THE IMPORTANCE OF RELIGION IN ADOLESCENTS ’ LIVES

R has shown that religion plays an important role in adolescents’ lives, positively impacting their academic performance, educational aspirations, worldview, and optimism about the future (Regnerus, Smith, & Fritsch, 2003). Being religious has also been associated with adolescent psychological well-being, positive self-concept, and good physical health (Donahue & Benson, 1995; Ellison, 1991; Oleckno & Blacconiere, 1991). Young people who are religious are less likely to engage in risky behaviors such as smoking, drugs, and alcohol use (Hays, Stacy, Widaman, DiMatteo, & Downey, 1986; Rohrbaugh & Jessor, 1975; Woodroof, 1985). Although researchers have investigated the relationship between religiosity and various outcomes, few studies have examined the mechanisms through which these relationships develop. Using data from an in-depth study of parents and their children, this paper examines the links among adolescents’ positive emotional and behavioral outcomes, religiosity in the home, and extracurricular participation at school.

R esearch has shown that religion plays an important role in adolescents' lives, positively impacting their academic performance, educational aspirations, worldview, and optimism about the future (Regnerus, Smith, & Fritsch, 2003).Being religious has also been associated with adolescent psychological well-being, positive self-concept, and good physical health (Donahue & Benson, 1995;Ellison, 1991;Oleckno & Blacconiere, 1991).Young people who are religious are less likely to engage in risky behaviors such as smoking, drugs, and alcohol use (Hays, Stacy, Widaman, DiMatteo, & Downey, 1986;Rohrbaugh & Jessor, 1975;Woodroof, 1985).Although researchers have investigated the relationship between religiosity and various outcomes, few studies have examined the mechanisms through which these relationships develop.Using data from an in-depth study of parents and their children, this paper examines the links among adolescents' positive emotional and behavioral outcomes, religiosity in the home, and extracurricular participation at school.

ADOLESCENTS AND RELIGIOSITY
Adolescence is a period of personal and religious identity formation.Between the ages of 13 and 18, teenagers begin to more explicitly articulate their sense of who they are, including how they relate to their parents and their peers (Csikszentmihalyi & Larson, 1984;Erikson, 1958).Adolescence is also the time when young people are perhaps most vulnerable with respect to their sexual identity, since they are adjusting to dramatic changes in their physical development (Csikszentmihalyi & Schmidt, 1998).While these emotional and physical changes are occurring, many adolescents are trying to clarify their spiritual beliefs, determine what values are important to them, and what moral principles should guide their behavior.Sometimes this can be difficult, especially because of societal and peer pressures that challenge family and religious beliefs.Adolescents often receive conflicting messages from their families, friends, and the media about what should be valued and how one should act.Increasingly, research suggests the importance of providing young people with opportunities to develop moral judgments and ethical behaviors that lead them to become responsible, caring, and civic-minded adults (Damon, 2002;Wilson, 2001).
Having a religious identity can help adolescents develop a more positive sense of self as they move through adolescence to adulthood.Religious identification can provide adolescents with an inner strength, often referred to as resilience, which can be especially helpful for coping with personal adversity.The stresses and pressures of peer groups often place adolescents in a precarious position, where they are presented with difficult choices.Religion can become a spiritual resource to draw upon when facing social problems, ambiguities with relationships, and personal, moral, and ethical decisions.
Belonging to a religious group, especially for adolescents who are becoming more socially and ethically conscious, can present opportunities for moral growth and personal responsibility.Religious participation, such as attending religious services, can be viewed as a form of social integration that reinforces values and attitudes upheld by the family and the religious community that one affiliates with, thus encouraging positive goals and behaviors (King & Elder, 1999;Regnerus & Elder, 2001).This seems particularly important since many public institutions such as schools have adopted a code of "moral relativism and ambivalence" (Damon, 2002, p. xii), where standards of honesty, fairness, compassion, and responsibility are rarely articulated or discussed.
The majority of teenagers report that they are religious to some extent (Regnerus et al., 2003;Smith, Denton, Faris, & Regnerus, 2002).Studies that examine religiosity among adolescents indicate that for some, being religious refers to direct participation in an organized religious group, whereas for others, it refers more generally to spirituality or their belief in a particular deity.Measures of how religious adolescents feel are also somewhat imprecise.Many studies that examine religiosity use survey items with response categories that typically range from not at all to very religious.When analyzing these responses, categories are often collapsed, making it difficult to distinguish those who are somewhat religious from others who are more or less religious.Regardless of what being religious means to different individuals, it is generally seen as important to the lives of American adolescents.
Religiosity can refer to a variety of factors, including religious beliefs, religious participation, and religious involvement in secular activities, such as volunteering with a church-sponsored youth group in a soup kitchen (Hill & Hood, 1999).Several scholars have attempted to define religiosity in terms of different aspects of religious commitment (Ellison, Gay, & Glass, 1989) or through a more social-psychological frame of religious identity (Horowitz, 1999).Dimensions of commitment tend to include personal faith, participation in organized religious activities, and identification with a particular religious denomination.Similarly, researchers emphasizing components of religious identity refer to the subjective assessment of spirituality in one's life, religious practice, and communal affiliation.The overlap between these two conceptions suggests that in defining religious commitment or identity, it is important to differentiate among ritual practice, religious affiliation, and a personal sense of one's religious beliefs.Merely examining participation or affiliation tends to underestimate the importance that religious identity has in an individual's life.

