Content uploaded by Tom O'Connor
Author content
All content in this area was uploaded by Tom O'Connor on Dec 18, 2017
Content may be subject to copyright.
The Relative Influence of Youth and Adult
Experiences on Personal Spirituality
and Church Involvement
THOMAS P. O’CONNOR
DEAN R. HOGE
ESTRELDA ALEXANDER
We surveyed 206 young adults who had grown up in middle-class churches in three denominations—Baptist,
Catholic, and Methodist—who were first studied at age 16 in 1976. The goal was to assess the relative strength of
youth and adult influences on their personal religious and institutional church involvement at age 38. The determi-
nants of these two outcomes at 38 varied widely. For personal spirituality such as private prayer, attending Bible
classes, and reading religious material, we found strong youth and adult determinants such as the denomination of
one’s youth, church youth group participation, having an experience since high school that changed their feelings
about the church, and attending church with one’s spouse. For church involvement, however, all but one of the
determinants occurred after age 16, mainly the experiences of being inactive in church after high school, switching
denominations, having children, and going to church with one’s spouse. Social learning theory was the best theory
for explaining these findings.
Church leaders of all Christian denominations want to know what factors produce Christian
adults who have an active spirituality and are engaged in their churches. How important are
childhood experiences for determining adult church involvement and personal religiosity? What
kind of adult experiences have the most long-lasting spiritual effect? What causes some young
adults to remain involved in church and others to depart? Do adult experiences supersede or
build on youthful influences in determining adult spirituality and church behavior? Do structural
influences such as denomination or gender interact with individual factors to produce greater or
lesser involvement in religious practice?
Questions such as these have stimulated much research, including the present article. We
report results of a longitudinal study of youth in Baptist, Methodist, and Catholic denominations
who were studied at age 16 and 22 years later at age 38. To explain the importance of our study
and situate it theoretically we need to review prevailing ideas of religious change and commitment
during youth and early adulthood. These are basically nascent theories based on past findings about
the importance of different influences on adult religious practice, especially church involvement.
They can be categorized into three types: family life cycle theory, social learning theory, and
cultural broadening theory.
Family life cycle theory states that the needs and tasks an individual must address during
different parts of the family life cycle determine his or her church involvement. Research has
found that church disaffiliation occurs most often in the teenage years and the early 20s, with less
disengagement after the early 20s and often a reentry into church life later when the young adult is
building a family (Roozen 1980). Adult church involvement is stronger for married persons than
single persons and stronger for persons with children than for those without (Mueller and Cooper
Thomas P. O’Connor is President of the Center for Social Research in Maryland and Administrator, Religious Services,
Oregon Department of Corrections. He is a doctoral candidate at Catholic University of America, Washington, DC. 1420
Court Street NE, Salem, OR 97301. Email: oconnortom@aol.com
Dean R. Hoge is Professor of Sociology, Catholic University of America, Washington, DC 20064. Email: hoge@cua.edu
Estrelda Alexander is Associate Dean for Community Life, Wesley Theological Seminary, 4500 Massachusetts Ave. NW,
Washington, DC. 20016. Email: EAlexander@Wesleysem.edu
Journal for the Scientific Study of Religion 41:4 (2002) 723–732
724 JOURNAL FOR THE SCIENTIFIC STUDY OF RELIGION
1986; Stolzenberg, Blair-Loy, and Waite 1995). Stolzenberg, Blair-Loy, and Waite (1995) suggest
that religious disaffiliation and reaffiliation are influenced by the strength of family ties, marital
history, and attitudes to family life. Other researchers point out that the influence of family life
may be gender specific. Since men’s religious roles are less institutionalized or socially defined
than women’s, men’s involvement in church may depend more on nonreligious factors such as
changes in family status (Wilson and Sherkat 1994).
Social learning theory views religious behavior as a learned behavior arising out of a particular
life context. As life contexts change, people change, primarily by observing role models and
practicing new behaviors (Bandura 1977). Thus, early socialization into a religious tradition is
a result of the model provided to children by parents and other adults (Hunsberger 1983; Hoge,
Johnson, and Luidens 1993). In the teenage years when peer influences tend to outweigh parental
influences, teenagers will become less church involved as they learn new behaviors from their
peers that help to relate them to their changing situations. Similarly, as an adult, a person’sown
set of attitudes, beliefs, and values and those of his or her adult friends will influence that person’s
religious and church involvement. Social learning theory would suggest that religious learning
will vary from one denomination or religious group to another depending on the kind of religious
modeling and practices that are most prevalent in each group.
