Folk wisdom informs us that “As the twig is bent so grows the tree” and that “The child is the father of the man.” Biblical verse (Proverbs 22:6) declares, “Train up a child in the way he should go: and when he is old, he will not depart from it.” The family is considered a primary force in shaping the values and attitudes of its members, especially its young members.

Children are typically socialized (or indoctrinated) by the religious beliefs and practices of their parents. In fact, many parents feel morally bound to instruct their children in religious matters. Just how effective is this indoctrination? Does the “tree” really grow as the “twig” is bent? If so, a strong positive correlation should exist between the religious involvement of parents and their offspring. Moreover, this correlation should occur not only when the offspring are young and under the supervision of their parents, but also after children become adults and leave their parents’ supervision.

Past Findings

Extant research typically supports the belief that parents do indeed influence the belief and practices of their offspring, but disagreement exists as to the strength of the influence. See, for example, reviews of research on the topic by Hyman (1959), Kalish and Johnson (1972), and Hoge et al. (1982). In his review of previous research, Hyman (1959) found generally “moderate” correlations between the values of parents and their offspring (the median Pearson’s r for the numerous studies cited was approximately .5). Bengtson (1975: 369) in a three-generation study reported weak to moderate value transmission and concluded that “generalizations concerning family influences on the development of values should be made with caution: similarities between parents and youth reflect their commonality of social location rather than direct transmission.” Necombe and Svehla (1937) found “high” correlations between parent and child attitudes for their heterogeneous religious sample but much weaker correlations in their subsamples of strongly religious people in specific denominations. They also noted declining correlations with increasing age of the offspring. A study of alcohol and drug usage among college students by Perkins (1985) found parental attitudes to have little direct effect on their offspring’s alcohol and other drug usage. Commitment to Judeo-Christian traditions was, however, found to be a significant moderating influence.

Focusing specifically on religious involvement, Hoge, Petrillo, and Smith (1982) report generally “weak” parent/child correlations. Furthermore, they found denomination to be a better prediction of adolescent religious involvement than parental religiosity. Kalish and Johnson (1972) found daughter/mother correlations on a religious value scale to be noticeably stronger than mother/grandmother correlations. They concluded that value similarities between generations are strongest between the adjacent generations who have most recently lived together. A recent study by Potvin and Sloane (1985) reports that adolescents whose parents are regular church attenders are five times more likely to score high on religious practice than adolescents whose parents are infrequent or nonchurch attenders. Furthermore, Hunsberger (1980) reports noticeably negative relationships between parental religious instruction and religious apostasy of the child. Dudley and Dudley (1986), however, report only moderate relationships between the religious values and attitudes of parents and those of their adolescent offspring.

Variances in the strength of the parent/child relationships cited can, in part, be explained by (1) the fact that parent/child behavior correlations are typically stronger than parent/child value correlations; (2) the subjective nature of such terms as weak, moderate, or strong (e.g. Bengtson, [1975], reports “weak to moderate” intrafamily value transmission with an r of .53 on one of his two scales, while Hoge and Petrillo [1978], report r’s of .48 and .60 as evidence of “very strong” parental influence); and (3) the age of the child—parental influences typically decline as the child becomes older and more independent.

Purpose of Study

This paper specifically investigates the short- and long-range influence of familial religious socialization among Southern Baptists and Mormons (members of The Church of Jesus Christ of Latter-day Saints). Both of these denominations have high retention rates among their members (Roof and Hadaway, 1979) and emphasize childhood religious indoctrination.

Mormon doctrine, for example, declares: “And again, inasmuch as parents have children in Zion, or in any of her stakes which are organized, that teach them not to understand the doctrine of repentance, faith in Christ the Son of the living God, and of baptism and the gift of the Holy Ghost by the laying on of the hands, when eight years old, the sin be upon the heads of the parents” (Doctrine and Covenants 68:25).

My data are derived from a systematic sample of Southern Baptists and Mormons in the St. Louis metropolitan area. Five hundred Mormons and five hundred Southern Baptists were mailed questionnaires; 238 Southern Baptists and 261 Mormons returned completed surveys. A list of the variables used in this study and how they are operationalized is provided in the appendix.

Analysis

As was already mentioned, children are typically socialized into the beliefs and practices of the parents. Hence, measures of parental religious involvement should provide useful indicators of religious socialization. Two such indicators in my data are the father’s and the mother’s church attendance during their child’s last year of high school. If these measures are viable indicators of familial religious socialization, we should expect a high correlation between the child’s church attendance during his or her last year of high school and his or her parents’ church attendance for the same period. The correlations are presented in Table 1 and, as expected, they are rather strong. The father/child church attendance correlation is .58 for the Mormons and .53 for the Baptists. The corresponding mother/child correlations are .66 and .59. Thus, parental church attendance is an excellent predictor of their child’s church attendance. Note also that the mother’s church attendance is a better predictor of the child’s church attendance than is the father’s church attendance.

