorce rates. Catholics also tend to divorce less and remarry after divorce less frequently than Protestants.

Perhaps no other societal institution has a closer link with religion than does the family (D’Antonio & Aldous, 1983; Hargrove, 1979). Yet recent changes in family formation have challenged Judeo- Christian ethics concerning family life (D’Antonio, 1980). Delayed marriage, rising divorce rates, and reduced family size all run counter to the traditional pronatal, pronuptial stance of most Christian religions. In light of these transformations in the typical family-lifecycle events that the average person can expect to experience, the persistence of the relationship between family and religion remains an important issue. Indeed, the influence of religion appears to be waning (Wilson, 1978). In order to understand the role of religion in society, it will be necessary to continue monitoring the relationship between religious and family-oriented behavior.

This report considers the interrelationship between religion and family by selecting a religion with a unique family-centered theology and comparing patterns of family formation among its membership with patterns for Catholics, Protestants, and persons with no religious preference. To the extent that religious differentials persist, we suggest that even amidst dramatic changes in nationwide patterns of family formation, religion still plays an important role in family life.

Aspects of family formation considered here include family size, marriage, divorce, and remarriage. Religious differentials in family size and contraceptive use have received considerable attention (Bouvier & Rao, 1975; Chamie, 1981; Westoff & Potvin, 1967; Westoff & Rider, 1977). This attention is justified by differing doctrinal views of procreation, as well as by the implications of fertility for church growth and age composition. Marriage, divorce, and remarriage also have profound implications in terms of the composition of religious groups. Delayed marriage and divorce result in larger populations of singles and female-maintained households, and in lower fertility; remarriage makes for more complicated kinship networks. There is considerable evidence that divorce varies by religion (Albrecht, Bahr, & Goodman, 1983; Bumpass & Sweet, 1972; McCarthy, 1979), but marriage and remarriage rates have received little attention. Given the dramatic change in the frequency and timing of these life-cycle events, the linkage between religion and family is incomplete without reference to these aspects of family formation.

Denominational Differences in Family Formation​

The social mechanisms linking religious ideology and familial behavior are varied and complex. The influence of religion begins when parents use religious values in socializing their children. Religious rites mark major events in the life-cycle including puberty, marriage, births of children, and death. Religion regulates premarital sexual behavior, male selection, family size, and marital stability. Religious orthodoxy on issues such as sexual behavior and male authority may also be a source of stress in family relationships (Christensen, 1972). Not only do people use religious teachings as a guide for behavior, but they also select religions that are consistent with their personal preferences. Individuals with unconventional attitudes about family life may reject involvement in organized religion. For example, the divorced or childless may feel out of place in family-oriented churches and decide to drop out. On the other hand, pronatalists may search for a religion where they can feel at ease with a larger family. Thus, the association between religion and family might more accurately be thought of as measuring interdependence between institutions rather than as a causal relationship. The task of sorting out and quantifying these sources of influence lies beyond the scope of this paper. Our major task is to document the overall magnitude of this influence. Absence of differences among religious groups implies that ideological differences have diminished and mechanisms of influence have become ineffective.

Mormon Families. The Church of Jesus Christ of Latter-day Saints (Mormon or LDS) espouses a theology that is both pronuptial and pronatal (B. L. Campbell & E. E. Campbell, 1977; Thomas, 1983). The key tenet in the Mormon theology of the family is that, given the proper circumstances, family relationships (i.e., husband-wife and parent-child) will be perpetuated in heaven. The proper circumstances include belief in Christ, performance of ordinances in a Mormon temple, and observance of a moral code of behavior. Marriage ceremonies performed exclusively within a temple are regarded as “the one and only perfect contract of matrimony” (Talmage, 1968, p. 84). Marriage and parenthood are considered to be duties of all who are physically able and to be a valuable, necessary experience leading to spiritual development (Talmage). Titles appearing in the Church’s official magazine, the Ensign, such as “Fundamentals of Enduring Family Relationships,” “Her Children Arise Up and Call Her Blessed,” “Prepare the Heart of Your Son,” “Bring Out the Best in Marriage,” and “Happy Parents Equal Happy Children” reflect this emphasis on family life. For Mormons, the theological conception of marriage is embodied in marriage ceremonies performed within Mormon temples. Thus, it is appropriate to compare patterns of family formation for those Mormons who choose a temple marriage ceremony with those who do not.