Cultural broadening theory (also called “localism theory”) considers church involvement
in terms of “plausibility structures”(Cornwall 1989). This perspective contends that in a highly
differentiated and pluralistic modern society, the maintenance of a particular religious worldview
requires that a community of people support one another’s belief in it in daily interaction. A
“local orientation”characterized by involvement with local institutions and networks of friends
and family (as opposed to a “cosmopolitan”orientation), should predict higher rates of church
involvement (Roof 1978). Cognitive broadening that occurs during high school and college can
produce “liberalized social attitudes, greater cosmopolitanism, religious skepticism, and a sense
of moral and religious relativity”(Hoge, Johnson, and Luidens 1993). College experience, re-
gional or geographic relocation, interfaith marriage, or other shifts in one’s life may disrupt one’s
interactions with earlier religious practices and church networks and so lead to lesser involvement
in the church (Roozen 1980; Hoge and Petrillo 1978a).
Past research into determinants of adult religious involvement has had limitations. Many of
the studies asked adult respondents to recall aspects of their childhood and youth rather than gather
independent data about those early years. Only a few, such as Myers’s study of the inheritance of
religiosity, Dudley’s study of Seventh Day Adventist youth from age 16 to 26, Dillon’s study of
religious involvement from adolescence through adulthood, and the present study, utilize infor-
mation gathered at widely different points of time (Myers 1996; Dudley 1995, 1999; Dillon 2000).
In addition, few studies were able to study denominational variations, and many studies focused
only on church involvement as an outcome to the neglect of personal religious involvement. The
present study overcomes these limitations because its design permits a longitudinal study of both
personal religious and church involvement across three denominations.
We tested three general hypotheses that were based in the theories of religious development
discussed above. (1) Family-related experiences during adulthood, such as marriage and having
children, will produce both greater church and personal religious involvement. (2) Due to social
learning, greater religious involvement as a child and greater religiosity of the family of origin will
foster adult church involvement and personal spiritual practice. (3) Adult experiences producing
cognitive broadening, such as college education and interfaith marriage, will reduce religious
commitment. We were also interested in exploring the relationship between personal religious
practice and institutional church involvement and the influence of childhood denomination on
adult religious practices. If attendance at church declines, does personal religious practice also
decline, or are personal religiosity and church attendance two independent religious constructs?
Do different denominations influence the adult religious behaviors of their young members in
different ways?
INFLUENCE OF YOUTH AND ADULT EXPERIENCE 725
METHOD
The study is a follow-up to a 1976 study by Dean Hoge and Gregory Petrillo. They gathered
extensive data on 451 Baptist, Catholic, and Methodist 10th graders regarding their faith de-
velopment, church involvement, and religious attitudes (Hoge and Petrillo 1978a, 1978b; Hoge,
Petrillo, and Smith 1982). The subjects in the original study, most of whom were aged 16, be-
longed to 35 churches in the Maryland suburbs of Washington, DC. Most were in white middle-
or upper-middle-class churches. Most also attended church regularly and participated in church
youth programs. Hoge and Petrillo set out to study all the 10th graders who were on the churches’
roles; 68 percent of the 10th graders responded. The original sample contained some bias, since
youth who participated in church programs were disproportionately cooperative in data collection.
We set out to interview all the subjects who had participated in the earlier study. The earlier
data and records had been preserved and we searched for the 451 target persons in a variety of
ways. We visited the churches again, and in all but one of them we were able to hire an informant
who had a long history of involvement with the congregation to help find the target persons.
In addition, we used Internet and phone book searches, and asked the subjects we located for
information about the other subjects in their church. Using these methods we found some current
information on 285 (63 percent) of the original subjects and conducted a phone interview with 206
of them in 1998 (a few interviews took place toward the end of 1997). Thirteen people declined
our interview offer, three were out of the country, two were deceased, and we were unable to make
contact with the remaining 61. Not counting the two deceased people, we interviewed 46 percent
of the target population. The present article is based on these 206 cases (61 original Baptists, 68
original Catholics, and 77 original Methodists).
