Role-modeling and conversations about giving in the socialization of adolescent charitable giving and volunteering.

Journal of Adolescence 37 (2014) 53–66

Contents lists available at ScienceDirect

Journal of Adolescence journal homepage: www.elsevier.com/locate/jado

Role-modeling and conversations about giving in the socialization of adolescent charitable giving and volunteering Mark Ottoni-Wilhelm a, b, *, David B. Estell c,1, Neil H. Perdue d, 2 a Department of Economics, Indiana University Purdue University Indianapolis (IUPUI), 425 University Boulevard, Indianapolis, IN 46202, USA b Indiana University Lilly Family School of Philanthropy, 550 West North Street # 301, Indianapolis, IN 46202, USA c Department of Counseling and Educational Psychology, Indiana University Bloomington, 201 North Rose Avenue, Bloomington, IN 47405-1006, USA d School of Psychological Sciences, University of Indianapolis, 1400 East Hanna Avenue, Indianapolis, IN 46227, USA

a b s t r a c t Keywords: Prosocial behavior Charitable giving Volunteering Positive behavior Modeling Socialization

This study investigated the relationship between the monetary giving and volunteering behavior of adolescents and the role-modeling and conversations about giving provided by their parents. The participants are a large nationally-representative sample of 12–18 yearolds from the Panel Study of Income Dynamics’ Child Development Supplement (n ¼ 1244). Adolescents reported whether they gave money and whether they volunteered. In a separate interview parents reported whether they talked to their adolescent about giving. In a third interview, parents reported whether they gave money and volunteered. The results show that both role-modeling and conversations about giving are strongly related to adolescents’ giving and volunteering. Knowing that both role-modeling and conversation are strongly related to adolescents’ giving and volunteering suggests an often overlooked way for practitioners and policy-makers to nurture giving and volunteering among adults: start earlier, during adolescence, by guiding parents in their role-modeling of, and conversations about, charitable giving and volunteering. Ó 2013 The Foundation for Professionals in Services for Adolescents. Published by Elsevier Ltd. All rights reserved.

Educators, community leaders, and policy-makers expend considerable effort to encourage adults to perform two particular prosocial behaviors – giving money to charities that help people and giving time in voluntary service. Adult giving and volunteering are most likely part of an ongoing pattern of behavior that began much earlier in development, and both theory and laboratory-based experimental work suggest that role-modeling and conversations about prosocial behavior will affect the development of prosocial behavior (Eisenberg & Fabes, 1998). Yet we know very little about the effectiveness of parental efforts to encourage these behaviors in nationally-representative population studies that describe U.S. adolescents in their natural settings. This represents a significant gap in our understanding. As Eisenberg and Mussen (1989, p. 156) point out: “It cannot be assumed that procedures that prove to be effective in laboratory studies (such as modeling; see chapters 6 and 7) will necessarily have a significant and lasting impact on behavior when introduced in a natural setting such as

* Corresponding author. Department of Economics, Indiana University Purdue University Indianapolis (IUPUI), 425 University Boulevard, Indianapolis, IN 46202, USA. Tel.: þ1 765 277 1229 (mobile). E-mail addresses: [email protected] (M. Ottoni-Wilhelm), [email protected] (D.B. Estell), [email protected] (N.H. Perdue). 1 Tel.: þ1 812 856 8308. 2 Tel.: þ1 317 788 3353. 0140-1971/$ – see front matter Ó 2013 The Foundation for Professionals in Services for Adolescents. Published by Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.adolescence.2013.10.010

54

M. Ottoni-Wilhelm et al. / Journal of Adolescence 37 (2014) 53–66

the home or school.. Thus, it is important to test in natural settings the effectiveness of those procedures that promote prosocial behavior in the laboratory.” There are no nationally-representative studies of the effectiveness of parents’ role-modeling of charitable giving, and although previous research has found evidence of an association between parental modeling of volunteering and adolescents’ volunteering (McLellan & Youniss, 2003), including a nationally-representative study (U.S. Department of Education, 1997), these studies could not control for conversations the parent may be having with the adolescent about prosocial behavior. In short, to our knowledge no previous study has estimated the strengths of role-modeling and conversations about prosocial behavior as separate influences on adolescents’ giving and volunteering. Furthermore, no study has previously estimated role-model and conversation associations while simultaneously controlling for parenting dimensionsdincluding parental warmth/support and behavioral controldthat theoretically are influences on prosocial behavior (Chase-Lansdale, Wakschlag, & Brooks-Gunn, 1995). Finally, little is known about whether the effectiveness of modeling, conversation, and parenting dimensions differs by the sex of the adolescent (however see Carlo, McGinley, Hayes, Batenhorst, & Wilkinson, 2007). Without this knowledge it is difficult to convince practitioners and policy-makers to shift some of their effort toward the encouragement of parents in their role-modeling and talking about giving and volunteering, thereby nurturing the giving and volunteering of tomorrow’s adults. Moreover, it is not known whether any advice to parents should be tailored differently depending on the sex of the adolescent. To address these gaps, the present research estimates the relative strengths of rolemodeling and conversations about prosocial behavior as separate influences on adolescents’ charitable giving and volunteering using a large nationally-representative sample. These effects are estimated while simultaneously assessing the association of parental warmth and support and parental behavioral control. Parenting, socialization, and prosocial behavior Parenting and prosocial behavior Theory suggests parenting is related to prosocial behavior. Baumrind (1971) distinguished three broad patterns of parenting based on levels of warmth/support and control. Authoritative parents are warm and provide appropriate rules and guidance. Authoritarian parents are less warm and tend to be directive in their control. Indulgent parents may be warm but provide few limits on their children. Maccoby and Martin (1983) built on this framework and identified four typologies of parents: Authoritative (high warmth and guidance-based control), authoritarian (low warmth, highly directive control), indulgent or permissive (high warmth, low control), and neglectful (little warmth or control). Authoritative parenting is thought to be an effective style for the development of prosocial behavior because it encourages empathy via modeling empathy, warmth, and respect for individuals, along with expectations of age-appropriate personal responsibility that include behaving prosocially toward others (Chase-Lansdale et al., 1995). Authoritative parents use reason-based explanations for their disciplinary practices, and this in turn promotes perspective taking and empathy in the child (Hoffman, 2000; Knafo & Plomin, 2006). The processes of respecting one’s offspring as an individual and granting increasing autonomy as the adolescent takes responsibility in turn promote the development of a strong identity that includes treating others with warmth and respect (Chase-Lansdale et al., 1995; Wuthnow, 1995). As such, authoritative parenting promotes both prosocial behavior in the near term as well as internalization of the value of prosocial behavior in the long term (Hastings, Utendale, & Sullivan 2007). While authoritative parenting is most closely associated with the development of prosocial behavior and, due to lack of either warmth or structure, neglectful parenting can be safely assumed to be least likely to promote prosocial behavior, it is less clear how systematically the other two parenting styles relate. As Hastings et al. (2007, pp. 643–644) argue, authoritarian parenting may make a child less prosocial by “modeling a lack of concern for the needs of others” and eliciting the child’s rejection of socialization attempts, but can be associated with prosocial behavior depending upon the disposition of the child and context of the socialization. Similarly, permissive behavior may be related to prosocial behavior due to modeling warmth, but the lack of enforced behavioral standards may inhibit long-term internalization of this as a norm. In short, authoritative parenting is most likely to lead to internalized, independent prosocial behavior, neglectful parenting least likely to foster the development of such behavior, but the other two styles may or may not lead to prosocial behavior depending on the child and the context. Supporting evidence has been found associating authoritative parenting to adolescent social responsibility (Gunnoe, Hetherington, & Reiss, 1999) and prosocial behavior (Baumrind, 1991; Knafo & Plomin, 2006; Padilla-Walker, Carlo, Christensen, & Yorgason, 2012). Other research examined the dimensions of these typologiesdwarmth and controldand their association with adolescent behavior. These dimensions are associated with empathy and perspective taking (e.g., Laible & Carlo, 2004; Soenens, Duriez, Vansteenkiste, & Goossens, 2007), children’s internalization of prosocial values (Hardy, Padilla-Walker, & Carlo, 2008; also see Hardy, Bhattacharjee, Reed, & Aquino, 2010), and prosocial behavior (Carlo, Mestre, Samper, Tur, & Armenta, 2011). While warmth and controldeither as separate constructs or in parenting typologiesdboth may relate to a variety of child outcomes including prosocial behavior, recent research has indicated that different forms of control have unique effects on children. Psychological controldthe use of manipulation, guilt, coercion, and contingent love, for exampledis an important construct that needs to be examined apart from behavioral control (Barber, 1996). Psychological control is related to lower