A related element of Mormon doctrine is belief in a “preexistence” where all inhabitants of the earth lived in spirit form before birth. Having children provides the opportunity for the “pre-mortal beings” to participate in earth life, which is essential for the salvation of these eternal beings (Hastings, Reynolds, & Canning, 1972; Spicer & Gustavus, 1974). Consequently, childbearing is stressed as a necessary step in carrying out God’s plan for “pre-mortal” children.

Analysts of religious fertility differentials have consistently found high fertility rates among Mormons (DeHart, 1941; Mineau, Bean, & Skolnick, 1971; Skolnick et al., 1978; Spicer & Gustavus, 1974; Thornton, 1979; Westoff & Potvin, 1967; Yoder, 1980). Their above-average fertility cannot be explained by differences in socioeconomic status (Thornton). Moreover, Yoder finds that the correlation between religious attendance and fertility is stronger for Mormons than for either Catholics or Protestants. Mormons also have conservative sex norms (Christensen, 1976). To date, however, little attention has been given to the influence of religiosity upon other aspects of family formation among Mormons. Existing analyses of Mormons are often based on relatively small, unrepresentative samples (Thomas, 1983). We expect to find Mormon/non-Mormon differences not only in family size, but also in rates of marriage, divorce, and remarriage corresponding to pronuptial elements of Mormon theology.

Catholic-Protestant Differences. Traditionally, Catholics have been characterized by higher fertility rates than Protestants. Although there has been substantial convergence during recent decades (Westoff & Jones, 1979), evidence from the most recent large-scale national surveys indicate that a significant difference still exists (Mosher & Hendershot, 1984). This difference is consistent with the pronatalist and anticontraceptive position taken by the Catholic Church. Catholics also take a more negative view of divorce and remarriage. Until recently, divorced Catholics who remarried were excommunicated, and although the church’s position has moderated, annulment is still necessary before a remarriage can be legitimized (McCarthy, 1979). Corresponding to the Church’s position, Catholics were found to have lower divorce rates and lower rates of remarriage following divorce than white Protestants, despite recent convergence in marriage patterns (Bumpass & Sweet, 1972; McCarthy). Less attention has been given to Catholic- Protestant differences in the proportion ever marrying. Analysis of a Rhode Island sample finds close similarities between Protestants and Catholics in marriage rates, but substantial ethnic variation appeared within the Catholic sample (Kobrin & Goldscheider, 1978).

There is also substantial variation in fertility among Protestant denominations (DeJong, 1965; Marcum, 1981; Yoder, 1980). Conservative Protestants have larger families, on average, than liberal Protestants. DeJong argues that among Protestants fertility issues are not so much a matter of church doctrine but rather of religious orientation. Conservative Protestants may consider contraceptive use as interference with God’s will. Conservative Protestants also take a more traditional position on issues such as abortion, homosexuality, premarital sex, and divorce (Hunter, 1983). A rural southern subcultural tradition of higher fertility may also exist among some conservative Protestants. On the other hand, liberal Protestants may be more apt to view overpopulation as a social problem, and fertility control as a moral responsibility. Thus, there is ample justification for a liberal–conservative Protestant distinction in our analysis of religious differentials in family formation.

Control Variables. Religiosity is not the only factor which may account for interdenominational differentials in family characteristics. Denominations also differ in other characteristics which might account for differences in family formation. For example, the explanation of religious differentials in fertility based on differences in socioeconomic composition has been referred to as the “social characteristics” hypothesis (Chamie, 1981). Here we suggest two approaches for verification of the importance of religion per se. First, if religious commitment makes a difference, then level of involvement should be related to patterns of family formation. Attendance will be used as our measure of involvement: frequent attenders should have the highest rates of marriage and remarriage, the lowest rates of divorce, and (in pronatalist religions) the highest fertility. Attendance is used because it reflects contact with a formal organization and a religious community. Other dimensions of religiosity should also be considered, but detailed analysis of relationships between dimensions of religiosity and family formation lie outside the scope and data limitations of this paper.