We checked on possible biases by comparing the youth data for the 206 adult subjects with
the 245 cases from the original sample whom we did not interview on gender, mother’s church
attendance, father’s church attendance, respondent’s church attendance as a youth, participation
in church youth programs, and belief in the importance of church teachings. We found that the two
groups did not differ significantly on these variables except that the 206 persons in the follow-up
study had reported slightly higher church participation by their mothers (p < 0.05). We conclude
that our data has a small bias in that the follow-up persons were reared in homes that were a little
more church involved than the original sample.
DESCRIPTION OF THE SAMPLE
In both the original and the follow-up samples 55 percent of the respondents were female
and 45 percent were male. The denominational breakdown across studies was also similar; in
1976 it was 33 percent Baptist, 34 percent Catholic, and 33 percent Methodist, and in 1978
it was 30 percent originally Baptist, 33 percent originally Catholic, and 37 percent originally
Methodist. Approximately half (55 percent) of the follow-up sample were female—original
Baptist 56 percent, Catholic 50 percent, and Methodist 58 percent. The post-high-school ex-
periences of the adults in the three denominations were also surprisingly similar. We found no
significant differences (p <0.05) among original Baptists, Catholics, and Methodists on gender,
race, age, education level, having attended a Christian college as opposed to a different college,
employment status, marital status, number of marriages, having children, and number of children.
Thus the three denominational groups had many demographic and life cycle variables in common
and this reduced the possible confounding effects of the demographic and life cycle variables on
our outcome variables.
Most of the subjects were 38 years old and nearly all (97 percent) were white. A majority of the
respondents (74 percent) were currently married, while 18 percent were never married, 8 percent
were divorced or separated, and 1 percent were living with a partner. A surprisingly high percent
(88 percent) of those who were ever married were still in their first marriage, 12 percent were in
726 JOURNAL FOR THE SCIENTIFIC STUDY OF RELIGION
TABLE 1
HISTORY OF AND CURRENT SPIRITUAL AND CHURCH INVOLVEMENT BY
ORIGINAL DENOMINATION
Baptist Catholic Methodist
(N =61) (N =68) (N =77)
%% %
∗Percent attending church (weekly or more) as a youth 82 83 58
∗Percent attending church (weekly or more) as an adult 51 38 30
Percent praying privately (very or somewhat frequently) as a youth 20 18 13
Percent praying privately (weekly or more) as an adult 62 43 38
∗Ever become inactive in church? Yes 69 81 86
Mean age for becoming inactive 21 22 20
Mean age for becoming active again 29 28 27
Currently active in church? Yes 72 69 60
Currently a member of a church? Yes 79 76 68
∗Mean hours per month spent on church affairs of any kind, including 11.7 7.2 8.9
worship (as an adult)
Ever have an event since high school that changed your feeling about 52 63 57
church? Yes
Result of event? (N =119)
Church was more important 63 50 70
Church was less important 31 43 21
No difference 6 7 9
Personal religious practice as an adult (weekly or more) in last year
∗Attended a prayer meeting 18 4 5
∗Attended a Bible study class 30 3 8
∗Read the Bible 53 16 26
∗Listened to Christian programming on radio or TV 46 10 17
Read a religious book other than the Bible 21 13 13
∗Significant at 0.05 by F-test or chi-square.
their second, and 1 percent were in their third. Seventy percent of the subjects had children, an
average of 2.2. The subjects were well educated; 15 percent had a high school diploma, 17 percent
had some college, and the rest had a bachelor degree (39 percent), a master degree (21 percent),
or a professional/doctoral degree (9 percent). Of those who went to college, 26 percent attended
a Christian college. However, only 14 percent were ever active in a Christian campus ministry
or group. Baptists were much more likely to have been active in campus ministries or groups
(27 percent of Baptists; 5 percent of Catholics; 7 percent of Methodists). Most (87 percent) were
employed (12 percent were homemakers, 2 percent were unemployed) and the occupations of the
subjects place them in the middle- or upper-middle class; the majority were college professors or
administrators, teachers, physicians, engineers, managers, business owners, accountants, lawyers,
sales workers, and government employees.