RELIGION AND THE FAMILY
For most adolescents, religious identity develops within the context of the family.Religion is a family activity, and religious participation is highest among families with school-age children (Stolzenberg, Blair-Loy, & Waite, 1995).It is therefore not surprising that religious practice and family processes are often associated with each other (Wilcox, 2001).Religious practices within the family can help parents guide their behaviors with each other and their children.In a recent study of adolescent religiosity, Schmidt (2003) finds that adolescents who report higher levels of religiosity, compared to those who consider themselves non-religious, perceive their families as more supportive and challenging.Support refers to family interactions that can be characterized as warm and caring, whereas challenge refers to interactions where children are encouraged to do their best.Family support and challenge have been shown to be indicators of "good parenting" that influence adolescent achievement and ambition (Csikszentmihalyi & Schneider, 2000;Schneider & Stevenson, 1999).It may be that being a good parent involves not only providing an environment that is supportive, nurturing, and goal-directed, but one that is influenced by beliefs deeply rooted within a particular religious tradition.Religious affiliation may strengthen a young person's sense of self by conferring a sense of belonging and group identity that is more stable than the fluidity of teenage peer groups (Schneider & Stevenson, 1999;Steinberg, Brown, & Dornbusch, 1997).
Within the family, mothers are more likely to engage in religious practices and to view themselves as religious than fathers.These behaviors carry over to their children, and mothers are found to directly influence their children's religious and moral development (Bao, Whitebeck, Hoyt, & Conger, 1999;Benson, Masters, & Larson, 1997).Schmidt (2003) found that mothers who report that they are religious in contrast to those who do not identify themselves as religious, are more likely to spend time with their children, have more frequent discussions of rules and values, and provide more supportive and challenging family environments.Additionally, adolescents who identify their mothers as being very religious tend to have a stronger sense of personal faith and well-being (Schmidt, 2003).
This research seems to suggest that parent religiosity influences adolescent religious identity.What is not clear is whether these parenting influences are the result of being religious or of being "good parents" who engage in positive behaviors with their children.Longitudinal data that examine religious identity through the life course may provide more definitive answers to this question (Elder & Conger, 2000).In examining adolescent religiosity, taking into account measures of positive and negative parenting styles, such as the amount of support and challenge provided to children, may help to distinguish religious effects from those of good parenting.

RELIGIOSITY, EDUCATION, AND PRO-SOCIAL BEHAVIOR
The links between religious practice, in the form of religious participation, and school-related outcomes, such as academic achievement, have tended to be modest (Muller & Ellison, 2001;Regnerus et al., 2003).It may be that religious participation is linked to other types of school-related behaviors that are enhanced by participating in extracurricular activities such as sports teams or service clubs.Such activities may facilitate the development of leadership, cooperation, and other pro-social behaviors.Increasingly, critics are urging public schools to become more active in promoting adolescent development through programs designed to enhance cooperation, responsibility, and caring for others.Beyond what is taught in the classroom, extracurricular activities are often seen as the appropriate place to teach students about fairness, sportsmanship, and ethnic and racial tolerance.Much like religion, participation in school-sponsored extracurricular activities has been shown to shape and form how adolescents define themselves (Guest & Schneider, 2003).Research indicates that students who participate in extracurricular activities are more likely to have a positive self-image (Eccles & Barber, 1999).It may be that participation in school-sponsored extracurricular activities also shapes pro-social and altruistic behaviors that are frequently attributed to religious participation.
The question of causality is inconclusive, however, for some adolescents may participate in extracurricular activities for reasons, such as portfolio-building for college, that are unrelated to their religious beliefs.Others may participate in activities because they are genuinely interested in helping others, altruistic endeavors that stem from parents' interests and values and that may be dependent or independent of family religious affiliation, practice, or spirituality.To understand the relationship between religiosity and adolescent development, it is important to examine school and community factors, independent of religion, that may be affecting adolescents' values, relationships to others, and sense of self-esteem.
The uplifting or hedonic experience of religion, that is, the enjoyment and heightened sense of self-esteem that accompanies certain types of religious practice, may also come from winning a sports game or singing in a chorus.It could be argued that the satisfaction of accomplishing one's objectives, whether that is performing, editing the school newspaper, or conducting an experiment, is stronger than feelings experienced during religious involvement.On the other hand, the happiness derived from religious practice or from a sense of spirituality may have a moral imperative that is fundamentally different from the hedonic experiences associated with more practical or worldly endeavors.If that is the case, religion may be filling an important void in the lives of young people, especially those who have few opportunities to become involved in school-sponsored activities.