Unfortunately, religious involvement at the time of the survey may not accurately reflect involvement at the time of the marriage. This is especially true for the divorced and remarried since divorce and remarriage may dramatically affect religious involvement. Retrospective data on religious involvement are generally not available, and reliability would need to be established before we could have confidence in such measures. For Mormons, however, temple marriage is an appropriate indicator. To qualify for a temple marriage, the prospective bride and groom must be living in accordance with the Church’s code of ethics (this includes items such as attendance at meetings, premarital chastity, and acceptance of Church authorities). Temple marriage implies an acceptance of Mormon theology and a commitment to establish a family unit that will persist in the eternities. If Mormonism is characterized by different patterns of family formation, the differences should be most evident among temple marriages. Therefore, divorce, remarriage, and fertility will be compared for temple and non-temple marriages.

Religious involvement, however, is also associated with social characteristics (Albrecht & Heaton, 1984; Argyle & Beit-Hallahmi, 1975; Davidson, 1977). Thus, comparisons by level of religious involvement cannot completely rule out the social characteristics hypothesis. The second approach will be to statistically control for social characteristics in our comparison of religious differences. Since education generally has as high a positive association with church attendance and denominational affiliation as any other social characteristic, we have selected education as a control variable. To the degree that religious differences in family formation persist after educational differences are taken into account, we gain confidence in the assertion that religion makes a difference in important dimensions of everyday living.

Methods

Data collection for the Mormon survey was initiated in the spring of 1981. In the first stage, questionnaires were mailed to a random sample (n = 7,446) of adults (age eighteen and older) from a computerized list of all members in the United States and Canada. A reminder post card was sent out two weeks later. These two mailings generated a response rate of 54 percent. The second stage involved asking bishops (local ministers) to send someone to interview respondents who had not returned questionnaires. With these personal follow-ups an additional 5 percent return was achieved. A third step was to ask bishops to complete questionnaires for nonresponding members by filling in information they had from personal knowledge and membership records. (Bishops have computerized membership records containing information on sex, age, marital status, marital history, current family size, and children ever born.) This provided information on an additional 15 percent of the sample. Finally, an attempt was made to interview nonrespondents by telephone, yielding an additional 7 percent for a total response rate of 81 percent. From all of these steps combined, only 4 percent of the original sample refused to respond. One percent had died or were no longer members of the LDS Church. The final 14 percent of the sample were unknown to local bishops and unavailable to telephone or mailing approaches. 1

We suspect some bias in the reported frequency of religious participation since those who refused and those who were not located are probably less involved in the LDS Church. Many in this group would probably not identify themselves as Mormons, however, implying comparability of our results with surveys which rely on self-reported preference to establish religious membership. The variables analyzed are observable historical events rather than religious attitudes or opinions. Thus, we believe that response bias due to Church auspices of the survey of patterns of religious involvement are minimal.

For analysis of Catholics, Protestants, and those with no preference, we utilized the National Opinion Research Center’s Cumulative General Social Surveys (1983). NORC provides one of the most widely analyzed data sets that includes information about religious preference, attendance, and family formation. Each year the survey is administered, approximately 1,500 respondents are interviewed. In order to augment the sample size, we have combined samples for different years. Changing patterns of family formation during the 1970s would result in unequal comparisons for the early 1970s, so we have selected the 1978,1980,1982, and 1983 samples to compare with the Mormon survey conducted in 1981. To eliminate the confounding influence of race, only whites are included in the analysis. Methodists, Episcopalians, Presbyterians, and Congregationalists (including the United Church of Christ) are classified as liberal. Conservative Protestant churches include Lutheran, Baptist, Church of God, Church of Christ, Assembly of God, Holiness, Jehovah’s Witness, Nazarene, Pentecostal, and Seventh Day Adventist (Glock & Stark, 1965; Marcum, 1981). 2 Some incomparability may exist due to differences in data collection techniques and wording of questions. For example, the NORC survey asks if the respondent has ever been divorced, whereas the Mormon survey asks if the first marriage and/or the most recent marriage ended in divorce. Overall, however, similarities in questions and correspondence with national data suggest that the data sets are suitable for comparative purposes.