Table 1 presents some of the findings about the history and current religious involvement of
our subjects. The first four rows of Table 1 show that for all three denominations the frequency
of church attendance drops from age 16 to age 38 but the frequency of private prayer actually
increases during this time. This is a very interesting finding because it gives us our first clue that
personal religious and church practice are different aspects of adult religiosity. Overall, 79 per-
cent of the sample said they had at some time been inactive in church, i.e., had attended church
INFLUENCE OF YOUTH AND ADULT EXPERIENCE 727
less than 12 times a year. Men and women were equally likely to become inactive. However,
the percentage that became inactive after high school was almost significantly different (p =
0.053) across the three denominations: for Baptists it was 69 percent, for Catholics 81 percent,
and for Methodists 86 percent. The age of becoming inactive varied, with most (68 percent)
becoming inactive by age 21; the average age for becoming inactive was also 21. Almost half
(48 percent) of those who became inactive said they did so for personal or motivational reasons
such as “being too busy,”“busy with family obligations,”“lack of interest,”and “feeling bored
or lazy;”about 20 percent named reasons that pertained to some conflict with faith or the church,
such as “disagreed with church,”“doubted the faith,”and “disliked worship style;”about 16 per-
cent named geographical or family reasons like “left home,”“moved away from family,”and
“no more parental pressure;”6 percent gave interpersonal reasons such as “disliked the pastor,”
“spousal influence,”and “influence of friends;”6 percent were unsure; and 4 percent gave other
reasons.
Over half (58 percent) of the respondents who had been inactive said they became active again
later, and by age 38 a majority of our sample (67 percent) were currently active: 72 percent of the
original Baptists, 69 percent of the Catholics, and 60 percent of the Methodists (not significant
atp<0.05). However, not all returned to the same denominations; 62 percent of the original
Baptists were actively Baptist, 81 percent of the Catholics stayed Catholic, and 62 percent of the
Methodists remained Methodist (significant at p <0.05). Most (62 percent) became active again
between the ages of 23 and 30; the average age was 28. The age pattern for becoming inactive
and active again is similar to that found by Hoge, Johnson, and Luidens (1993) and by Hoge et al.
(2001). The main reasons the respondents gave for becoming active again were related to marriage
and children or to spiritual needs. About 25 percent said “children,”22 percent said “spiritual
need”or “marital problems,”15 percent said “influence of spouse,”6 percent said “influence of
others,”6 percent said “conversion,”and the rest gave other responses.
We asked, “Since you completed high school, has there been any particular event, experience,
or relationship which has changed your feelings about the church?”Over half said yes (Table 1).
Baptists, Catholics, and Methodists were equally likely to have had such experiences. What
were the experiences? About 24 percent said it was marriage and/or having children. About 22
percent told of a conflict with the church or a minister or priest, usually about sexual issues,
marriage, remarriage, or being gay. About 21 percent said it was the death of a family member
or loved one, illness of a loved one, or their own illness. About 14 percent spoke of learning
that had enlarged their horizons, of exposure to other religions, or of their investigations about
religion. About 10 percent told of positive experiences with pastors or religious friends. The effect
of these experiences varied depending on what happened. Respondents who told of marriage,
children, deaths, or illnesses tended to say the experience made the church more important to
them. Respondents who told of conflict with church leaders or disgust at church irregularities said
the experiences made the church less important. On balance, the events made the church more
important. Table 1 also shows that original Baptists compared to Catholics and Methodists spent
the highest number of hours per month on church affairs and were also significantly more likely
to engage in adult personal religious practices such as attending prayer meetings, Bible study
classes, reading the Bible, and listening to Christian programming on radio or TV.
DEPENDENT AND INDEPENDENT VARIABLES
Dependent Variables
We identified three outcome or dependent variables: (1) ever became inactive in church life;
(2) personal religious involvement; and (3) church involvement. Our first outcome variable—
becoming inactive in church—could be defined as an interim outcome variable because it took
728 JOURNAL FOR THE SCIENTIFIC STUDY OF RELIGION
place prior to the time of our adult survey but after our youth survey. Our second and third
outcome variables were the level of personal religious involvement or spirituality and the level
of church involvement at the time of our adult survey at approximately age 38. We constructed
our two final outcome measures by factor-analyzing a wide range of available outcome mea-
sures at age 38 into three different dimensions. Using the variables that fell into the first two
dimensions, we constructed two outcome scales or measures, with higher scores indicating: (1)
greater personal religious involvement; and (2) greater church involvement. These two outcome
measures—personal religious involvement and church involvement—intercorrelated at 0.63.