55

levels of empathy, teacher-reported altruism, and a variety of prosocial behaviors in children such as sharing and helping with classmates (Krevans & Gibbs, 1996). This indicates that psychological control may undermine any positive association of warmth and/or behavioral control on prosocial behavior, though more research is needed examining these links. While parenting and prosocial behavior are linked, there are mixed findings about the strength of the association. For example, some studies have not found an association (see the reviews in Eisenberg, Fabes, & Spinrad, 2006 and Eisenberg, Morris, McDaniel, & Spinrad, 2009; also Soenens et al., 2007 found supporting evidence for maternal warmth/support, but not behavioral control) while other studies have found indirect, but not direct, effects (e.g., Barry, Padilla-Walker, Madsen, & Nelson, 2008). Carlo et al. (2007) suggest that the mixed evidence may be because the dimensions of parenting styles are constructs that are too broad when used to investigate specific prosocial behaviors, and that it may be more fruitful to examine specific socialization techniques. They found several specific techniquesdincluding conversations about prosocial behaviordwere associated with the Prosocial Tendency Measure (PTM) to a much greater degree than broader parenting styles. Associations with role-modeling could not be investigated in their study because the measurement of modeling did not survive exploratory factor analysis. Direct socialization of prosocial behavior The theoretical importance of role-modeling in the development of prosocial behavior follows from social learning theory (Bandura, 1976), as well as more recent neurological research indicating that the basis of imitative behavior (mirror neurons) may provide the neural substrate of empathy (Preston & de Waal, 2002). The most robust evidence from the experimental literature on children’s prosocial behavior is that having an adult experimenter role-model the desired behavior increases children’s prosocial behavior (Eisenberg & Fabes, 1998). If adult strangers have an impact, the regular, long term close relationships parents often have with their children should make parents particularly effective modelers of prosocial behavior (Hastings et al., 2007). The body of experimental research on the more specific modeling of prosocial behavior referenced above is supplemented by experimental work investigating gender’s relationship to modeling in general. Specifically, there is evidence supporting the occurrence of same-sex modeling in the form of an increased tendency to imitate a same-sex model beginning in early childhood (Bussey & Bandura, 1984). Although Bussey and Bandura’s sample was comprised of children between two and six years old, there was a stronger indication of same-sex imitation the older the children were. The gender intensification hypothesis (Hill & Lynch, 1983) suggests that adolescence is a time of intensified socialization of boys and girls into gender stereotypic roles, and there is even some evidence outside of laboratory settings that adolescent boys and girls may show an increase in gender stereotyped behaviors (Crouter, Manke, & McHale, 1995). However, little is known about whether there are neither different associations between the sex of the adolescent and parental modeling of prosocial behavior in general, nor about different sex-based associations with the modeling of charitable giving and volunteering in particular. After role-modeling, the next strongest evidence from the experimental literature is that children’s prosocial behavior can be increased through verbal communication (Eisenberg & Fabes, 1998). For example, most (though not all) experiments have found that other-oriented induction has a positive effect on children’s donations (see, e.g., Dlugokinski & Firestone, 1974; Eisenberg-Berg & Geisheker, 1979; Grusec, Sass-Kortsaak, & Simutis, 1978; cf. Lipscomb, Bregman, & McAllister, 1983). Carlo et al. (2007) find evidence that conversations emphasizing empathy are positively associated with prosocial behavior among adolescents. Although there has been extensive work on sex differences in the amount of prosocial behavior performed (see Eisenberg et al., 2006 for a review), fewer results are available about sex differences in the effectiveness of socialization techniques. Stukas, Switzer, Dew, Goycoolea, and Simmons (1999) find evidence that parental modeling has a stronger effect on girls’ altruistic self-image, suggesting perhaps a stronger effect on girls’ prosocial behavior. However, Carlo et al. (2007) find no evidence of sex differences in their PTM associations with parenting techniques and styles. Though social learning theory would suggest same-sex models should have a greater impact, more work is needed exploring these issues. Nationally-representative studies of socializing adolescent prosocial behavior Giving money to charity and volunteering are two key prosocial behaviors that receive emphasis in adulthood because they are important components of civil society. Educators, community leaders, and policy-makers expend much effort to encourage adults to give and volunteer. There are extensive literatures about adults’ charitable giving (Vesterlund, 2006) and volunteering (Musick & Wilson, 2008). Volunteering in adulthood is part of an ongoing pattern of behavior that for many was already underway during adolescence (Hart, Donnelly, Youniss, & Atkins, 2007; Janoski, Musick, & Wilson, 1998). There is a large literature about adolescents’ volunteering examining the influences of socio-economic status, extra-curricular activities, and religion, and investigating positive outcomes associated with volunteering (Eisenberg et al., 2009). This literature includes nationallyrepresentative studies (e.g., Youniss, McLellan, Su, & Yates, 1999), but to our knowledge there has been only one nationally-representative study that has estimated the association between a parental role-model of volunteering and adolescent volunteering (U.S. Department of Education, 1997). This study concludes that an adolescent whose parent volunteers is significantly more likely to volunteer, but did not examine parent–adolescent conversations about prosocial behavior. The study also did not investigate whether the role-modeling association was of similar magnitude for both girls

56


and boys. Finally, and again to our knowledge, there have been no nationally-representative studies of the association of adolescents’ volunteering with parenting dimensions or styles. However, Zaff, Moore, Papillo, and Williams (2003) found that parental monitoring of their child’s time, TV watching, and homework completion in the eighth gradedindicators of behavioral controldwere positively associated with the child’s volunteering six years later as a 19–20 year-old. This finding suggests there likely are associations with behavioral control and volunteering during adolescence. Charitable giving observed in adulthood most likely was also underway during adolescence, but charitable giving by adolescents has been studied much less intensively than adolescents’ volunteering.3 While Kim, LaTaillade, and Kim (2011) found an association between conversations about giving and adolescents’ giving, role-modeling was not considered and differences in the associations between boys and girls were not examined. The present study The present research estimates two modelsdone for the probability that the adolescent gives to charity and the other for the probability of volunteeringdin which parental role-modeling of volunteering and giving as well as conversations about giving are separate influences. The models also contain parenting dimensions of warmth, behavioral control, and psychological control as influences on adolescents’ giving and volunteering. Our hypotheses are that adolescents’ giving and volunteering are positively associated with (a) parental modeling and (b) conversations about giving. Because giving and volunteering are often thought of as closely related prosocial behaviors, our third hypothesis is that there are “cross-over” effects, specifically that parental modeling of giving is positively associated with adolescents’ volunteering and vice versa. While positive parenting provides an overall context that influences adolescent outcomes (Darling & Steinberg, 1993) and undoubtedly encourages the development of general prosocial behaviors, rolemodeling and talking about a specific prosocial behavior, like charitable giving, are likely to be more directly related to the development of that specific prosocial behavior. With this in mind, and in light of the mixed evidence about broad parenting dimensions and specific prosocial behaviors, our fourth hypothesis is that adolescents’ giving and volunteering are more strongly associated with role-modeling and conversations about giving than with the overall context of parenting dimensions. Our investigation of these hypotheses offers several contributions. First, as already noted we investigate role-modeling, conversations about prosocial behavior, and parenting dimensions as separate influences on adolescents’ giving and volunteering in a multiple regression framework. For example, among parents who talk to their children about giving many also model giving. Our analysis of separate associations avoids attributing to conversations about giving an association that is really due to modeling giving. Equally important, our analysis will reveal the importance of role-modeling, conversations about giving, and broader parenting styles. Second, our analysis is based on a nationally-representative sample of U.S. adolescents. We can learn whether results similar to those about modeling and conversations about prosocial behavior obtained from experimental studies with non-parents and smaller, more homogeneous samples also obtain in a population study. Third, we investigate whether modeling, conversations about giving, and parenting dimension associations differ by sex of the adolescent. In our models of adolescent giving and volunteering we control for family income and the education of the primary caregiver. Janoski and Wilson (1995) explain how adolescents and young adults may be more likely to volunteer not only because they were socialized to do so by their parents, but also because parents transmit socio-economic status to their children and, in turn, people with higher status are more likely to volunteer. Therefore, an empirical association between a parent’s volunteering and an adolescent’s volunteering may not reflect socialization due to the parent’s role-model, but rather a status transmission channel through which high-status parents are more likely to volunteer and also to transmit socioeconomic advantage to adolescents that makes it easier for them to volunteer. This possibility applies equally well to charitable giving because adolescents of high-income parents likely have access to more money from which they can give. Method Participants Participants are from the Child Development Supplement (CDS) of the Panel Study of Income Dynamics (PSID). The PSID is a long-term longitudinal study of household economics. Economic, health, and demographic data have been collected annually 1968–1997 and biennially since then in core main family interviews. Major funders of the PSID are the National Science Foundation, the National Institute on Aging, and the National Institute of Child Health and Human Development (NICHD); the CDS is funded by NICHD. In 1997 the PSID families with children age 12 and younger were selected for the first wave of the CDS