Family formation variables to be analyzed include proportion ever married, proportion of those ever married who have divorced, proportion of those divorced who have remarried, and children ever born. When examining the proportion ever married, we restrict our sample to those over age thirty to allow for differences in the timing of marriage. Various types of statistical analysis will be performed. First, we wish to compare religious groups. Mormons, Catholics, conservative Protestants, liberal Protestants, and those with no religious preference will be ranked on each variable, and differences between groups will be evaluated statistically. Second, the differences between frequent and infrequent attenders will be explored, giving attention both to the overall difference between religious groups and to the attendance-religion interactions. Third, education will be introduced as a control factor in our examination of religious differentials in aspects of family formation. Many of the family formation variables are proportions most appropriately analyzed with nonparametric statistics such as log linear models (Fienberg, 1980). For children ever born, we will use analysis of variance.

Results​

Interfaith Comparisons. Comparisons among Catholics, liberal and conservative Protestants, Mormons, and those who reported no religion on selected aspects of family formation are reported in Table 7.1. Nearly everyone eventually marries, but there are differences according to religious preference. Those with no religion are least likely to have married: 81 percent of men and 87 percent of women have married. Among the religious groups, Mormons have the highest rates of marriage, exceeding 97 percent for males and females. Conservative Protestants follow closely with values of 96 percent, and liberal Protestants lag behind slightly with percentages of 93 for men and about 95 for women. Catholics are another step lower with percentages of 89 for men and 91 for women. We have ranked each group from highest to lowest and computed statistical tests for the difference between each pair (see Table 7.2). Not all possible comparisons are made, but adjacent G2 (likelihood ratio chi-square) values are additive. For example, the separate differences between Mormon men, conservative Protestant men, and liberal Protestant men are not significant, but the test for differences among the three groups is significant (G2 = 2.90 + 3.00 = 5.90, df = 2,p = .05). Statistical tests (Table 7.2) indicate some differences among men—the greatest between liberal Protestants, Catholics, and those with no preference. For women, the contrasts are generally not as great, but there is a marked gap between liberal Protestants and Catholics.

 

Table 7.1: Patterns of Family Formation for Catholics, Protestants, and Mormons

 

Sex

Catholics

Liberal Protestants

Conservative Protestants

Mormons

No Religion

% of those over age 30 who have ever married

Male

88.6

93.3

96.1

97.5

81.0

Female

91.2

95.4

96.3

97.2

86.7

% of ever married persons who have ever divorced

Male

19.8

24.4

27.7

27.7

39.2

Female

23.1

30.8

30.9

18.8

44.7

% of ever divorced persons who are currently remarried

Male

49.5

62.9

61.8

66.6

48.4

Female

35.2

50.0

55.1

53.0

37.3

Children ever born to women (first marriage intact, standardized for marital duration)

 

2.38

2.03

2.27

3.31

2.02

Sample Size

Population over age 30

Male

446

374

512

1625

147

Female

616

548

751

1740

83

Ever married persons

Male

469

398

596

2180

158

Female

713

597

981

2458

114

Ever divorced persons

Male

93

97

165

311

62

Female

165

184

272

462

51

Women (first marriage intact)

 

452

312

464

1327

61

 

The high divorce rate characterizing the United States does not appear to be endemic to any particular religion. Those with no religious preference, however, have by far the highest divorce rates: 39 percent of the men and 45 percent of the women have experienced at least one divorce. Mormons have the lowest percentage ever divorced with values of 14 for men and 19 for women. Catholics are next in order with percentages of 20 for men and 23 for women. Divorce among Protestants is substantially higher with percentages of 25 for men and 30 for women. Significance tests indicate little difference between liberal and conservative Protestants, but the other differences among the other religions tested are significant (see Table 7.2).

 

 Table 7.2: Rank, Order, and Statistical Comparison of Religious Groups on Family Formation Variables)

Rank

 

Sex

1

 

2

 

3

 

4

 

5

% of those over age 30 who have ever married

M

Mormon

 

Conserv. P.