SOME ASSUMPTIONS
This study uses data from the Alfred P. Sloan 500 Family Study, a national study of middle-and upper-income families, all of whom have children.Research indicates that religious participation is often tied to income and education and that families with limited resources are more likely to be religious than those with high family incomes and levels of educational attainment (Schmidt, 2003).Since families with school-age children are more likely to be religious, the dominant family structure of the Sloan sample may attenuate the effects of income and education on religious participation.In these families one would expect to find the following: Hypothesis 1: Religious participation and affiliation are likely to positively influence adolescent well-being as measured by pro-social and hedonic experiences, altruistic participation, and family duty.The social integration of religious life in the family with other activities, such as volunteering, is likely to positively influence adolescents' sense of themselves as being cooperative, responsible, hardworking, and productive.It is expected that the pro-social nature of these experiences carries over into many aspects of adolescents' lives.Such behaviors not only socialize young people to engage with others and to set goals and meet them, but also contribute to positive feelings about themselves.Religious participation may also contribute to adolescents' sense of social and family responsibility as exhibited through their involvement in household chores and altruistic activities.
Hypothesis 2: Positive parenting styles, characterized by high levels of support and challenge, are associated with adolescent well-being, pro-social experiences, and participation in family chores and altruistic activities.As shown in previous research, family support and challenge are positively related to adolescent well-being.It is expected that these same relationships to be evident in this population.One might expect that "good parenting" has positive effects on adolescent well-being and social behaviors independent of parents' religious participation.It is likely that parents with strong moral and ethical principles may also be religious, but this is not necessarily the case for all families.For those that are religious, the influence of religious participation on adolescent well-being may be realized through parents providing a challenging and supportive environment for their children.
Hypothesis 3: Participation in school-sponsored extracurricular activities is likely to be associated with measures of adolescent well-being.Extracurricular participation, much like religious practice, requires adherence to particular rules and self-discipline.During adolescence when young people are actively seeking new identities and understandings of themselves and their worlds, these activities, much like the consistent actions of their parents, help to set benchmarks of acceptable behavior and social responsibility.Families with religious beliefs may encourage their children to participate in extracurricular activities, considering such involvement as providing another venue for reinforcing values and prosocial activities.For those adolescents whose families are not religious, participation in extracurricular activities may promote pro-social behaviors and contribute a positive sense of self.

METHOD SAMPLE
Analyses in this paper are based on data from a subsample of parents and adolescents who participated in the Alfred P. Sloan 500 Family Study.The Sloan study is a national study examining parents, adolescents, and kindergarten children in middle-class, dual-earner families (Schneider & Waite, 2003).Families were recruited from eight middle-and upper-middle-class communities across the US; approximately 300 families have teenagers, 200 have kindergarteners, and 28 families have both adolescents and kindergarteners.For a more detailed description of the Sloan study (see Hoogstra, 2003).
Of the 300 adolescent families, 257 had teenagers who had completed both the adolescent survey and the Experience Sampling Method (ESM).The survey included several items about religious participation and school life, whereas the ESM explored the emotional experiences of young people at school and in religious activities.Only teenagers who had completed both instruments were selected for analysis.Additionally, because the analyses incorporate some variables from the parent surveys, it was necessary to exclude adolescents whose parents had missing data on key variables used in analyses.The final sample includes 225 adolescents who had complete ESM and survey data, and whose parents had completed all of the survey items of interest.Several analyses were undertaken to ensure that the remaining sample was not different from the excluded sample on several of the outcome and predictor variables.Independent-sample t-tests and chi-square tests were performed between the included and excluded samples for the outcome and predictor variables used in the analyses, along with several demographic variables such as age, gender, family income, parent educational attainment, college aspirations, and grades.No significant differences were found between the included and the excluded sample on these variables (all t's < 1.93, χ 2 < .02).
The final sample includes 108 males and 117 females, whose average age is 15 years and 5 months (males are slightly older than the females: males = 15.6;females = 15.2).Ethnically, the sample is fairly homogenous; 81% of adolescents identify themselves as White, non-Hispanic.Most of the parents have completed college, and many of them have earned master's or professional degrees.Education and academic achievement are highly valued by the parents and their children.Approximately 70% of the mothers expect their children to continue their education after receiving a college degree.Even taking grade inflation into account, the grades of these adolescents are quite high, with the average student receiving mostly A's and B's on his or her most recent report card.All of the teenagers in the study attend public schools.

MEASURES
As previously noted, the measures for this study were taken from the adolescent and parent surveys and the adolescent responses to the Experience Sampling Method (ESM).Both surveys include a variety of items that have been used in other studies such as the General Social Survey and the National Education Longitudinal Study of 1988-92 that ask about religious participation and school experiences.Survey items measuring parenting style are based on similar items in the ECLS-Kindergarten Spring Parent Interview Questionnaire (http://nces.ed.gov/ecls/ kindergarten/questionnaires.asp) and the Child's Report of Parent Behavior Inventory (Shaefer, 1965).Questions about basic demographic characteristics of the respondents and family activities were also included in the surveys.
The ESM, created by Csikszentmihalyi and colleagues (Csikszentmihalyi & Larson, 1987), is a form of time diary that examines the daily lives of individuals, including what they are doing and how they feel.The ESM was administered to both parents and adolescents; only the responses of the adolescents are of interest here.The method involves preprogramming wristwatches that beep randomly throughout the day.In this study, adolescents were beeped eight times a day over the course of a week.When beeped, adolescents report the time of day, place, activity, who they are with, and how they feel at the time they are signaled.Out of a possible 56 responses, adolescents in the sample responded to 32.5 beeps on average.In relation to other ESM studies and other forms of time diaries, the average response rate is considered to be robust and comparable to other national surveys (Mulligan, Schneider, & Wolfe, 2002).