3 Wilhelm, Brown, Rooney, and Steinberg (2008) provide evidence of an association between the dollar amounts contemporaneously given to charitable organizations by adult-aged children and their parents. There also is evidence of an association between the volunteering of adult-aged children and their parents: Bekkers (2007) analyzes a binary volunteering variable; Janoski and Wilson (1995) analyze the ordinal intensity of volunteering; Mustillo, Wilson, and Lynch (2004) analyze the amount of hours volunteered. Bekkers (2011) finds evidence that adult’s volunteering and giving are associated with their parents’ past volunteering, retrospectively reported by the adults.


57

(CDS-I) and 88% participated. In 2002–2003 the CDS-I families who had continued in the core PSID (94% continued) were selected for re-interview in the second wave (CDS-II) and 91% participated. Among the CDS-II participating families our analysis focuses on adolescents 12–18 years of age (n ¼ 1468) because the child-reported warmth/support and psychological control constructs are asked only of adolescents. The response rate among adolescents is 89% (n ¼ 1312). There are a few additional sample exclusions. We do not include adolescents if their parent’s giving and volunteering data were not collected (n ¼ 40). A few adolescents did not answer questions about their giving and volunteering (n ¼ 10) or had missing data on other variables (n ¼ 18). The final analysis sample is n ¼ 1244. Of these, 634 are girls and 610 are boys. The number of African–American adolescents is 556. The number of African–American adolescents is disproportionally large because the original 1968 PSID design combined a nationally-representative sample of American families with a low-income oversample drawn mainly from the urban North and rural South. The low-income oversample yielded a large number of African–American families. The results we present use the weights provided by the PSID to account for the differential probability of selection of families in the original 1968 sample and for subsequent attrition. The weighted percentage of African–American adolescents in our sample is 22%.4

Measures Data for our study were collected in three separate interviews. First, adolescents were interviewed in person using computer-assisted personal interviewing (CAPI). For more sensitive questions about warmth/support and psychological control, as well as the volunteering questions, the mode was Audio Computer-Assisted Self-Interviewing (ACASI). Second, the adolescent’s primary caregiver was interviewed either in person or over the telephone (CATI) based on the caregiver’s preference. Third, either the primary caregiver or her spouse were interviewed in the PSID’s core main family interview using CATI. The large majority (91.8%) of the primary caregivers is the mother or step-mother of the adolescent. Another 2.5% of the primary caregivers are other women, typically a grandmother. Just under 6% of the primary caregivers are men, and almost all of these (95%) are biological fathers. Adolescents’ giving and volunteering In CDS-II adolescents were asked “did you give some of your money last year – if only a few pennies – to a church, synagogue, or another charity that helps people who are not part of your family?” Adolescents also were asked “were you involved in any volunteer service activities or service clubs in the last 12 months?” Responses to both questions were a binary “yes/no”. The two questions parallel those asked of the mother/step-mothers (or spouses) in the 2001 PSID core main family interview. Role-modeling Beginning in 2001, the PSID’s core main family interview contains data on parents’ giving and volunteering that form the role-modeling variables central to our analysis. The mother/step-mother/spouse reported if the family gave during the last calendar year more than $25 to religious or charitable organizations. These organizations were defined to the respondent to be non-profit organizations that “serve a variety of purposes such as religious activity, helping people in need, health care and medical research, education, arts, environment, and international aid”. The PSID’s measurement of giving compares well with other nationally-representative measures (Wilhelm, 2006a). Also reported in the core main family interview was whether the mother/step-mother did volunteer work through organizations, such as “coaching, helping at school, serving on committees, building and repairing, providing health care or emotional support, delivering food, doing office work, organizing activities, fundraising, and other kinds of work done for no pay”. Because the survey question asked if the mother/step-mother did any volunteer work “that totaled 10 h or more” during the previous calendar year (2000), lesser amounts of volunteering would have gone unreported. Despite the 10 h screen, the volunteer rate of women 25 years and older in the PSID (Wilhelm, 2006b) is just slightly less than in the Census Department’s Current Population Survey (Boraas, 2003): 29 versus 32%. For men the rates from both studies are 25%. Nesbit (2010) also finds that volunteer rates from the PSID are similar to those from the CPS. Note that both the giving and volunteering questions were about activity in calendar year 2000, implying that these variables describe role-modeling one year (or more) prior to the measurement of adolescent giving and volunteering in CDS-II.

4 The 22% is a little higher than the percentage from the Census. The difference arises because we use the adolescent’s self report of her/his own race (the Census uses the parent’s report of the adolescent’s race), combined with a method for resolving the adolescent’s race when her/his report is inconclusive (the adolescent reports “Hispanic”, so we know ethnicity, but not race; “multi-racial”, but we do not know which specific races; or does not answer the race question). In these inconclusive cases we use parental race to determine the adolescent’s race. Parents were allowed up to four mentions of their race, and if any one of those mentions was “African–American” we coded the adolescent as African–American.In the first wave of the CDS (1997) primary caregivers reported the race of their children, then all aged 0–12 years old. The percentage African–American non-Hispanic from these reports is 16%. Our calculations based on the 2000 Census indicate that the percentage of African–American children among all children in the 0–13 age group is between 15.2% (counting only children whose parent reported that the child had only one race; Census, 2000a) and 16.6% (also counting children whose parent reported “African– American” along with another race; Census, 2000b). Hence the 16% estimate using the CDS primary caregiver reports is right in-between the two comparable measures from the Census. The primary caregiver 1997 reports of children’s race match the CDS-II adolescent reports in all but 3% of the cases. Reestimating our models using the primary caregiver 1997 reports as the race control results in only negligible changes in the estimates.