 

Liberal P.

 

Catholic

 

None

F

Mormon

 

Conserv. P.

 

Liberal P.

 

Catholic

 

None

% of ever married persons who have ever divorced

M

None

 

Conserv. P.

 

Liberal P.

 

Catholic

 

Mormon

F

None

 

Conserv. P.

 

Liberal P.

 

Catholic

 

Mormon

% of ever divorced persons who are currently remarried

M

Mormon

 

Liberal P.

 

Conserv. P.

 

Catholic

 

None

F

Conserv. P.

 

Mormon

 

Liberal P.

 

None

 

Catholic

Children ever born

F

Mormon

 

Catholic

 

Conserv. P.

 

Liberal P.

 

None

Statistical Test for the Difference Between:a

 

 

1 and 2

 

2 and 3

 

3 and 4

 

4 and 5

 

 

 

 

G2

 

G2

 

G2

 

G2

 

 

% of those over age 30 who have ever married

M

2.51

 

3.41

 

5.59*

 

5.19*

 

 

 

F

1.61

 

.56

 

8.33**

 

1.59

 

 

% of ever married persons who have ever divorced

M

7.65**

 

1.36

 

2.59

 

8.73**

 

 

 

F

8.46**

 

0.00

 

9.78**

 

6.40*

 

 

% of ever divorced persons who are currently remarried

M

.44

 

.03

 

3.70

 

.02

 

 

 

F

.31

 

.48

 

2.63

 

.07

 

 

 

 

F

 

F

 

F

 

F

 

 

Children ever born

F

113.19***

 

1.19

 

4.01*

 

2.18

 

 

*

p < .05

 

 

 

 

 

 

 

 

 

**

p< .01

 

 

 

 

 

 

 

 

 

***

p < .001

 

 

 

 

 

 

 

 

 

a There is one degree of freedom for each test.

 

 

 

 

 

 

 

 

 

 

                                         

 

 A majority of those who divorce eventually remarry. As with marriage, remarriage is most likely among Mormons and Protestants, while Catholics and those with no religion have the lowest percentages remarried following divorce. Moreover, compared with differences in proportions ever married, the percentage differences among religious groups in remarriage are substantially greater, some falling in the range of 15 percentage points. But statistical tests are less likely to be significant because of the smaller sample sizes. The only gap which approaches significance is between Protestants and Catholics or those with no preference.

As with the previous variables considered, fertility also varies across religions. Mormons, with a value of 3.31 children, clearly have the largest families (standardized for marital duration). Catholics ran a distant second with a value of 2.38, and conservative Protestants are close behind. Liberal Protestants and those with no preference have the smallest values of just over 2.00. Statistical tests indicate that the Mormon/non-Mormon difference is by far the most significant, and that conservative Protestants may be higher than liberal Protestants, but other comparisons are not significant.

Religious Attendance. Among those who do state a religious preference, frequent attenders (two or more times per month) are different from infrequent attenders. Table 7.3 shows that among Catholics, infrequent attenders are more likely to marry than frequent attenders. The pattern tends to be reversed among Protestants and Mormons: frequent attenders have slightly higher rates of marriage. The effect of religion on marriage is statistically significant for both men and women, but the effect of attendance and the attendance-religion interaction are statistically significant only for men (see Table 7.4).

Within each religious group, divorce is much lower among frequent attenders than among infrequent attenders, regardless of sex. Statistical tests indicate that differences between the religious groups are clearly significant (Table 7.4). Moreover, attendance has a significant effect on divorce, especially for men. The attendance-religion interaction is not significant. Thus, religious identification is associated with divorce, and frequent attenders are characterized by lower divorce rates than infrequent attenders regardless of religious preference.

With the exception of liberal Protestant women, remarriage is higher among frequent attenders than among infrequent attenders. Moreover, the effect of attendance tends to be greater among Mormons than Catholics or Protestants. The attendance effect is significant for women and especially for men, but the attendance-religion interaction is significant only for women.