Dependent Measures
Four outcome measures are developed for this study, two of which focus on involvement in positive social experiences-pro-social experience and altruismone that measures positive feelings-hedonic experience-and one that captures involvement in family life-family duty.The pro-social construct includes items from the ESM that measure whether the adolescent feels cooperative, responsible, hardworking, productive, and believes that he or she is meeting the expectations of others.The altruism construct, derived from the survey, includes how often the adolescent and his or her parents are involved in volunteer and charity work, how much time the adolescent spends in volunteer work or community service outside of school, whether the adolescent participates in a service club at school, and if he or she has won an award for community service.Hedonic experience, created from the ESM, is a "feel good" measure constructed from items measuring how happy and cheerful adolescents feel, how much they enjoy the activity and feel that they are succeeding, and the extent to which they feel good about themselves.For the two measures constructed from ESM items (pro-social and hedonic experience), a composite variable was computed at the "beep" level that was an average of the ESM items.This composite variable was then aggregated to the individual level to obtain an overall value for each individual.Finally, family duty was constructed as a sum of two measures from the survey that examine how much time adolescents and their parents work on household chores together and how often the adolescent spends time on household tasks, including shopping, babysitting younger siblings, cooking, washing dishes, doing laundry, and yard and home maintenance.

Independent Measures
An affirmative association is expected between positive feelings and behaviors experienced by adolescents and religious practice by the mother and the adolescent.Child's religious practice is based on survey items that examine how often the adolescent and one or both parents attend religious services together, how much time the adolescent spends in religion or heritage classes outside of school, and how often the adolescent attends religious services.Mother's religious practice is based on a similar set of items.
As described earlier, how mothers and fathers parent may also affect adolescent well-being.Therefore a measure of parenting style was created from 10 survey items that include parents' evaluations of statements regarding the quality of parent-child interactions and how the reality of parenting fits with their original expectations of parenting.Two other parenting variables, family support and family challenge, are also used to examine adolescent family interactions.Support refers to feelings of acceptance, understanding, and love that adolescents attribute to their parents.Challenge refers to the goals, expectations, and ambitions that adolescents perceive their parents hold for them.Support and challenge have been used in several studies of family interaction and were found to be related to adolescent ambitions, academic performance, and well-being (Csikszentmihalyi & Schneider, 2000;Schneider & Stevenson, 1999).C.
The last two measures are derived from school-based survey items.Extracurricular participation was created from nine survey items regarding the adolescent's participation in school activities throughout the school year.The proportion of graduating seniors attending 4-year colleges comes directly from the high schools attended by the adolescents.This measure is used as an indicator of a high school's orientation to academic achievement.It is expected that schools with a stronger emphasis on academics will also be more likely to encourage students to participate in extracurricular activities that promote social service.

Analytic Approach
Three separate analytic strategies are employed to understand what factors are associated with pro-social behaviors and emotional well-being among adolescents.The first set of analyses investigates differences in adolescents' pro-social and hedonic experiences when they are engaged in religious thoughts or activities, versus other activities, by religious affiliation.The second set of analyses examines the influence of individual, family, and school factors on the adolescent wellbeing and pro-social behaviors, using a series of ordinary least squares regressions.These analyses are conducted for all adolescents in the sample as well as for Catholic and Protestant students.Finally, to separate and quantify individuallevel from family-level effects, two-level hierarchical linear models (HLM) are constructed for two of the outcome measures: family duty and altruistic participation.