Table 1. Correlations* between Past and Present Measures of Religious Involvement for Mormons and Southern Baptists

*Pearson's r is the measure of correlation used. For justification of its use with ordinal data see Labovitz (1970).
a = P < .05, b = P < .01, c = P < .001
LDS = Mormons, SBC = Southern Baptists

Another indicator in my data of childhood religious socialization is the respondent’s self-reported religiosity during his last year of high school. Childhood religiosity correlates with the father’s church attendance at .37 for the Mormons and at .32 for the Baptists; correlations with the mother’s church attendance are .44 and .35 respectively. Notice also that the two measures of childhood religious involvement—church attendance and religiosity—correlate strongly together at .68 for the Mormons and .62 for the Baptists.

These findings clearly support the belief that parents play a dominant role in the religious socialization of their offspring. The child’s religious involvement definitely tends to mirror his or her parents’ religious involvement. But what is the temporal efficacy of this childhood socialization? Are patterns of religious involvement acquired as a child maintained as an adult?

An analysis of the respondents’ adult religious involvement with both their childhood religious involvement and their parents’ church attendance should provide a tentative answer to this question.

Notice in Table 1 that parental church attendance correlates only weakly with respondents’ adult church attendance: the father/adult respondent correlations are .05 for the Mormons and .22 for the Baptists, and the mother/adult respondent correlations are .13 and .10 respectively. Also weak and typically nonsignificant are the correlations of parental church attendance with the other measures of their adult offspring’s religious involvement (that is, testimony, prayer, tithing, and self-reported religiosity). It appears that the impact of parental religious socialization is rather transitory, waning considerably after the child becomes an adult.

The relatively temporal effect of parental religious socialization is further supported by the fact that only mild correlations exist between the respondents’ childhood and adult attendance (Mormon .16, Baptist .12) and between childhood and adult religiosity (.29 and .18 respectively). In addition, these two measures of childhood religiosity correlate only weakly with three other measures of adult religious involvement: testimony, tithing, and prayer.

To this point this analysis has focused on the family of orientation, and specifically, parental influence. Let us now look briefly at the family of procreation and specifically at the husband/wife religious involvement relationship. Regretfully, only one variable in my data set—spouse’s religiosity—measures the spouse’s religious involvement. Additional measures would be desirable, but this measure does provide tentative information about the relationship. Among the Mormons, four of the five measures of adult religiosity correlate with spouses’ religiosity at .50 or higher, while correlations between spouses’ religiosity and the two measures for the respondents’ childhood religious involvement are in the teens. The Baptist correlations, though not as strong, reflect the same pattern. Typically the couples in the analysis are religiously quite similar. While it is possible that marital selection accounts for this similarity (measures of religious involvement at the time of engagement are lacking), the weak correlations between the respondents’ past religious involvement and spouses’ present religious involvement suggest that marital socialization is the more likely cause.

The final variable in Table 1 which remains to be examined is the number of best friends who are of the same faith as the respondents. Correlations between “friends” and the respondents’ five adult religious measures range from .42 to .61 for the Mormons and from .30 to .49 for the Baptists, and are all significant at the .001 level.

The correlations between “friends” and the respondents’ childhood religious involvement measures are, however, noticeably weak. In short, childhood religious involvement is a poor predictor of the faith of the adult respondents’ best friends. Yet the number of present best friends who are of the same faith as the respondent is an excellent predictor of the respondent’s present religious involvement. These several findings make eminent sense if we view parents, spouse, and friends as comprising primary groups for the respondents. Primary groups, because of their intimate, face-to-face, holistic nature, strongly influence the beliefs and practices (whether sacred or secular) of their members. In fact, this is why they are called primary groups, because they play a primary role in shaping the beliefs and values of their members.

When an individual leaves one primary group and joins another, the influence of the former group on the individual wanes while the influence of the new group increases. Hence, the parent/child, spouse/spouse, and friends/adult respondent correlations are all substantial.

While the parent/adult child, spouse/child respondent, “friends”/child respondent, and adult/child correlations are typically weak, the substantial correlations reflect primary group relations for the respondents. The weak correlations reflect relationships in which the respondent/other were not a primary group at the time of measurement.