Frequency of attendance appears to have very little impact on the fertility of Catholics or Protestants, but among Mormon women frequent attenders have .68 more children on average than infrequent attenders. Even so, low attending Mormons have higher fertility than Protestants or Catholics. Analysis of variance (Table 7.4) indicates that both the attendance effect and the attendance-religion interaction are significant.

In summary, in five of the six comparisons the difference between frequent and infrequent attenders was statically significant. Church attendance is associated with lower rates of nonmarriage and divorce, higher probabilities of remarriage after marital dissolution, and, for Mormons, higher fertility. Nevertheless, introducing attendance as an additional variable generally does not result in the disappearance of religious differences. These findings support the conclusion that not only religious identification, but also formal religious involvement are associated with patterns of family formation.

Temple–Non-Temple Comparisons. As an additional measure of religious involvement within the Mormon Church, we are able to distinguish between couples by level of commitment at the time of the marriage by asking whether the ceremony was performed in a temple. Differences between temple and non-temple marriages generally parallel the differences that were observed between frequent and infrequent attenders: temple marriages are characterized by lower divorce and larger families (see Table 7.5). In fact, the differences in divorce and family size are even greater if temple marriage is the comparison criteria rather than if attendance is the criteria. Non-temple marriages are about five times more likely to end in divorce than are temple marriages. Remarriage is an exception, however. Those divorced from temple marriages are about as likely to remarry as are those divorced following non-temple marriage. In some cases, having been married in the temple may actually discourage remarriage since high level clearance must be obtained before a remarriage can be performed in the temple. Overall, however, results support conclusions regarding the association between religious involvement and family formation. 3

Controlling Education. Including education as a control variable (Table 7.6) does not negate any of the above conclusions regarding religious differences in family formation. Education is coded as less than twelve years, twelve years (completing high school), some college, or four or more years of college. Mormons have the highest levels of education (54.9 percent of men and 45.3 percent of women have had some college) followed by liberal Protestants (46.1 and 41.7 percent for men and women respectively), and then Catholics (41.4 for men, 29.9 for women). Conservative Protestants rank the lowest (27.0 and 22.2 percent for men and women respectively). Because of the small sample size, those with no religious preference are not included in this part of the analysis. Religious differences in divorce are reduced when education is controlled. Education is negatively associated with divorce, and Mormons have the lowest values on these variables such that when education is included, the religion effect is diminished. In other cases, however, education acts as a suppressor variable—religious differences are even greater after education is controlled. Education is negatively associated with marriage and family size among females, and with remarriage among males; Mormons with equivalent education score higher on marriage, remarriage, and family size. Thus, including education as a control actually accentuates the Mormon/ non-Mormon differences in these aspects of family formation. In sum, religious differences in family formation cannot be explained away as a simple by-product of religious differences in educational attainment.

Discussion​

Religious groups are characterized by different patterns of family formation in the United States. Identification with any religion is associated with a stronger traditional family orientation. Those reporting no religion are less likely to marry or remarry, are more likely to divorce, and to have smaller families than those stating a religious preference. Moreover, there is significant variation between religious groups. Compared with Catholics and Protestants, Mormons have higher rates of marriage and remarriage, lower divorce rates, and larger families. These differences are consistent with the pronatalist, pronuptialist Mormon theology of family continuity in the hereafter. Catholic-Protestant differences are also evident. Catholics are less likely to marry or remarry than Protestants but have less divorce and larger families. These differences are consistent with the Catholic position against divorce and birth control. Lower rates of marriage and remarriage for frequently attending Catholics could be a consequence of a more ambiguous theology of marriage. Celibacy is seen as superior to marriage, and remarriage is sometimes equated with adultery. Liberal and conservative Protestants are similar in most respects except perhaps in family size. In sum, religious identification appears to be associated with decisions people make about family life.

Introducing frequency of attendance as a control variable underscores the importance of religious involvement. Those who attend church frequently are more likely to marry (except among Catholics) and less likely to divorce; they have larger families and marry at more mature ages than do infrequent attenders. These patterns are also confirmed among Mormons when we compare temple and non-temple marriages.