The Personal Religious Involvement Scale was made from seven items (Cronbach’s alpha =
0.86). Six were self-reports of religious activities within the last year: attended a prayer meeting;
attended a Bible study class; read the Bible; listened to Christian programming on radio or TV;
had a personal time of devotion or prayer; and read a religious book other than the Bible. They
were scored: weekly or more =3; occasionally =2; and never =1. The seventh question asked
about how many times the person has prayed privately in the last seven days. The responses were
scored ; more than once a day =3; about once a day =2.5; several times =2; about once =1.5;
and never =1.
The Church Involvement Scale was composed of four items (Cronbach’s alpha =0.81):
church attendance in the last year, scored in three levels; number of hours spent in an average
month on church affairs including time for meetings, committee work, travel, study, and worship,
scored from 0 hours to 23 hours; “How often in the last year have you given money to your
church?”and “How often in the last year have you volunteered for a religious activity?”scored:
weekly or more =3; occasionally =2; and never =1. All responses were to range from 1 to 3.
INDEPENDENT OR PREDICTOR SCALES
In our analysis to predict our three outcomes we used a wide range of individual items from
the youth and the adult surveys such as gender, denomination, education level, and marital status.
In addition we created four scales that were theoretically appropriate to our analyses. From the
youth data we created a seven-item Youth Creedal Assent Scale (from King and Hunt 1975)
that measures general adherence to traditional Christian creeds (Cronbach’s alpha =0.84). For
example, two items state, “I believe in eternal life,”and “I believe that God revealed Himself
to man in Jesus Christ.”This scale correlated –0.46 with the Youth Religious Relativism Scale,
which was constructed from five items, such as “All the great religions of the world are equally
good and true,”and “not all the churches have God’s truth; many are in serious error”(Cronbach’s
alpha =0.83).
We also created a two-item Youth Traditional Sexual Beliefs Scale composed of two strongly
agree to strongly disagree items scored from 5 to 1: “Publications that dwell on and illustrate
sex acts should be banned from newsstands”and “It is morally right for a couple committed
to each other to have premarital sexual intercourse.”After the scoring for the latter question
was reversed, the two items correlated at 0.37. The Youth Traditional Sexual Beliefs Scale score
correlated 0.48 with the Youth Creedal Assent Scale and –0.39 with the Youth Religious Rela-
tivism Scale. This measure of sexual beliefs as a 16 year old proved to be important in predicting
whether individuals dropped out of church life. Finally, the youth data contained seven differ-
ent measures for youth group participation (Bible study groups, retreats, trips, service projects,
choirs, lectures/courses, and youth worship services) and from these we created a Youth Program
Participation Scale composed of the mean of the seven measures coded: very much involved
=3, quite involved =2, slightly =1, not at all =0, and not offered =0 (Cronbach’s alpha =
0.85). The Youth Program Participation Scale correlated 0.39 with the Youth Creedal Assent Scale,
–0.32 with the Youth Traditional Sexual Beliefs Scale, and 0.19 with the Youth Traditional Sexual
Beliefs Scale.
INFLUENCE OF YOUTH AND ADULT EXPERIENCE 729
PREDICTORS OF BECOMING INACTIVE IN CHURCH
We have already described how a large majority of the sample (79 percent) said they had,
at some time, become inactive in church, i.e., had attended church less than 12 times a year. To
see what youthful variables would predict this intermediate outcome (dropping out from church)
between age 16 and 38 we used logistic regression to analyze the influence of our individual and
scale youth variables identified above. We entered each of the youth variables into a forward step
(likelihood ratio) logistic regression to see if they would predict which respondents had never
become inactive versus those who had. Surprisingly, only three of the youth variables remained
in the equation at the 0.05 significance level: a yes/no question that read “Has there ever been a
period in your life when you reacted either partially or wholly against the beliefs taught you?”;
frequency of church attendance; and the Youth Traditional Sexual Belief Scale. On the question
regarding reacting against beliefs, the odds on those who had never rebelled staying active after
adolescence were 2.9 times higher than for those who had rebelled. On church attendance, the
odds on those who attended weekly or more staying active were three times higher than for those
who had attended less than weekly. On traditional sexual beliefs, the odds on those who held more
traditional sexual beliefs staying active increased in a linear fashion as subjects scored higher on
this five-point scale.