58


Conversations about giving In CDS-II the primary caregiver was asked the binary response question: “Do you ever talk to (NAME of CHILD) about giving some of (his/her) money– if only a few pennies – to a church, synagogue, or another charity?” The question obviously parallels the giving question asked to adolescents. Unfortunately the time budget of CDS-II was not sufficient to add another question specifically about conversations concerning volunteering. Therefore we use the “talk-about-giving” response as a proxy measurement of the parent’s conversations about prosocial behavior in both models. Parenting dimensions In our main analysis we used the parenting dimension framework with measurements of three dimensions: (a) warmth and support (b) behavioral control, and (c) psychological control (Crockett & Hayes, 2011). We also conducted analyses replacing the dimension framework with parenting styles defined by interacting warmth/support with behavioral control to classify authoritative, authoritarian, permissive, and neglectful styles; this follows the procedure from Lamborn, Mounts, Steinberg, and Dornbusch (1991). Warmth and support were measured using 4 items developed from the Child Report of Parent Behavior Inventory (Schaefer, 1965). The adolescent reported the degree to which his/her mother/step-mother (a) “cheers me up when I am sad” (warmth and emotional support), (b) “gives me a lot of care and attention” (warmth and nurturance), (c) “often praises me” (support), and (d) “enjoys doing things with me” (warmth and supportive involvement). Item responses were on a 3-point scale: not like her, somewhat like her, a lot like her. To ensure privacy the responses were collected during the ACASI portion of the adolescent interview. Cronbach’s a ¼ .78. Behavioral control was measured using the mother/step-mother report of the Household Rules measure (Alwin, 1997).The measure has 9 items, each using a 5-point scale, describing the frequency that the mother/step-mother enforces rules specifically for the adolescent. The 9 items are about which friends the adolescent can spend time with, how late the adolescent can stay up, the use of time after school, when homework is done, if a place is set where homework is done, whether homework is checked, the eating of snack foods, which TV programs can be watched, and how much time can be spent watching TV (a ¼ .84). Psychological control was measured with the Psychological Control Scale-Youth Self Report, a set of 6 items based on the CRPBI items used by Barber (1996). The items are the adolescent’s report on the degree to which his/her mother/step-mother (a) “blames me for other family members’ problems”, (b) “changes the subject whenever I have something to say”, (c) brings up past mistakes when she criticizes me”, (d) “often interrupts me”, (e) “If I have hurt her feelings, she stops talking to me until I please her again”, and (f) “is always trying to change how I feel or think about things”. Item responses are on the 3-point not like her, somewhat like her, a lot like her scale (a ¼ .75). Demographic and economic variables Adolescents reported their race. In the present study African–American adolescents were the only minority group with sufficient numbers in the sample to include a variable designating their race. Race is coded as African–American ¼ 1 and all others ¼ 0. Sex and date of birth were collected when the participant was born. Sex is coded male ¼ 0, female ¼ 1. In the core main family interview, the mother/step-mother (or spouse) answered over 100 questions about components of family income. We use the PSID’s measure “total family income” that includes income from all major sources (e.g., labor earnings from both parents, business income, asset income, and transfers from the government). The PSID’s total family income compares well with income measured in the CPS (Gouskova & Schoeni, 2007). To further improve the quality of the income data as a measurement of the family’s long-run economic resources we took a five-year average of income using the 1995–1997, 1999, and 2001 core family interviews. The primary caregiver’s education is measured as less than high school, high school (the reference category), some college, and college or more. Empirical approach We estimate two logit models, one for the probability the adolescent gave money to a religious congregation or a charity and one for the probability the adolescent does volunteering. From these models we report the marginal effects so that the reported estimates can be directly interpreted as effects on the probabilities of giving and volunteering. The logit estimates are from weighted maximum likelihood. Results Table 1 presents the means, standard deviations, and pairwise correlations of the analysis variables. Among the adolescents, 72% gave and 41% volunteered. Similar percentages of their primary caregiversd69% and 38%drole-modeled these two prosocial behaviors. The 38% volunteer rate among primary caregivers may at first seem high relative to the CPS estimate for all adults (27.6%; Boraas, 2003), but the 38% rate is very close to the CPS estimate for women with children under 18 (39.8%; Boraas, 2003), like the large majority of primary caregivers in our sample. Whether or not the 41% volunteer rate for the adolescents is too high is less straightforward to judge. The rate for 16–18 year-olds in the CDS is indeed higher than the rate for 16–19 year-olds in the CPS (45% versus 30.9%; we compare the rates for older adolescents because the CPS does not estimate volunteer rates for younger adolescents). However, there are important

M 1. Adolescent giving .72 2. Adolescent volunteering .41 3. Role-model giving .69 4. Conversation about giving .62 5. Role-model volunteering .38 6. Warmth and support 1.44 7. Psychological control .48 8. Behavioral control 1.76 9. Age 15.3 10. Female .50 11. African–American .22 12. Log family income 10.79 Education of the primary caregiver 13. Less than high school .13 14. High school .36 15. Some college .27 16. College or more .24

SD

1

.45 .49 .46 .48 .49 .54 .45 1.12 1.93 .50 .41 .78

–

2

3

4

5

6

.22** .18** .18** .13** .07* .07* .08** .02 .08** .06* .09**

.23** .13** .23** .10** .06* .03 .08** .12** .10** .19**

.12** .40** .10** .06* .08** .01 .00 .18** .48**

.16** .11** .06 .24** .06* .07** .01 .07**

.13** .10** .03 .04 .04 .18** .35**

– .44** .08** .06* .03 .08** .17**

.34 .48 .44 .43

.12** .06* .05 .11**

.12** .16** .13** .14**

.39** .13** .12** .33**

.04 .15** .01 .12**

.28** .17** .13** .28**

.08** .05 .00 .12**

7

8

9

10

11

.07* .01 .04

.07* .03

.03 .02 .01 .01

.06* .04 .01 .00

.02 08** .06 .17**

12

13

14

15

.39** .23** .10** .46**

– .30** .24** .22**

– .45** .42**

– .34**

– – – – – .04 .11** .03 .08** .08** 08** .03 .02 .01

– .49** .07* .20** .14** 09** .01 .02 .08**

–

.39**


Table 1 Descriptive statistics and correlations.

Notes: The estimates use sample weights. *p .05,**p .01.

59

60


methodological differences between the two surveys that arise because the CDS was designed specifically to collect information from teenagers, whereas the CPS volunteering supplement was designed for adults but then also administered to older teens. These methodological differences suggest that future research is needed to make a definitive judgment about which method is better-suited to measure adolescent volunteering.5 With the methodological differences in minddand keeping in mind the closeness of the PSID and CPS estimates of volunteer rates for adults overall and for women with children under 18 like the primary caregivers in our sampledit seems to us that 41% is not an unreasonable estimate of the volunteer rate among adolescents. Sixty-two percent of parents talked to their adolescent about giving. The modeling and conversations about giving variables are each significantly correlated with adolescent giving and volunteering. Warmth and support are also significantly correlated with giving and volunteering, but the magnitudes are lower than the associations with the socialization variables. Similarly, psychological control is significantly (negatively) correlated with giving and volunteering, and the behavioral control–giving correlation is significantly positive, but these correlations are smaller than the modeling and conversations about giving correlations. The parent’s model of giving and model of volunteering are correlated (the correlation is among the strongest in the table) and both modeling variables are correlated with conversations about giving. This justifies the cotemporaneous investigation of both types of modeling along with conversations about giving in our empirical models. Not surprisingly, warmth and support are negatively correlated (fairly strongly) with psychological control. Both parenting dimensions are correlated with behavioral control, but the correlations are small (only one is significant). The correlations between the parenting dimensions and modeling/conversations about giving are in the anticipated direction with one exception (behavioral control–model of giving), and seven out of nine are significant. There are two other patterns to note. First, the fairly strong correlations between (the logarithm of) family income and both types of role-modeling imply that controlling for income is necessary to avoid misattributing an income association to a role-modeling association. The same applies to the primary caregiver’s education and the role-modeling variables. Second, parents talk about giving to older adolescents less and to girls more – the correlations are significant but small. Table 2 presents estimates of the logit models. In column 1, the logit model for adolescent giving is significant (c2ð16Þ ¼ 64:3, p < .001). The estimates indicate that the parental role-model of giving is associated with a 13 percentage point higher probability that the adolescent gives (p ¼ .005). Conversation about giving is associated with an 11 percentage point higher probability (p ¼ .002). The modeling association is large relative to the 72% baseline probability that the adolescent gives: comparing two adolescents who are identical in terms of the observed characteristics we control, except that one has a parent who gives and the other has a parent who does not, the adolescent with the parent role-modeling giving has a probability of giving that is 18% higher, relative to the baseline probability of giving (.13/.72 ¼ .18). The conversation about giving association also is large relative to baseline: an adolescent whose parent talks to him about giving has a 15% higher probability of giving (.11/.72). Hence, an adolescent whose parent both role-models and talks about giving has a probability of giving that is 33% higher than an adolescent whose parent does neither. In column 2, the logit model for adolescent volunteering is significant (c2ð16Þ ¼ 83:4, p < .001). The parental model of volunteering is associated with an 11 percentage point higher probability that the adolescent volunteers (p ¼ .008), and talking about giving is associated with an 8 percentage point higher probability of volunteering (p ¼ .050). These, too, are large associations, representing 27% and 20% higher probabilities of volunteering relative to the baseline 41% of adolescents who volunteer. Together these estimates imply that an adolescent whose parent both role-models volunteering and talks about giving has a probability of volunteering that is 47% higher than an adolescent whose parent does neither. The “crossover” parental model of giving is associated with a 16 percentage point higher probability that the adolescent volunteers (p < .001) – corresponding to a 39% higher probability of volunteering relative to baseline. The 11 percentage point association with the role-model of volunteering is large even though role-modeling of, and talking about, giving are entered as separate