It is possible that my analyses are inadequate. Intervening variables could be masking the true relationships. Such a variable could be convert status. It is well-known that both Baptists and Mormons are quite active in proselyting. In my sample, 56 percent of the Mormons and 53 percent of the Baptists are converts. If the converts were raised, by and large, in homes in which there was little religious involvement but after conversion as adults became religiously active, a strong correlation would exist between parent and child religious involvement, but a weak or possibly negative correlation would exist between parent and adult-child religious involvement. If the parent/child and the parent/adult-child correlations are both strong for the lifelong members, the negative or weak parent/adult-child convert association would mask over the strong association for the lifelong members.

In order to test for this possible masking effect, converts and lifelong members were analyzed separately (see Tables 2 and 3). As expected, childhood church attendance correlates strongly with parents’ church attendance for both converts and lifelong members. Of greater interest, however, is that the parent/adult-child associations are still typically weak for both the converts and the lifelong members. Thus, even for those individuals who were reared in their present faith, parental religious involvement is a poor predictor of adult religious involvement.

Table 2. Correlations* for Lifelong and Convert Mormons

*Pearson's r is the measure of correlation used. For justification of its use with ordinal data see Labovitz (1970).
a = P < .05, b = P < .01, c = P < .001
LLM = Lifelong Member, CON = Convert

Table 3. Correlations* for Lifelong and Convert Southern Baptist

*Pearson's r is the measure of correlation used. For justification of its use with ordinal data see Labovitz (1970).
a = P < .05, b = P < .01, c = P < .001
LLM = Lifelong Member, CON = Convert

Controlling for convert status, however, does make a difference in the child/adult correlations, particularly for the Mormons. In Table 1 the correlations of childhood church attendance with the five measures of adult religious involvement range from .14 to .19 for the Mormons and from .07 to .18 for the Baptists. When converts and lifelong members are differentiated (Tables 2 and 3), the Mormon correlations range from .36 to .44 for the lifelong members but from only .03 to .08 for the converts; corresponding correlations for the Baptists range from .14 to .36 and from -.02 and .16 respectively. Childhood religiosity correlations follow the same pattern. Thus, for the converts of both denominations childhood religious involvement shows little relationship to adult religious involvement; but among the lifelong members the childhood/adult religiosity measures are substantial for the Mormons and typically moderate for the Baptists.

Interestingly, lifelong member correlations between measures of childhood religious involvement and best friend’s religion or spouse’s religiosity are, in general, moderately strong. Contrary to my earlier conclusions, this suggests that for lifelong members the selection of spouse and friends (as well as socialization after selection) influences the high respondent/spouse and respondent/friends correlations.

Summary and Conclusions

Measures of the respondents’ childhood religiousness correlated moderately to highly with measures of their parents’ religious involvement. Neither childhood nor parents’ religious involvement, however, associated even moderately with the respondents’ adult religious involvement. Adult religious involvement did, on the other hand, correlate highly with spouse religiosity and with the number of best friends who were of the same faith as the respondent. Taken together, these findings suggest that the primary groups which individuals participate in strongly influence their religious involvement. The family of orientation, the family of procreation, and closest friends are classical examples of primary groups. Religious measures for each of these groups correlate strongly with measures of the respondents’ religious involvement during the time the respondents were participating members of these groups. But these correlations decline substantially when the respondents are not part of the group.

Due to possible differences in childhood socialization, lifelong members and converts were analyzed separately. Controlling for convert status had no noticeable effect upon the parent/adult-child correlations, but it did significantly alter the child/adult correlations. Among the lifelong members, child/adult correlations were moderately strong for the Mormons and mildly strong for the Baptists, but for the converts of both groups the correlations were negligible.

Assuming causality between the variables, my data suggest the following relationship: (1) parents’ religious involvement strongly influences the childhood religious involvement of their offspring; (2) childhood religious involvement negligibly influences adult religious involvement of both Mormons and Baptist converts: (3) the influence of childhood religious involvement on the adult religious involvement of lifelong members is moderate for Mormons and mild for Baptists; (4) parents’ religious involvement only weakly influences their child’s adult religious involvement; (5) a strong positive reciprocal relationship exists between the respondents’ adult religious involvement and (a) their spouses’ religiosity and (b) the faith of their best friends; and (6) childhood religious involvement of lifelong members weakly to moderately influences religiosity of spouse and faith of best friends (probably through both selection and socialization).

These findings support the vast majority of research on socialization. My emphasis on the critical role of temporal primary groups helps us to understand why parent/offspring correlation on values and behavior vary so much in strength from study to study. As the child ages, the primacy of the family of orientation wanes. It becomes less and less a primary group for the child, while other groups (friends, spouse, etc.) become more primary. Thus parent/child correlation tends to vary from strong for children to weak for adult offspring.