This unexpected finding about the influence of sexual beliefs alerts us to the importance of
sexual issues for high school youth. Possibly, sexual morals and sexual problems may be more
important than parental religious practices and theological issues in determining the future church
involvement of teenagers.
PREDICTORS OF PERSONAL RELIGION AND CHURCH INVOLVEMENT AT AGE 38
In a first test of the importance of the youth factors on the two main adult outcomes—personal
religious and institutional church involvement at age 38—we constructed numerous correlations.
Of the two outcome scales, experiences and attitudes during youth tended to be more often and
more strongly correlated with personal religious behavior at age 38 than church involvement.
Gender, frequency of church attendance, and frequency of private prayer were significantly cor-
related with personal religious practice but not with church involvement. Denomination, Bible
knowledge, church youth program participation, youth creedal assent, youth religious relativism,
and youth traditional sexual beliefs were significantly correlated with both personal religiosity
and church engagement.
Having examined the relationships between the variables of interest at a bivariate level, we
turned our attention to a multivariate analysis of the impact of the youthful and adult independent
variables on the two main dependent variables—the 1998 measures of personal religious involve-
ment and church involvement. First we selected the five most powerful youth predictors and the
six most powerful adult predictors in our data from the results of preliminary regression analyses
involving a wide selection of youth and adult variables that could have potentially predicted the
outcome variables. In this fashion we eliminated several of the youth variables, including: years
of Catholic school and of CCD (for Catholics), years of Sunday school (for Protestants), parental
church attendance, respondent’s church attendance as a youth, Bible knowledge, and the Youthful
Relativism Scale. We also eliminated several of the adult variables, including interfaith marriage
and current marital status. Having selected the 11 most powerful youth and adult predictors we
again used regression analysis to regress the predictor variables on our two outcome variables.
Table 2 depicts the regression solutions.
Table 2 shows two models for each of the outcome variables, first a test of the predictive
power of the youth variables at age 16, then a test of all variables gathered at age 16 and at
age 38. From the adjusted R2figures we see that the youth variables predict personal religious
involvement at age 38 (adjusted R2=0.297) much better than they predict church involvement
730 JOURNAL FOR THE SCIENTIFIC STUDY OF RELIGION
TABLE 2
REGRESSION COEFFICIENTS FOR PREDICTORS OF THREE ADULT
OUTCOMES (BETAS)
Personal
Involvement Religious Church Involvement
Youth All Youth All
Youth Variables
Gender (female =1, male =0) 0.14∗0.06 0.12 0.02
Original Baptist vs. other (Baptist =1, others =0) 0.31∗0.31∗0.11 0.06
Frequency of private prayer 0.14∗0.07 0.02 0.04
Church Youth Program Participation Scale 0.27∗0.32∗0.22∗0.24∗
Creedal Assent Scale 0.07 0.05 0.06 −0.04
Adult Variables
Ever inactive in church life? (yes =1, no =0) −0.19∗−0.26∗
Particular event or experience that changed your 0.16∗0.02
feelings about the church? (yes =1, no =0)
Years of education 0.00 0.00
Joined a different denomination (yes =1, no =0) 0.26∗0.30∗
Have children (yes =1, no =0) −0.03 0.13∗
Respondent and spouse attend church together (yes =0.08 0.27∗
1, no =0)
Adjusted R20.297 0.494 0.094 0.395
∗Significant at 0.05.