5 There were three methodological differences between the two surveys: question wording, proxy reporting, and interview mode. In the CDS, after a series of questions about extra-curricular activities, such as homework, sports, school clubs/student government, and scouts/hobby clubs, the adolescent is asked: “Were you involved in any volunteer service activities or service clubs in the last 12 months?” (Emphasis in the original). In the September 2002 CPS, after the usual monthly questions about labor force activity in the prior week, the respondents were asked: “Since September 1st of last year, (have you/ has NAME) done any volunteer activities through or for an organization?” (Emphasis in the original). If the response to this question was “no”, the respondent was asked a follow-up: “Sometimes people don’t think of activities they do infrequently or activities they do for children’s schools or youth organizations as volunteer activities. Since September 1st of last year, (have you/has NAME) done any of these types of volunteer activities?” The CDS’s placement of its volunteering question in the context of extra-curricular activities seems well-suited to aid the adolescent in recalling any volunteering activity they might have done. The CPS’s placement of its volunteering questions in the context of labor force activity seems better-suited to respondents who have already made the transition to adulthood. In addition, the adult-orientation of the first CPS volunteering question seems to evoke a context less well-suited to how teenagers think about volunteering, and the follow-up question about volunteering in “children’s schools or youth organization” seems more relevant to the volunteering done by parents, rather than that done by teens.Second, all of the CDS adolescents self-reported their own volunteering. In the CPS, the overall rate of proxy reporting was 20%, and the proxy reports of volunteering are 40% lower than the self-reports (Nesbit, 2010). Whether proxy reporting was more prevalent for the 16–19 year-olds selected in the CPS sample is an important topic for future research.Third, in the CDS the volunteering questions were asked using the Audio Computer-Assisted Self-Interview mode. The CPS mode was Computer-Assisted Telephone-Interview. If there is any difference in social presentation effects, they would likely be weaker in ACASI relative to CATI (Tourangeau & Smith, 1998).Finally, the 45% rate among 16–18 year-olds in the CDS is very close to the 44% rate during the junior and senior years in high school estimated from the National Education Longitudinal Study of 1988 (NELS:88) (U.S. Department of Education, 2003). Like the CDS, the NELS:88 volunteering questions were designed specifically for teenage respondents.


61

Table 2 Logit models of the probability that the adolescent gives, volunteers.

Role-model giving Conversation about giving Role-model volunteering Warmth and support Psychological control Behavioral control Age Female African–American Log family income Education of the primary caregiver Less than high school Some college College or more Log-likelihood n

Adolescent giving

Adolescent volunteering

.13**(.05) .11**(.04) .03(.04) .00(.04) .05(.04) .04*(.02) .02(.01) .08*(.03) .09*(.04) .01(.03)

.16***(.05) .08*(.04) .11**(.04) .05(.04) .03(.05) .01(.02) .03**(.01) .11**(.04) .05(.05) .01(.03)

.07(.06) .03(.04) .08(.05) 12,256 1244

.03(.08) .16**(.05) .10(.06) 13,638 1244

Notes: The estimates are marginal effects from logit models, estimated using the sample weights. Each model contains a dummy variable indicating when the adolescent did not live with a mother/step-mother, a dummy variable indicating when the adolescent’s race is missing, and a dummy variable indicating when the primary caregiver’s education is missing. The reference category for the education of the primary caregiver is high school completion. Standard errors are in parentheses. *p .05,**p .01,***p .001.

effects; however, if those effects are removed the role-model of volunteering association rises to 15 percentage points (not shown in Table 2). Although the parenting dimension associations with adolescent giving and volunteering in columns 1 and 2 are in the expected direction, the estimates are small and the only statistically significant association with a parenting dimension is behavioral control in the giving model (p ¼ .038). We also estimated models that replaced the parenting dimensions with authoritative, authoritarian, permissive, and neglectful parenting styles as main effects, but the styles were not significant. Because parenting dimensions/styles may act as moderators (Steinberg, Lamborn, Dornbusch, & Darling, 1992), we also estimated different versions of the models in which dimensions (warmth, behavioral control) and styles (authoritative) moderated the effects of role-modeling and conversations about giving, but these interaction effects were also not significant. The estimates for demographic variables indicate that, controlling for role-modeling, conversation about giving, and parenting dimensions, girls and African–American adolescents have higher probabilities of giving by 8 percentage points (p ¼ .015) and 9 percentage points (p ¼ .022), respectively. While age and giving are not significantly associated, there is a significant association between age and volunteering (3 percentage points per year; p ¼ .007). Adolescent girls have an11 percentage point higher probability of volunteering (p ¼ .003). Race is not related to volunteering, and family income is not significantly related to giving or volunteering. Education of the primary caregiver is positively associated with both giving and volunteering, but the three indicator variables are jointly significant only in the volunteering model (p ¼ .017). Table 3 presents estimates of the logit models separately by sex. Giving model estimates in columns 1 and 2 show that the role-model association is much stronger for girls (20 percentage points, p ¼ .004) than boys (8 percentage points; p ¼ .204),

Table 3 Logit models of the probability that the adolescent gives, volunteers by sex. Adolescent giving

Role-model giving Conversation about giving Role-model volunteering Warmth and support Psychological control Behavioral control Age African–American Log family income Education of the primary caregiver Less than high school Some college College or more Log-likelihood n Notes: See notes to Table 2. *p .05, **p .01, ***p .001.