RESULTS
To determine whether adolescents' emotional experiences are more positive during religious activities, initial analyses examine the pro-social and hedonic experiences of adolescents when engaged in religious activities or when thinking about religion and spirituality compared with their emotions during all other activities.Results of independent-sample t-tests show significant differences between "religious" and "non-religious" activities for both pro-social and hedonic experience (t (9066) =-3.169, p < .01 and t (8879) = -3.449,p < .01,respectively).Adolescents had more positive pro-social and hedonic experiences when they were engaged in religious activities than when engaged in all other activities.For pro-social experience, the mean during religious activities is 1.29 compared to 1.11 for non-religious activities.For hedonic experience, the mean for religious activities is 2.09 compared to 1.91 for other activities.
The next set of analyses examines whether these kinds of experiences, along with the other outcome measures, religious practice in the home, parenting styles, and extracurricular participation, differ by religious affiliation.Within the sample, there are five major religious groups: Roman Catholic, Protestant, Other Christian, Jewish, and Eastern and one category for those teenagers who reported no religious affiliation.Table 1 lists the means and standard deviations.
To test whether there were significant differences among these groups, individual one-way analysis of variance (ANOVA) tests were performed on each variable by religious affiliation.There are significant main effects of religious affiliation for three of the variables: hedonic experience, family challenge, and mother's religious practice.Post-hoc analyses (with Bonferroni corrections) indicate that the main effect for hedonic experience is driven by the difference between the Eastern religion and "no religion" categories.Adolescents in the Eastern religion category have significantly higher values for hedonic experience than those in the "no religion" category.However, there were no differences among religious affiliations for family challenge once Bonferroni corrections were made.Results for mother's religious practice appear to rest on the difference between adolescents who are affiliated with a religion and those who have no religious affiliation.(Analyses not shown.) Additional comparisons were made between those who identified themselves as having no religious affiliation and the rest of the sample.Independent t-tests resulted in significant differences between these two groups for child's religious practice and mother's religious practice (t (218) = 7.29, p < .001and t (223) = 6.90, p < .001,respectively).These significant differences are not unexpected given that an individual with no religious affiliation would be unlikely to participate in religious activities.Significant differences are also found between the religious and non-religious adolescents with respect to pro-social and hedonic experiences (t (223) = 2.43, p < .05 and t (223) = 2.46, p < .05,respectively).Adolescents who are affiliated with a particular religion have significantly higher pro-social and hedonic experiences than those who have no religious affiliation.
To further explore the relationships among outcome and predictor variables, a series of bivariate correlations were conducted (see Table 2).
For the most part, the variables are not highly correlated with the exceptions of family support and challenge (r =.770) and between mother and adolescent's religious practice (r =.790).With respect to support and challenge, subsequent analyses were conducted to disaggregate these variables into more meaningful measures: low challenge-low support; high support-low challenge or high challenge-low support; and high challenge-high support.These three categories were created by taking responses to each measure that fell above and below the sample mean.In the following regression models, the low support-low challenge group is the referent category.With respect to religiosity, only mother's religious practice is used in analyses to avoid multicollinearity.

UNDERSTANDING RELATIONSHIPS BETWEEN RELI-GIOSITY AND POSITIVE EMOTIONS AND WELL-BEING
To investigate the relationship between family religiosity and the four outcome measures, four sets of ordinary least squares regression analyses were conducted.The standard model for all four of the outcome measures includes three familylevel variables-parenting style, the support and challenge, and mother's religious practice, and two school-level variables-extracurricular participation and proportion of seniors going to a 4-year college.For each outcome variable, the first set of analyses is conducted on the full sample of 225 teenagers and then two partial regressions are conducted separately for those adolescents who identified themselves as Catholic or Protestant.Ideally, regressions would also have been performed on the Eastern religion group (the group whose means were most similar to both the Protestant and Catholic groups) and on those who identified themselves as having no religious affiliation.However, the limited sample sizes of these two groups prevented conducting regressions that would produce valid and robust estimates.
Age, gender, and other demographic variables are not included in the analyses.Regressions were initially performed with these control variables, and their inclusion did not significantly change the results of the models.Because the sample size is limited, especially when examining the religious affiliations separately, these variables are excluded from all subsequent models.

Pro-Social Experience
As shown in Table 3, the first regression model examines the relationship between individual, family, and school influences and pro-social behaviors-that is, the extent to which adolescents feel hardworking, responsible, cooperative, and productive.For this analysis, hedonic experience, altruism, and family duty are included at the individual level, since it is expected that these feelings and behaviors are related to pro-social behaviors.For adolescents, there is a significant positive association between hedonic experience and pro-social activities, (i.e., activities that evoke feelings of cooperation, responsibility, and productivity).For every unit increase in hedonic experience, there is nearly a half unit increase in pro-social experience.Similarly, the positive association between family duty and pro-social experience is not unexpected.An adolescent who more frequently performs household chores, is more likely to feel hardworking and responsible, whereas the relationship between altruism and pro-social although marginally significant may stem from a different motivation.
Of the family measures, only mother's religious practice is significantly associated with pro-social experience for the full sample.A one unit increase in religious practice is associated with approximately one-eighth of a unit increase in pro-social experience.Stepwise regression analysis of pro-social experience revealed a potential interaction between mother's religious practice and altruism: the coefficient for altruistic participation was reduced when mother's religious practice was included as a predictor.For the Catholics and Protestants, however, there are no family variables that are significantly associated with pro-social experience.
Extracurricular participation has a significant positive association with prosocial experience for the full sample.The effect of extracurricular participation is similar in magnitude to that of mother's religious practice.Secondary regressions that were performed with the school-level variables by themselves resulted in a larger coefficient for extracurricular participation than is seen in the full model.Examination of the correlations between extracurricular participation and the other adolescent measures reveals a strong positive correlation with altruism, which may contribute to this reduction.As was the case with the family measures, for the Catholic and Protestants there are no significant school-level predictors associated with pro-social experience.
For the Protestant adolescents, there is a similar pattern of association for both hedonic experience and family duty.However, whereas the hedonic experience coefficient is similar in magnitude to that of the overall sample, the coefficient for family duty is over four times larger for the Protestant students than it is for the full sample.This strong, positive association between family duty and pro-social experience may reflect the Protestant work ethic; that is, Protestant adolescents who perform household chores more frequently are more likely to feel responsible, hardworking, productive, and cooperative.This idea gains support when the coefficient for family duty is examined for Catholics.Here, there is no association between family duty and pro-social experience.In fact, for the Catholic teenagers hedonic experience is the only experiential variable that is even marginally associated with pro-social experience, although the coefficient is in the same direction as it is for the full sample and for the Protestant adolescents.