All this is not to say that the causal direction goes directly from religion to family behavior. Those who, for whatever reason, do not conform to the religious ideal when it comes to family life undoubtedly feel less comfortable in family-oriented religious groups. They may be more inclined to drop out or to attend services less often than those whose family life more closely matches the norm. For example, the strong association between divorce and infrequent attendance may result because divorced people stop going to church, rather than because people who do not go to church get divorced. Likewise, singles and the childless may feel more out of place at religious gatherings than their married, child-rearing counterparts. Thus, it appears that religious organizations offer a form of social and theological support for family-oriented individuals.

We doubt that religious belief, in itself, would have a strong influence on family formation unless buttressed by social structural supports. To illustrate, endogamous marriage is facilitated by dances and other social functions that support the marriage market; child rearing is facilitated by neighborhood sharing of child care; and divorce is discouraged by advice on how to make marriages succeed. Detailed analysis would, we think, identify close linkages between profamily theology and a social structure which rewards profamily behavior.

Finally, there is the possibility that some other factors, such as socioeconomic differentiation, might explain religious differences in family formation. Education, it turns out, does explain part of the religious differentials in divorce, but it also accentuates some interdenominational differentials. Thus, differences in social characteristics may underlie some religious differences; it may mask others. Apparently the social characteristics hypothesis is a two-edged sword. Focus on religion and the interface between religious belief and the social structure of religious organizations deserves equal attention with social characteristics in our attempts to explain the association between religious behavior and family formation.

The family formation differentials we have documented in this paper will reinforce the religion-family connection. Higher fertility within religious groups will lead to perpetuation and growth of these groups. Differences in marriage, divorce, and remarriages result in congregations dominated by married people, and relationships with spouse and children will continue to be among the central concerns of a majority of church members. Thus, we predict that the interdependence between religion and family will persist. A major challenge facing religions, however, will be how to include the singles, divorced, and childless without relinquishing their meaning for families. To fail on either count will entail either loss of generality toward growing segments of our population, or weakening of a symbiotic relationship which is critical for the perpetuation of religious organizations.

Notes

Helpful comments from Howard Bahr, Dan Lichter, Darwin Thomas, and anonymous reviewers are gratefully acknowledged.

1 Cell sizes will be substantially smaller than the total number of respondents because some individuals do not qualify (e.g., they are too young or have not been married), and there are missing values on some questions.

2 Other denominations coded in the NORC data are less easy to classify on the basis of objective criteria, but they constitute fewer than 10 percent of all Protestants, so their exclusion will not greatly influence our results.

3 Reviewers questioned how much of the Mormon pattern was due to Mormon concentration in Utah. Moreover, temple marriage differentials could be due to the later age at marriage and the smaller percentage of premarital conceptions among the temple married. To address these issues, state of residence (Utah/non-Utah), age at marriage, and premarital conception (birth before or after 7 months of marriage) were considered as control variables. Slightly over one-third of the Mormon sample lived in Utah so that the patterns could not be dominated by Utahns. Utah residence has very little effect on marriage, divorce, remarriage, and fertility of Mormons where other factors are considered. The only exception is that Utah men are more apt to remarry than non-Utah men. Inclusion of age at first marriage, residence, and premarital conception in a model of female divorce (premarital conception was not ascertained for men) does not diminish the temple/non-temple divorce differential. Although these factors may influence patterns of family formation, they cannot account for the distinctiveness of Mormons.

References

Albrecht, S. L., Bahr, H. M., & Goodman, K. L. (1983). Divorce and remarriage. Westport, CT: Greenwood Press.

Albrecht, S. L., & Heaton, T. B. (1984). Secularization, higher education, and religiosity. Review of Religious Research, 26, 43–58.

Argyle, M., & Beit-Hallahmi, B. (1975). The social psychology of religion. London: Routledge and Kegan Paul.

Bouvier, L. F., & Rao, S. L. N. (1975). Socioreligiousfactors infertility decline. Cambridge: Ballinger Publishing Company.

Bumpass, L. L., & Sweet, J. A. (1972). Differentials in marital instability: 1970. American Sociological Review, 37, 754–766.

Campbell, B. L., & Campbell, E. E. (1977). The Mormon family. In C. H. Mindel & R. W. Haberstein (Eds.), Ethnic families in America (pp. 379–412). NY: Elsevier.