Adolescent volunteering

Girls

Boys

Girls

Boys

.20**(.07) .07(.05) .03(.05) .03(.05) .03(.06) .03(.02) .03*(.01) .14***(.04) .02(.04)

.08(.06) .16**(.05) .04(.06) .04(.06) .08(.06) .05(.03) .01(.02) .03(.07) .06(.05)

.17*(.07) .04(.06) .18**(.06) .08(.06) .07(.07) .01(.03) .03*(.02) .01(.07) .05(.05)

.17**(.06) .13**(.05) .06(.06) .21***(.06) .05(.06) .01(.03) .02(.01) .07(.06) .01(.04)

.01(.07) .00(.05) .05(.06) 5690 634

.15(.10) .06(.07) .10(.07) 6350 610

.16(.10) .11(.07) .02(.08) 7162 634

.19*(.09) .17*(.07) .13(.08) 6063 610

62


but the opposite pattern is evident for conversation about giving: for girls 7 percentage points (p ¼ .134) and for boys 16 percentage points (p ¼ .003). Estimates for the parenting dimensions are not significant for either girls or boys, though the .05 point estimate for boys has p ¼ .082.As for the demographic variables, there is an age-giving association for girls. The association between African–American race and giving seen in Table 2 is driven by a significant pattern among girls, not boys. The volunteering model in columns 3 and 4 shows a similar pattern: the role-model association is much stronger for girls (18 percentage points, p ¼ .003) than boys (6 percentage points, p ¼ .266), but the opposite for conversation about giving: for girls 4 percentage points (p ¼ .545) and for boys 13 percentage points (p ¼ .010).The “cross-over” role-model of giving association is equally strong for girls and boys (17 percentage points, ps ¼ .012 and .004). Warmth and support are significantly associated with boys’ volunteering (.21, p < .001). The volunteering-age association is due to the significant pattern among girls. The association between adolescent volunteering and the primary caregiver’s education is due to the association among boys (the p-value for the joint test of the three indicator variables being zero is .013). We checked the sensitivity of the results in Tables 2 and 3 to include a control for the gender of the primary caregiver, and the results were virtually unchanged. We estimated the models using only the primary caregivers who were mothers/stepmothers of the adolescents, and again the results were virtually unchanged. Because the number of male primary caregivers was so small in our sample we do not have the statistical power to investigate whether female caregivers are more efficacious role-models than male caregivers. However, for adolescents who lived with two parents, or a mother and her cohabiting partner, we can create woman-specific and man-specific volunteering-role-model variables and use those in place of the rolemodel of the primary caregiver. When we do this the results are little different from the weak .06 role-model result for boys’ volunteering reported in Table 3, column 4. The point estimate for the man-volunteering-role-model for boys’ volunteering is .07 (s.e. ¼ .07, p ¼ .293). This point estimate is a little larger than the .04 for the woman-volunteering-role-model (s.e. ¼ .06, p ¼ .550), but neither estimate is significantly different from zero, nor are the estimates significantly different from each other. Hence, the weak role-model result for boys’ volunteering in Table 3 does not appear to be an artifact of almost all of the primary caregivers being women. The girls’ volunteering result in Table 3 column 3 is largely, but not entirely, driven by the woman-volunteering-role-model (point estimate ¼ .14, s.e. ¼ .06, p ¼ .035); the man-volunteering-role-model is positive but not significant (point estimate ¼ .06, s.e. ¼ .07). Discussion The main results are that both parental role-modeling and conversations about giving are associated with adolescents’ giving and volunteering. An adolescent whose parent role-models giving and talks about giving has a probability of giving that is 33% higher than an adolescent whose parent does neither. An adolescent whose parent role-models volunteering and talks about giving has a probability of volunteering that is 47% higher than an adolescent whose parent does neither. Additional evidence of modeling is the “cross-over” parental role-model of giving association with adolescent volunteering: an adolescent whose parent models giving has a 39% higher probability of volunteering, and this is in addition to the associations due to the parent’s model of volunteering and talking about giving. Our findings are significant because they are the first validation using nationally-representative data describing charitable giving by U.S. adolescents in their natural settings of earlier laboratory evidence from smaller, more homogeneous samples regarding the impact of modeling and conversations about giving. Similarly, the associations we obtain for adolescents’ volunteering validate earlier experimental evidence about prosocial behavior that involves physical action, such as helping a confederate pick up dropped objects. Together the giving and volunteering results extend earlier work that showed parents as strong influences on the prosocial behavior exhibited by exemplars (e.g., Hart & Fegley, 1995; Oliner & Oliner, 1988): the new evidence suggests that parents strongly influence the kinds of prosocial behavior exhibited by a wider fraction of the population. Our results have further significance because they imply that role-modeling has a strong association with giving, even if conversation about giving is included in the analysis (and vice versa). That is, parental role-modeling and conversations about giving have separate associations with giving. Likewise, role-modeling and conversations about giving have separate associations with volunteering. The present evidence deepens our understanding of the previous regression evidence of the association between a parental role-model of volunteering and adolescent volunteering (McLellan & Youniss, 2003; U.S. Department of Education, 1997) to say that the association holds, albeit smaller in magnitude, even including rolemodeling of, and talking about, giving as separate effects. These associations of adolescent giving and volunteering with the specific socialization techniques are much stronger than associations with the broad parenting dimensions of warmth and support, behavioral control, and psychological control. These results are in line with Carlo et al.’s (2007) similar findings regarding the PTM and more direct socialization as compared to broader dimensions of parenting. Of course this does not imply that parenting dimensions are unimportant to the development of prosocial behavior broadly speaking, just that more direct socializing of specific behaviors may have a greater impact in encouraging those specific forms of behavior. And in one analysis – that of adolescent boys’ volunteering – we found warmth and support strongly related to volunteering. The implication is that the importance of warmth and support in the development of volunteering documented by Wuthnow (1995) is especially important for adolescent boys. Moreover, parenting dimensions may emerge as more important influences on prosocial behavior when adolescents transition into adulthood. It may be that parenting dimensions such as warmth and support affect the internalization of prosocial behaviors, and that relationship is better revealed by the giving and volunteering chosen after the adolescent


63

establishes her/his independent household. Future work studying these issues in emerging adulthood would help to clarify these processes. The models estimated separately by sex provide additional evidence about differences in the associations for girls and boys: role-model associations for girls are stronger than for boys, and conversations about giving are more highly associated for boys. The stronger role-model associations for girls are consistent with the literature including Stukas et al.’s (1999) evidence that parent modeling has a stronger effect on girls’ altruistic self-image than on boys’, and the recent explorations into parents’ civic and political engagement and adolescents’ own involvement in those areas in which a stronger association between parental participation and adolescent participation was found for female adolescents than for males (Cicognani, Bruna, Fournier, Gavray, & Born, 2012). Pinpointing the reasons these gender-based differences occur is challenging because gender may serve as characteristic that captures a variety of differences related to both environmental influences of socialization and nature-based predispositional differences between males and females (Hastings et al., 2007). As differences in behavior between males and females are found, they typically differ in such a way that supports gender stereotypes (Canary & Emmers-Sommer, 1997), and social role theory argues that males and females behave in a way that validates stereotypes regarding gender because they adhere to the social roles they perceive pertain to them (Eagly, 1987). Additionally, parents may treat children differently out of a desire to promote characteristics they associate as ideal for their children based on the child’s gender (Zahn-Waxler, 2000). Furthermore, children are more likely to imitate a same-sex model than a model of the opposite sex; although this result is tempered when the opposite-sex model behaves in a way more reflective of the opposing gender stereotype (e.g., a male child would be more likely to imitate a female role-model displaying a power-assertive behavior than a female role-model displaying a stereotypic female pattern of behavior) (Bussey & Bandura, 1984). Hence, the stronger association between parental role-modeling and female adolescents’ prosocial behavior in our study could be related to (1) more of the parental participants being female and thus a same-sex model for the adolescent girls and/or (2) an assumption by the adolescents or the caregiver that volunteering and giving may be more stereotypic of female roles than male and thus more appealing or important for female adolescents to imitate than their male counterparts. Although our results cannot definitively resolve this question, our sensitivity analysis that found a weak and insignificant association between men’s role-model of volunteering and boy’s volunteering tends to undercut the first explanation. Another possibility that might explain our sex-based findings is that the higher-level moral reasoning sometimes exhibited by girls (Eisenberg et al., 2006, p. 698) could imply that girls are more responsive to modeling, whereas boys have to be more explicitly, and verbally, encouraged to give and volunteer. In any event, these results call for the theoretical development of predictions about sex differences in the effectiveness of modeling and conversations about prosocial behavior that can be tested in future research. Such predictions may need to be situation dependent because effectiveness seems to depend on the type of role-modeling (Table 3 shows that the “cross-over” parental role-model of giving has an equivalent association with volunteering for boys and girls). They may also depend on the type of prosocial behavior: Ribar and Wilhelm (2006) find that role-modeling prosocial behavior that helps those inside the family (co-residence with elderly parents and financial help of elderly parents) has stronger effects on adult sons’ attitudes about the respective prosocial behavior than on adult daughter’s attitudes. There are a few limitations of our study that should be kept in mind while interpreting the results. The most important is that the magnitudes of the associations we estimate using cross-sectional data are rather weak evidence of any sort of causality. For example, a parent who is strongly empathic would likely be warm and supportive to her child, and more likely to role-modeling volunteering, while her child (also empathic) may be more likely to volunteer. While the previous experimental evidence leads us to think that some portion of the associations we estimate represent causal effects, longitudinal data controlling for prior socialization and parenting characteristics would be stronger tests of real-world influences. A second limitation is that we did not have a direct measure of whether the parent talks to the adolescent about volunteering. Had we had a direct conversation measure in the volunteering model, rather than the conversation about giving we used as a proxy, the estimated association may have been stronger. Our last set of limitations also has to do with measurement. Giving and volunteering are binary variables, and the restricted variance can present problems for estimation. Despite their limited range the associations with parental rolemodeling and conversations are strong enough that we obtained significant results. Nevertheless, the binary range does mean that we do not know, for example, if larger amounts given by parents are associated with larger amounts given by adolescents (see the studies in endnote3 for results about amounts, albeit for adult children, not adolescents). Another measurement limitation is that we do not have a measurement that indicates the adolescent’s involvement in the parent’s prosocial behavior. This issue could change the results or interpretation in three ways. First, while we think that most adolescents are aware of whether or not their parents give to charity or religious congregations – and parent volunteering would seem to be even harder for the adolescent to miss – we cannot provide evidence of this. However, to the extent that any such measurement errors are false negatives (adolescents do not know it, but their parents give), our estimates of the associations are likely too small, implying the actual associations are even larger than we have estimated. Second, it is possible that parents’ giving and volunteering are in fact a unified activitydgiving and volunteering at church services, for example. This would help to explain the cross-over effect because giving and volunteering may not be separate events. Attendance at church services itself is often a unified activity in which both the adolescent and parent participate. Adding church service attendance is beyond the scope of this paper, but because attendance in the U.S. context is important, it is a promising topic for future research. Third, the adolescent’s behavior may arise because the parent compels him/her to give and/or volunteer to a