Family Duty
To examine the influence that individual, family, and school variables have on adolescents' level of familial responsibility, the same series of models were performed with family duty as the outcome variable.For this model, hedonic experience, pro-social experience, and altruism are entered as individual measures.As Table 4 indicates, there are few significant associations between the predictors and family duty for the full sample, suggesting that the model is incomplete.Only pro-social experience has a positive, significant association with family duty.A one unit increase in pro-social experience is associated with one-sixth of a unit increase in family duty.
The models for the Catholic and Protestant groups are stronger.For the Protestant adolescents, both hedonic experience and pro-social experience are significantly associated with family duty, but these associations are in opposite directions.Whereas Protestant teenagers show a strong positive association between pro-social experience and family duty, there is a negative association between hedonic experience and family duty.Feelings of responsibility, cooperation, and hard work are positively associated with performance of household chores, but Protestant adolescents who spend more time engaging in these tasks do not enjoy, feel happy, or feel good about themselves.This is not the case for Catholics.There is no association between pro-social experience and family duty for Catholic adolescents, and in contrast to Protestants, there is a positive, though marginal association between hedonic experience and family duty.Of the family measures, mother's religious practice has a significant, positive association with family duty for Catholic adolescents.For every unit increase in mother's religious practice there is a 44% unit increase in family duty.Catholic students whose mothers frequently engage in religious practices are thus more likely to perform household chores.

Hedonic Experience
The next model examines associations between individual, family, and school variables and hedonic experience (feeling good, cheerful, happy, and enjoying what one is doing).Results show that pro-social experience is the only individual measure significantly associated with hedonic experience for the full sample.A one unit increase in pro-social experience is associated with a two-fifths unit increase in hedonic experience.
As Table 5 demonstrates, several family measures have significant associations with hedonic experience for the full sample.Parenting style, high support or high challenge, and high challenge-high support are all positively associated with hedonic experience.
If adolescents have parents who have a more positive parenting style, they experience a one-seventh unit increase in hedonic experience.Additionally, if adolescents are in a family environment that is either supportive or challenging (compared to adolescents in a low support-low challenge environment) they experience nearly the same increase in hedonic experience as is seen with a positive parenting style.However, if adolescents are in a family environment that is both supportive and challenging, they experience an increase in hedonic experience that is nearly twice that of either a positive parenting style or a family environment that is high in support or high in challenge.Therefore, being in an environment where the parents are supportive, challenging, or more positive in their interactions with adolescents is associated with positive adolescent well-being; however, if parents provide a family environment that is not only high in support, but also high in challenge, the adolescents have even stronger emotional experiences of overall well-being.
For the Catholics, the pattern for the family measures is similar to that for the full sample, although the coefficients are much larger in magnitude.For example, Catholic adolescents who have parents with a positive parenting style see an association with hedonic experience that is nearly twice that of the overall sample.Additionally, if these adolescents are in family environments that are both supportive and challenging (compared to Catholic adolescents in families that are low in support and challenge), they see a nearly half-unit increase in hedonic experience.For Catholic students, there are also marginal effects for high support or high challenge and mother's religious practice.Again, the magnitude of the coefficient for high support or challenge is roughly half the value of the coefficient for a family environment that is both high in support and challenge and is nearly twice the value of the coefficient for high support or challenge for the full sample.
The marginal negative association between mother's religious practice and hedonic experience is interesting.A one unit increase in mother's religious practice is associated with a nearly one-third unit decrease in hedonic experience.Potentially this association is indicative of the discipline that is often associated with a strong religious practice, especially among those of the Catholic faith where rituals associated with religious attendance and atonement are strongly encouraged.It is possible that this strong discipline is associated with lower levels of enjoyment and happiness.
For the Protestants, both pro-social experience and family duty are significantly associated with hedonic experience.For these adolescents, however, there is a larger coefficient for pro-social experience.A one unit increase in pro-social experience is associated with a three-fifths unit increase in hedonic experience.Similar to the results for the regression for family duty, there is a negative association between family duty and hedonic experience.Protestant adolescents who participate more frequently in family chores appear to experience less happiness, enjoyment, and feel less good about themselves; a one unit increase in family duty is associated with over a one-third unit decrease in hedonic experience.
For Protestant adolescents, there are no significant associations between family measures and hedonic experience.There is a marginal, positive association between a family environment that is high in support and high in challenge (compared to a Protestant family environment that is low in support and challenge) and hedonic experience.Also, although not significant, the magnitude and direction of the coefficient for a high support or high challenge family environment is similar to that seen for the full sample.It appears that feelings of general well-being are more strongly associated with individual experiences for Protestants, whereas for Catholics these feelings are more directly related to family interactions.