(adjusted R2=0.094). It is striking that no measures of parental religious involvement remained
in either of the models and that denomination and gender (two structural variables) remained in the
model predicting personal religious involvement from the youth variables. This finding reinforces
afinding from the original baseline study. The study in 1976 gathered questionnaire data from the
parents of the youth and tested various models of value transmission from parents to youth but
found weak parent-youth correlations on a whole series of variables. The strongest determinant
of the young people’s religious beliefs and behavior was the denomination they belonged to, not
the specific beliefs or behaviors of their parents (Hoge, Petrillo, and Smith 1982).
The adjusted R2for the youth variables on personal religion was 0.297, and the adult variables
added another 0.197, for a total of 0.494 compared to a total R2of 0.395 (0.094 for the youth
variables plus 0.201 for the adult) for church involvement. The predictors of adult personal
religious behavior were original denomination, participation in church programs as a youth,
never becoming inactive, changing denomination as an adult, and having spiritually important
experiences as an adult. The predictors of adult church involvement were participation in church
youth programs, never becoming inactive, joining a different denomination, having children, and
attending church with one’s spouse. Thus, the strongest predictors of church involvement occurred
after age 16 but the predictors of personal religious involvement occurred before and after age 16.
We were surprised how weak the youth variables were as determinants of church involvement
at age 38. The only aspect of youth that proved to be a significant predictor was the amount of
involvement in church youth programs. Parents’church attendance (as reported by youth), years
of Catholic school, years of Sunday school, private prayer, and church attendance as a youth turned
out to be nonsignificant. The type of youth group involvement was not very crucial; involvement
in choir or folk groups and participation in trips and Bible studies were as predictive of age 38
INFLUENCE OF YOUTH AND ADULT EXPERIENCE 731
church involvement as were lectures, courses, and youth worship services. Table 2 shows that
events after age 16 were by far the most important determinants of church involvement at age
38: never being inactive, joining a different denomination, having children, and attending church
with one’s spouse.
Are there differences between men and women in the determinants of personal religious
and church involvement at age 38? We calculated regressions for each gender separately and
found that the predictive power of the combined variables was consistently stronger for the men.
In particular, two adult variables had stronger betas for the personal religious involvement of
men than for women. The first variable was whether the men were ever inactive in church life.
Having been inactive in church life seemed to have had a more lasting impact on reducing men’s
personal religious involvement and church involvement. Second was whether the men had joined
a different denomination (and this move was no doubt due partly to the influence of their spouses).
The men who had switched were relatively stronger in personal religious involvement and church
involvement at age 38. While the exact interpretation of these findings is unclear, we were able to
predict men’s personal religious involvement at age 38 somewhat better than women’s, and the
difference was in the adult variables. This finding agrees with Stolzenberg, Blair-Loy, and Waite
(1995) and Wilson and Sherkat (1994).
CONCLUSION
Our longitudinal study of youth in the suburbs of Washington, DC provides new insights into
questions and theories about what factors produce Christian adults who have an active spiritual life
and are involved in church. First, how pervasive is the pattern of dropping out and returning? It is
clearly the norm. Overall, 79 percent dropped out at some time, and of them, 56 percent returned
by the time we re-interviewed them. Second, youth and adult variables differentially condition
personal spirituality and institutional church practice. Adult influences have a greater impact on
a person’s church involvement in their late 30s than on their spiritual practices. However, both
youth and adult experiences influence personal religious involvement and the combined impact of
the youth and adult variables means that personal religious involvement is more predictable than
church involvement. Church involvement is mostly determined by adult experiences, not earlier
religious upbringing, practices, or beliefs.
Third, the finding that almost none of the youth variables gathered at age 16 predicted church
going at age 38 was unanticipated, especially because we had a wide variety of variables to
test on these persons from when they were 16 years old. But with one exception—youth group
participation—none of the youth variables could predict the person’s church involvement when
they were 38 years old.
Fourth, do denominational differences at age 16 continue until the persons are in their late 30s?
Yes. Compared with subjects who were originally Methodist or Catholic, original Baptists were
more personally religious in their adulthood. However, original denomination was not predictive
of the level of church involvement as an adult.