64


particular cause that the parent is already engaged in (Fletcher, Elder, & Mekos, 2000). In this case, it is not so much rolemodeling and as it is mutual activity. Finally, we do not have any measure of the content of the conversations about giving that parents had with the adolescents. Law and Shek (2009) studied family influencesdincluding support, modeling, and the value placed on volunteeringdamong Chinese adolescents. Of these, family values were the strongest predictor of adolescent volunteering. While the intergenerational transmission of values must involve conversations about those values, not all conversations about giving or volunteering necessarily include valuesdparents may be discussing that they give (i.e., drawing the adolescent’s attention to the parental role-model), not why they give. Future work assessing content in conversations would be able to investigate the transmission of specific values, thus allowing for greater precision in examining the social forces that promote prosocial behavior. Keeping these limitations in mind, the present study finds nationally-representative evidence that parental role-modeling and conversations about prosocial behavior have relatively strong associations with adolescents’ giving and volunteering. These population-based results, in combination with the previous experimental results establishing causal effects of nonparental modeling and verbal communication, have a significant practical implication: practitioners and policy-makers, whose efforts traditionally have been to encourage present-day adults to give and volunteer, should begin those efforts earlier by guiding parents in their role-modeling and talking about giving and volunteering, in order to nurture potentially lifelong prosocial behavior among their adolescents. A second practical recommendation derived from the results is that conversations about giving and volunteering should be maintained during adolescence, especially for boys. Given that the percentage of parents that talk to their children about giving declines through adolescence – from 72% among parents of 10–12 year-olds to 60% among parents of 16–18 year-olds in our data – our findings strongly imply that shifting this pattern may lead to significant increases in prosocial behavior among older adolescents. Acknowledgments This work was funded by a grant from the Notre Dame Science of Generosity project (http://generosityresearch.nd.edu/). The Indiana University Lilly School of Philanthropy Research Fund 54-921-12 also provided financial support. Grants from the Atlantic Philanthropies, the Bill & Melinda Gates Foundation, the Charles Stewart Mott Foundation, the Corporation for National and Community Service, the Fidelity Investments Charitable Gift Fund, and the John Templeton Foundation support the Center on Philanthropy Panel Study – the PSID’s prosocial behavior module. An early version of this paper was presented at the March 2010 meeting of the Society for Research on Adolescence in Philadelphia, USA. References Alwin, D. F. (1997). Detroit area study, 1997: Social change in religion and child rearing. Ann Arbor, MI: ICPSR04120-v1. Bandura, A. (1976). Social learning theory. Englewood-Cliffs, NJ: Prentice-Hall. Barber, B. K. (1996). Parental psychological control: revisiting a neglected construct. Child Development, 67(6), 3296–3319. http://dx.doi.org/10.2307/ 1131780. Barry, C., Padilla-Walker, L. M., Madsen, S. D., & Nelson, L. J. (2008). The impact of maternal relationship quality on emerging adults’ prosocial tendencies: indirect effects via regulation of prosocial values. Journal of Youth and Adolescence, 37(5), 581–591. http://dx.doi.org/10.1007/s10964-007-9238-7. Baumrind, D. (1971). Current patterns of parental authority. Developmental Psychology, 4(1), 1–103. http://dx.doi.org/10.1037/h0030372. Baumrind, D. (1991). Parenting styles and adolescent development. In J. Brooks-Gunn, R. M. Lerner, & A. C. Peterson (Eds.), Interpersonal and Sociocultural Factors: Vol. 2. Encyclopedia of Adolescence (pp. 746–758). New York, NY, US: Garland Publishing, Inc. Bekkers, R. (2007). Intergenerational transmission of volunteering. Acta Sociologica, 50(2), 99–114. http://dx.doi.org/10.1177/0001699307077653. Bekkers, R. (2011). Charity begins at home: How socialization experiences influence giving and volunteering. Unpublished manuscript. Amsterdam: VU Amsterdam. Boraas, S. (2003). Volunteerism in the United States. Monthly Labor Review, 126(8), 1–11. www.bls.gov/opub/mlr/2003/08/art1full.pdf. Bussey, K., & Bandura, A. (1984). Influence of gender constancy and social power on sex-linked modeling. Journal of Personality and Social Psychology, 47(6), 1292–1302. http://dx.doi.org/10.1037/0022-3514.47.6.1292. Canary, D. J., & Emmers-Sommer, T. M. (1997). Sex and gender differences in personal relationships. New York: Guilford Press. Carlo, G., McGinley, M., Hayes, R., Batenhorst, C., & Wilkinson, J. (2007). Parenting styles or practices? Parenting, sympathy, and prosocial behaviors among adolescents. The Journal of Genetic Psychology: Research and Theory on Human Development, 168(2), 147–176. http://dx.doi.org/10.3200/GNTP.168.2.147-176. Carlo, G., Mestre, M. V., Samper, P., Tur, A., & Armenta, B. E. (2011). The longitudinal relations among dimensions of parenting styles, sympathy, prosocial moral reasoning, and prosocial behaviors. International Journal of Behavioral Development, 35(2), 116–124. http://dx.doi.org/10.1177/0165025410375921. Census. (2000a). PHC-T-9. Population by age, sex, Race, and Hispanic or Latino Origin for the United States: 2000. Table 1. Total population by age, race and Hispanic or Latino Origin for the United States: 2000. http://www.census.gov/population/www/cen2000/briefs/phc-t9/tables/tab01.pdf. Census. (2000b). PHC-T-9. Population by age, sex, race, and Hispanic or Latino Origin for the United States: 2000. Table 4.Population by race alone or in combination and age for the United States: 2000. http://www.census.gov/population/www/cen2000/briefs/phc-t9/tables/tab04.pdf. Chase-Lansdale, P. L., Wakschlag, L. S., & Brooks-Gunn, J. (1995). A psychological perspective on the development of caring in children and youth: the role of the family. Journal of Adolescence, 18(5), 515–556. http://dx.doi.org/10.1006/jado.1995.1037. Cicognani, E., Bruna, Z., Fournier, B., Gavray, C., & Born, M. (2012). Gender differences in youth’s engagement and participation. The role of parents and of adolescents’ social and civic participation. Journal of Adolescence, 35(3), 561–576. http://dx.doi.org/10.1016/j.adolescence.2011.10.002. Crockett, L. J., & Hayes, R. (2011). Parenting practices and styles. In B. Brown, & M. Prinstein (Eds.), Interpersonal and sociocultural factors: Vol. 2. Encyclopedia of adolescence (pp. 241–248). San Diego, CA US: Academic Press. http://dx.doi.org/10.1016/B978-0-12-373915-5.00077-2. Crouter, A. C., Manke, B. A., & McHale, S. M. (1995). The family context of gender intensification in early adolescence. Child Development, 66(2), 317–329. http://dx.doi.org/10.2307/1131580. Darling, N., & Steinberg, L. (1993). Parenting style as context: an integrative model. Psychological Bulletin, 113(3), 487–496. http://dx.doi.org/10.1037/00332909.113.3.487. Dlugokinski, E. L., & Firestone, I. J. (1974). Other centeredness and susceptibility to charitable appeals: effects of perceived discipline. Developmental Psychology, 10(1), 21–28. http://dx.doi.org/10.1037/h0035560.