Altruistic Participation
The final series of regression models identifies the factors that are associated with social responsibility.As Table 6 reports, there are no significant individual-level variables associated with altruistic participation.Only pro-social experience has a marginal, positive association with adolescents' participation in altruistic activities.There are also no significant associations between individual-level variables and altruistic participation for either the Protestant or Catholic students.Both groups of adolescents have coefficients for pro-social experience that are slightly larger than that seen for the full sample, but neither is significant.For Catholics, family duty has a marginal positive association with altruistic participation.Again, this could be tapping into an overall sense of responsibility that strengthens both family responsibility and social responsibility.
Of the family measures, only mother's religious practice has a significant association with altruistic participation.A one unit increase in religious practice is associated with a nearly one-fifth unit increase in altruistic participation.If adolescents have mothers with higher levels of religious practice, they are more likely to participate in altruistic activities.As was the case for the individual-level variables, there are no family measures variables that are significantly associated with altruistic participation for either the Protestant or Catholic groups.The Protestant adolescents have a marginal, positive association between a high sup-port or high challenge family environment and altruistic participation.Compared to Protestant adolescents in a low support-low challenge environment, those with either high support or high challenge have an almost one-third unit increase in altruistic participation, although this effect is marginal.
A large portion of the variation in altruistic participation is explained by school measures.For the full sample, both extracurricular participation and proportion of students going to a 4-year college are significantly associated with altruistic participation.For every unit increase in extracurricular participation, adolescents experience an almost one-fourth unit increase in altruistic participation.Also, for every unit increase in the proportion that goes to a 4-year college, there is an associated one-fifth unit increase in altruistic participation.Therefore, schools that encourage both extracurricular participation and that focus on preparation for 4-year colleges increase the likelihood that students will engage in altruistic behaviors.Although the school-level coefficients for Catholic students are not significant, they are similar in magnitude to those seen for the full sample.For Protestant students, there is a large positive association between going to a 4-year college and altruistic participation.For every unit increase in the proportion of students at the school who go to a 4-year college, there is a two-fifths unit increase in altruistic participation.This effect combined with the marginal effect of a high support or high challenge family, supports the idea of a strong work ethic.Protestant students who are in a supportive or challenging family environment or who are in a college preparatory school environment are more likely to participate in altruistic endeavors.

DISTINGUISHING INDIVIDUAL AND FAMILY INFLUENCES: HIERARCHICAL LINEAR ANALYSIS
Results from the regression models examining the four outcome measures suggest that family duty and hedonic experience show the strongest differences in the influence of individual versus family effects for Protestants and Catholics.This was not the case for pro-social experience, where most of the variation is explained by individual-level variables.To further investigate the different contributions of individual and family experiences, two-level Hierarchical Linear Modeling (HLM) was used.The initial unconditional model for hedonic experience did not indicate sufficient variance to justify a two-level model.However, the unconditional model for family duty did have sufficient variance to warrant further analysis.
In the altruistic regression models, the primary associations appear to be at the family and school levels.However, because there were not enough cases to conduct a three-level HLM, students' participation in school-sponsored extracurricular activities, which is based on students' survey responses, was entered as a level-one variable and family variables were entered at level two in a two-level HLM, making it possible to examine the different contributions of school and family experiences.For all multi-level analyses, the statistical package HLM (Raudenbush, Bryk, & Congdon, 2000) was used.Model development followed standard protocol by beginning first with the simplest model (unconditional) followed by the continual addition of predictors at both level one and level two based on results from previous models.

Family Duty
The best fitting HLM model for family duty incorporates pro-social experience (grand mean centered) and dummy variables identifying Catholic and Protestant affiliations (uncentered) at level one and mother's religious practice (grand mean centered) on the Catholic and Protestant slopes at level two (see Table 7).The variance component for family duty in the unconditional model was significant, which justified the addition of predictors: χ 2 = 275.91,p < .001.The level-one coefficient for the intercept (2.43) is significant (p < .001),as is the coefficient for pro-social experience (0.48, p < .05).Therefore, a one unit increase in pro-social experience increases the coefficient for family duty from 2.43 to 2.91.The Catholic variable is marginally significant at the intercept (-0.44, p = .058).Compared to all other students, Catholic students see a marginal decrease in the family duty coefficient from 2.43 to 1.99.However, religious practice is significant at level two on the Catholic slope (.31, p < .01).For Catholic students, as mother's religious practice increases by one unit, the coefficient for family duty changes from 1.99 to 2.30.Therefore, Catholic students whose mothers more frequently engage in religious practices see a positive increase in family duty.The Protestant variable is not significant at level one, nor is mother's religious practice significant for Protestants at level two.Therefore, the results of the HLM provide support for the regressions presented in Table 4. Overall, there are significant associations between family duty and student-level variables, but for Catholic students, family duty gains more from the family level in the form of mother's religious practice.