Chamie, J. (1981). Religion and fertility. Cambridge: Cambridge University Press.

Christensen, H. T. (1972). Stress points in Mormon family culture. Dialogue, 8 (4), 20–34.

Christensen, H. T. (1976). Mormon sexuality in cross-cultural perspective. Dialogue, 10 (2), 62–75.

D’Antonio, W. V. (1980). The family and religion: Exploring a changing relationship. Journal for the Scientific Study of Religion, 19 (2), 89–104.

D’Antonio, W. V., & Aldous, J. (Eds.). (1983). Families andreligions: Conflict and change in modern society. Beverly Hills, CA: Sage Publications.

Davidson, J. D. (1977). Socio-economic status and ten dimensions of religious commitment. Sociology and Social Research, 61 (4), 462–465.

DeHart, W. A. (1941). Fertility of Mormons in Utah and adjacent states. American Sociological Review, 6(6), 818–829.

DeJong, G. F. (1965). Religious fundamentalism, socio-economic status, and fertility attitudes in the Southern Appalachians. Demography, 2, 540–548.

Fienberg, S. E. (1980). The analysis of cross-classified categorical data. Cambridge: MIT Press.

Glock, C. Y., & Stark, R. (1965). Religion and society in tension. Chicago: Rand McNally.

Hargrove, B. (1979). The sociology of religion. Arlington Heights, IL: AHM Publishing.

Hastings, D. W., Reynolds, C. H., & Canning, R. R. (1972). Mormonism and birth planning: The discrepancy between church authorities’ teachings and law attitudes. Population Studies, 26 (\), 19–28.

Hunter, J. D. (1983). American evangelicalism. New Brunswick, NJ: Rutgers University Press.

Kobrin, F. E., & Goldscheider, C. (1978). The ethnic factor in family structure and mobility. Cambridge: Ballinger Publishing.

Marcum, J. P. (1981). Explaining fertility differences among U.S. Protestants. Social Forces, 60 (2), 532–543.

Mineau, G, Bean, L., & Skolnick, M. (1971). Mormon demographic history II: The family life cycle and natural fertility. Population Studies, 33, 429–446.

McCarthy, J. (1979). Religious commitment, affiliation and marriage dissolution. In R. Wuthnow (Ed.), The religious dimension: New directions in quantitative research. NY: Academic Press.

Mosher, W. D., & Hendershot, G. E. (1984). Religious affiliation and the fertility of married couples. Journal of Marriage and Family, 46 (3), 671–677.

Skolnick, M., Bean, L., May, D., Arbon, V., DeNevers, K., & Cartwright, P. (1978). Mormon demographic history I: Nuptiality and fertility of once-married couples. Population Studies, 32, 5–19.

Spicer, J. C., & Gustavus, S. O. (1974). Mormon fertility through half a century: Another test of the Americanization hypothesis. Social Biology, 21 (1), 70–76.

Talmage, J. E. (1968). The house of the Lord: A study of holy sanctuaries, ancient and modern. Salt Lake City, UT: Deseret Book.

Thomas, D. L. (1983). Family in the Mormon experience. In W. V. D’Antonio & J. Aldous (Eds.), Families and religions: Conflict and change in modern society (pp. 267–288). Beverly Hills, CA: Sage Publications.

Thornton, A. (1979). Religion and fertility: The case of Mormonism. Journal of Marriage and the Family, 41 (1), 131–142.

Westoff, C. F., & Potvin, R. H. (1967). College women and fertility values. Princeton: Princeton University Press.

Westoff, C. F, & Ryder, N. B. (1977). The contraceptive revolution. Princeton: Princeton University Press.

Westoff, C. F, & Jones, E. F. (1979). The end of “Catholic” fertility. Demography, 16 (2), 209–217.

Wilson, J. (1978). Religion in American society. Englewood Cliffs, NJ: Prentice- Hall, Inc.

Yoder, M. L. (1980). Religion as a determinant of fertility among white Americans, 1965. Unpublished doctoral dissertation, University of Wisconsin-Madison.