While a tree may grow as “the twig is bent,” humans are more supple than old wood. Present “bending” more than past “bending” directly shapes present religious involvement. Present influences, however, are naturally shaped by past influences.

As Bengtson (1972: 369) points out: “The family serves as an important mediating link in selecting or orienting the child to the multiple reference groups to which he or she can turn for value development in a pluralistic society.” Parental religious indoctrination, hence, has an indirect, if not a direct, influence on the religious involvement of adult offspring.

APPENDIX

ITEMS USED TO MEASURE THE VARIABLES

Childhood Religiosity

Which category best described you during your last year of high school (or when you were 17 years old)?

(5) very religious

(4) fairly religious

(3) mildly religious

(2) not very religious

(1) not at all religious

Adult Testimony

Have you received spiritual confirmation of the truthfulness of the gospel?

(3) I’m certain I have

(2) uncertain

(1) I have not

Best Friend’s Religion

Of your five closest friends, how many are members of your denomination?

(1) none

(2) one

(3) two or three

(4) four

(5) five

Adult Tithing

Last year what percentage of your income did you contribute to religion?

(1) zero or almost nothing

(2) 1 percent to 3 percent

(3) 4 percent to 9 percent

(4) 10 percent

(5) 11 percent or more

Adult Prayer

How often do you pray privately?

(1) I never pray

(2) I pray only on special occasions

(3) I pray once in a while but not regularly

(4) I pray several times a week

(5) I pray once a day or more

Adult Church Attendance

How often do you attend Sunday worship services?

(5) every week or just about

(4) two or three times a month

(3) several times a year

(2) once or twice a year

(1) I don’t go to church

Childhood Church Attendance

Father’s Church Attendance

Mother’s Church Attendance

How frequently did you, your father, and your mother attend church during your last year of high school (or when you were 17 years old)?

Adult Religiosity

Spouse Relgiosity

Which category best describes you (your spouse)?

Bibliography

Bengtson, V. L. 1975. “Generation and Family Effects in Value Socialization.” American Sociological Review 40: 358–71.

Doctrine and Covenants. 1981. Doctrine and Covenants of The Church of Jesus Christ of Latter-day Saints. Salt Lake City: The Church of Jesus Christ of Latter-day Saints.

Dudley, R. L., and M. G. Dudley. 1986. “Transmission of Religious Values from Parents to Adolescents.” Review of Religious Research 28: 3–15.

Hoge, D. R., and G. H. Petrillo. 1978. “Determinants of Church Participation and Attitudes Among High School Youth.” Journal for the Scientific Study of Religion 17: 359–79.

Hoge, D. R., G. H. Petrillo and E. I. Smith. 1982. “Transmission of Religious and Social Values from Parents to Teenage Children.” Journal of Marriage and the Family 44: 569–80.

Hunsaker, B. 1985. “Parent-University Student Agreement on Religious and Nonreligious Issues.” Journal for the Scientific Study of Religion 24: 314–20.

Hunsberger, B. 1980. “A Reexamination of the Antecedents of Apostasy.” Review of Religious Research 21: 158–70.

Hunsberger, B., and L. B. Brown. 1984. “Religious Socialization, Apostasy and the Impact of Family Background.” Journal for the Scientific Study of Religion 23: 239–51.

Hyman, H. H. 1959. Political Socialization: A Study in the Psychology of Political Behavior. Glencoe, IL: The Free Press.

Kalish, R. A., and A. I. Johnson. 1972. “Value Similarities and Differences in Three Generations of Women.” Journal of Marriage and the Family 34: 49–54.

Labovitz, S. 1970. “The Assignment of Numbers to Rank Order Categories.” American Sociological Review 35: 515–24.

Newcomb, T., and G. Svehla. 1937. “Intra-Family Relationships in Attitude.” Sociometry 1: 180–205.

Perkins, H. W. 1985. “Religious Traditions, Parents and Peers as Determinants of Alcohol and Drug Use Among College Students.” Review of Religious Research 27: 15–31.

Potvin, R. H., and D. M. Sloane. 1985. “Parental Control, Age, and Religious Practice.” Review of Religious Research 27: 3–14.

Roof, W. C., and C. K. Hadaway. 1979. “Denominational Switching in the Seventies: Going Beyond Stark and Glock.” Journal for the Scientific Study of Religion 18: 363–77.

Troll, L. E., B. L. Neugarten, and R. I. Kraines. 1969. “Similarities in Values and Other Personality Characteristics in College Students and Their Parents.” Merrill-Palmer Quarterly of Behavior and Development 15: 323–36.