What religious aspects of childhood and youth have the most long-lasting effects on the
personal adult religious involvement of middle-class youth? To our surprise it was not parents’
church attendance, the amount that parents talked to their children about religion, frequency
of church attendance as a youth, years of Catholic school, years of Catholic CCD, or years
of Protestant Sunday school. It was: (1) the culture of the denomination in which the person
was raised, including its teachings and habits; and (2) the amount of involvement in church
youth programs. Baptist youth were more involved in personal religious practices as adults than
either Methodist or Catholic youth. Something in their Baptist upbringing that we could not
specifically measure had a global effect on them. Youth of all denominations who were more
involved in church youth groups of any kind were also more involved in personal religious
practice.
732 JOURNAL FOR THE SCIENTIFIC STUDY OF RELIGION
The youth in our sample were from middle-class families, and many had achieved a high
level of education. We had suitable information to assess any impact of higher education and
cultural broadening on them after age 16, and our measures of these factors all proved weak. The
cultural broadening theory does not appear to have explanatory power for middle-class youth.
Rather, factors such as the overall religious culture of the family and denomination, church youth
involvement, and later experiences of becoming inactive or joining a different denomination
proved more important. Therefore, two theories—social learning theory and family life cycle
theory—were somewhat supported by our study and each one proved most useful for a particular
outcome. Social learning theory was very pertinent for explaining both personal religious and
church involvement, as most of the important factors, e.g., participation in church youth groups,
becoming inactive in church, joining a different denomination, can be understood as occasions
within which religious modeling and learning takes place. Family life cycle theory was not really
pertinent for explaining personal religious involvement and was only somewhat related to church
involvement because of the importance of having children. Future researchers need to include
denomination as a variable of interest and distinguish between personal and public adult religious
outcomes when assessing which factors in childhood and young adulthood predict adult religiosity.
ACKNOWLEDGMENTS
The authors would like to thank Crystal Parikh for her collaboration on this project. The study was made possible
by a generous grant from the Louisville Institute.
REFERENCES
Bandura, A. 1977. Social learning theory. Englewood Cliffs, NJ: Prentice-Hall.
Cornwall, M. 1989. The determinants of religious behavior: A theoretical model and empirical test. Social Forces
68(2):572–92.
Dillon, M. 2000. Tracking religious involvement over the life course: Evidence from a 70-year longitudinal study. Paper
read at American Sociological Association. Washington, DC.
Dudley, R. L. 1995. Grace, relevancy, and confidence in the future: Why Adventist young adults commit to the church.
Journal of Psychology and Christianity 14(3):215–27.
———. 1999. Youth religious commitment over time: A longitudinal study of retention. Review of Religious Research
14(3):218–27.
Hoge, D. R., W. Dinges, B. Johnson, and J. Gonzales. 2001. Young adult Catholics: Religion in a culture of choice. Notre
Dame, IN: University of Notre Dame Press.
Hoge, D. R., B. Johnson, and D. A. Luidens. 1993. Determinants of church involvement of young adults who grew up in
a Presbyterian churches. Journal for the Scientific Study of Religion 32(3):242–55.
Hoge, D. R. and G. H. Petrillo. 1978a. Determinants of church and participation and attitudes among high school youth.
Journal for the Scientific Study of Religion 17(4):359–79.
———. 1978b. Development of religious thinking in adolescence: A test of Goldman’s theories. Journal for the Scientific
Study of Religion 17(2):139–54.
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(3):569–80.
Hunsberger, B. 1983. Apostasy: A social learning perspective. Review of Religious Research 25(1):21–38.
King, M. B. and R. A. Hunt. 1975. Measuring the religious variable: National replication. Journal for the Scientific Study
of Religion 13:13–22.
Mueller, D. P. and P. W. Cooper. 1986. Religious interest and involvement of young adults: A research note. Review of
Religious Research 27(3):325–45.
Myers, S. M. 1996. An interactive model of religiosity in inheritance: The importance of family context. American
Sociological Review 6:858–66.
Roof, W. C. 1978. Commitment and community. New York: Elsevier Press.
Roozen, D. A. 1980. Church dropouts: Changing patterns of disengagement and re-entry. Review of Religious Research
21(4):427–50.
Stolzenberg, R. M., M. Blair-Loy, and L. J. Waite. 1995. Religious participation in early adulthood: Age and family life
cycle effects on church membership. American Sociological Review 60:84–103.
Wilson, J. and D. E. Sherkat. 1994. Returning to the fold. Journal for the Scientific Study of Religion 33(2):148–61.