Altruistic Participation
Mother's religious practice and extracurricular participation are both strongly associated with adolescents' degree of participation in altruistic activities.The best fitting HLM model for altruism includes extracurricular participation (grand mean centered) at level one and religious practice (grand mean centered) on the intercept at level two (see Table 7).(The variance component for altruistic participation in the unconditional model was significant (χ 2 = 233.33,p < .01),which provided justification for further model development.)In this model, the influence of extracurricular participation and mother's religious practice are separated between the level-one and level-two coefficients.At level one, both the intercept (1.13) and extracurricular participation (0.19) are significant at p < .001.A one unit increase in extracurricular participation is associated with an increase from 1.13 to 1.32 in the coefficient for altruistic participation.At level two, the coefficient for mother's religious practice is significant (0.07, p < .05).Therefore, a one unit increase in mother's religious practice is associated with a change in the coefficient for adolescents' altruistic participation from 1.13 to 1.20.With this analysis, it is apparent that extracurricular participation has an effect that is nearly three times that of the family-level variable of mother's religious practice on altruistic behavior.This result seems to suggest that participation in school-sponsored activities can promote positive social behaviors independent of family religious practices and parenting style.

DISCUSSION
The primary goal of this paper was to examine individual, family, and school influences on adolescent well-being, pro-social experiences, and family and social responsibility.Initially, it was expected that religious practices in the home would positively influence these adolescent feelings and behaviors.However, this appeared not to be the case for all the measures.Mother's religious practice was significantly and positively associated with pro-social experience and altruistic participation but not family duty or hedonic experiences.The influence of family religiosity on adolescent behaviors and emotional well-being is complex and varies by affiliation for this limited sample.As others have found, the relationship between religious practices and adolescent outcomes is positive but not pronounced.The nuanced effects of religion can be seen in some of the results reported in this paper.
For example, among Protestant adolescents, feeling productive, hardworking, and responsible is strongly related to performing family chores; but Protestant adolescents who more frequently perform chores report lower levels of happiness, enjoyment, and self-esteem.For Catholics, there is no significant relationship between family chores and pro-social experiences, and, compared to Protestants, Catholic adolescents have much more positive feelings about themselves when engaged in family chores, although this relationship is marginal.In addition, mother's religiosity is strongly associated with family duty for Catholic adolescents but not for Protestants.
Overall, school influences are most strongly associated with altruistic participation, and this effect is even stronger for Protestant adolescents than it is for the full sample.For Catholics, the direction of the association is also positive but not significant.Although mother's religiosity is significantly associated with altruistic participation in the full sample, the effects are considerably weaker than the school measures.Moreover, mother's religious practice is not significant for Catholic or Protestant adolescents.The findings regarding the influence of school measures on altruistic participation are further strengthened by the HLM results.The HLM analysis demonstrates that the influence of extracurricular involvement on altruistic participation is nearly three times that of mother's religious practice, suggesting that participation in school-related activities may fill a void for adolescents, particularly if families are not religious.Therefore, a school that emphasizes student participation in extracurricular clubs and activities could aid in the development of altruistic behaviors for adolescents, even if those teenagers do not come from families with a strong religious emphasis.
Family influences are particularly strong for predicting adolescent wellbeing.Overall, adolescents experience a more positive sense of self when their families are supportive and challenging.For the Catholic adolescents, these influences are more pronounced than they are for the full sample; however, they are not significant for the Protestant students.These results are consistent with ideological and historical differences among Protestants and Catholics.Protestants appear to be more influenced by individual experiences, whereas Catholics appear to be more family-oriented, and are more influenced by family religious practices and parenting styles.Thus, it may be that one of the reasons that it is so difficult to find strong effects for religion is because these effects vary by religious affiliation, and small sample sizes make it difficult to examine differences in behaviors and beliefs by religious affiliation.
These results are based primarily on middle-class students and students from two prominent religious backgrounds.Research that incorporates a wider variety of religious backgrounds and a more racially and ethnically diverse sample could provide a more nuanced understanding of the relationships among religious, family, and school influences on adolescent pro-social behavior and emotional wellbeing.In addition, there may be some family interactions that are not captured by the measures used in this paper.There are potentially interactions within the family and household, such as parent-child activities, communication styles, family dynamics, that contribute to both pro-social behaviors and overall well-being, but are not adequately captured by the measures used.
What the findings suggest is that family religiosity may serve as a conduit for the transmission of goals and values, some that are religious, others that promote ethical and moral behavior in other domains.The structure of religious life may provide young people with boundaries and expectations of behavior that act as countervailing influences to adolescent peer groups and images projected through the media.This may be the case especially for Catholic adolescents, the majority of whom now attend public schools in contrast to several decades ago when most attended private religious schools (National Center for Education Statistics, 1996).Religious practice in Catholic families may receive heightened importance for the parents as they attend to their children's moral and social development.But the findings also show that there are other influences, such as participation in extracurricular activities, which also can promote pro-social behaviors, altruism in particular.These results support the recommendations by Damon (2002) and others that schools need to promote character development and provide young people with opportunities to exercise moral judgment, ethical behavior, and care and concern for others.
Barbara Schneider is co-director of the Alfred P. Sloan Center on Parents, Children, and Work at NORC and the University of Chicago.Holly Rice is the data manager and a research analyst, and Lisa Hoogstra is the director of research services and a senior research analyst for the Alfred P. Sloan Center on Parents, Children, and Work.Correspondence concerning this article should be addressed to Dr. Barbara Schneider, University of Chicago, 1155 E. 60 th Street, Room 369, Chicago, IL 60637.