65

Eagly, A. H. (1987). Sex differences in social behavior: A social-role interpretation. Hillsdale, NJ: Erlbaum. Eisenberg, N., & Fabes, R. A. (1998). Prosocial development. In N. Eisenberg, W. Damon, & N. Eisenberg (Eds.) (5th ed)Social, emotional, and personality development: Vol 3. Handbook of child psychology (pp. 701–778). Hoboken, NJ, US: John Wiley & Sons Inc. Retrieved from EBSCO host. Eisenberg, N., & Mussen, P. H. (1989). The roots of prosocial behavior in children. New York, NY, US: Cambridge University Presshttp://dx.doi.org/10.1017/ CBO9780511571121. Eisenberg, N., Fabes, R. A., & Spinrad, T. L. (2006). Prosocial development. In N. Eisenberg, W. Damon, & R. M. Lerner (Eds.) (6th ed.)Social, emotional, and personality development: Vol. 3. Handbook of child psychology (pp. 646–718). Hoboken, NJ, US: John Wiley & Sons Inc. Retrieved from EBSCO host. Eisenberg, N., Morris, A., McDaniel, B., & Spinrad, T. L. (2009). Moral cognitions and prosocial responding in adolescence. In R. M. Lerner, & L. Steinberg (Eds.) (3rd ed.)Individual bases of adolescent development: Vol 1. Handbook of adolescent psychology (pp. 229–265). Hoboken, NJ, US: John Wiley & Sons Inc. Retrieved from EBSCO host. Eisenberg-Berg, N., & Geisheker, E. (1979). Content of preachings and power of the model/preacher: the effect on children’s generosity. Developmental Psychology, 15(2), 168–175. http://dx.doi.org/10.1037/0012-1649.15.2.168. Fletcher, A. C., Elder, G. H., & Mekos, D. (2000). Parental influences on adolescent involvement in community activities. Journal of Research on Adolescence, 10(1), 29–48. http://dx.doi.org/10.1207/SJRA1001_2. Gouskova, E., & Schoeni, R. F. (2007). Comparing estimates of family income in the panel study of income dynamics and the March current population survey, 1968–2005. Technical Series Paper #07-01. Ann Arbor: University of Michigan Institute for Social Research http://psidonline.isr.umich.edu/Publications/ Papers/tsp/2007-01_Comparing_Estimates_PSID.pdf. Grusec, J. E., Sass-Kortsaak, P., & Simutis, Z. M. (1978). The role of example and moral encouragement in the training of altruism. Child Development, 49(3), 920–923. http://dx.doi.org/10.2307/1128273. Gunnoe, M. L., Hetherington, E. M., & Reiss, D. (1999). Parental religiosity, parenting style, and adolescent social responsibility. Journal of Early Adolescence, 19(2), 199–225. http://dx.doi.org/10.1177/0272431699019002004. Hardy, S. A., Bhattacharjee, A., Reed, A., & Aquino, K. (2010). Moral identity and psychological distance: the case of adolescent parental socialization. Journal of Adolescence, 33(1), 111–123. http://dx.doi.org/10.1016/j.adolescence.2009.04.008. Hardy, S. A., Padilla-Walker, L. M., & Carlo, G. (2008). Parenting dimensions and adolescents’ internalisation of moral values. Journal of Moral Education, 37(2), 205–223. http://dx.doi.org/10.1080/03057240802009512. Hart, D., & Fegley, S. (1995). Prosocial behavior and caring in adolescence: relations to self-understanding and social judgment. Child Development, 66(5), 1346–1349. http://dx.doi.org/10.2307/1131651. Hart, D., Donnelly, T. M., Youniss, J., & Atkins, R. (2007). High school community service as a predictor of adult voting and volunteering. American Educational Research Journal, 44, 197–219. http://dx.doi.org/10.3102/0002831206298173. Hastings, P., Utendale, W., & Sullivan, C. (2007). The socialization of prosocial development. In J. E. Grusec, & P. D. Hastings (Eds.), Handbook of socialization: Theory and Research (pp. 638–664). New York: Guilford Publications. Hill, J. P., & Lynch, M. E. (1983). The intensification of gender-related role expectations during early adolescence. In J. Brooks-Gunn, & A. C. Petersen (Eds.), Girls and puberty: Biological and psychosocial perspectives (pp. 201–228). New York: Plenum. Hoffman, M. L. (2000). Empathy and moral development: Implications for caring and justice. Cambridge, UK: Cambridge University Press. Janoski, T., & Wilson, J. (1995). Pathways to voluntarism: family socialization and status transmission models. Social Forces, 74(1), 271–292. http://dx.doi. org/10.2307/2580632. Janoski, T., Musick, M., & Wilson, J. (1998). Being volunteered? The impact of social participation and pro-social attitudes on volunteering. Sociological Forum, 13(3), 495–519. Retrieved from http://www.jstor.org/stable/684699. Kim, J., LaTaillade, J., & Kim, H. (2011). Family processes and adolescents’ financial behaviors. Journal of Family Economic Issues. http://dx.doi.org/10.1007/ s10834-011-9270-3. published on-line. Knafo, A., & Plomin, R. (2006). Parental discipline and affection and children’s prosocial behavior: genetic and environmental links. Journal of Personality and Social Psychology, 90(1), 147–164. http://dx.doi.org/10.1037/0022-3514.90.1.147. Krevans, J., & Gibbs, J. C. (1996). Parents’ use of inductive discipline: relations to children’s empathy and prosocial behavior. Child Development, 67(6), 3263– 3277. http://dx.doi.org/10.2307/1131778. Laible, D. J., & Carlo, G. (2004). The differential relations of maternal and paternal support and control to adolescent social competence, self-worth, and sympathy. Journal of Adolescent Research, 19(6), 759–782. http://dx.doi.org/10.1177/0743558403260094. Lamborn, S. D., Mounts, N. S., Steinberg, L., & Dornbusch, S. M. (1991). Patterns of competence and adjustment among adolescents from authoritative, authoritarian, indulgent, and neglectful families. Child Development, 62(5), 1049–1065. http://dx.doi.org/10.1111/j.1467-8624.1991.tb01588.x. Law, B. M. F., & Shek, D. T. L. (2009). Family influence on volunteering intention and behavior among Chinese adolescents in Hong Kong. Adolescence, 44(175), 665–683. Retrieved from http://ezproxy.lib.indiana.edu/login?url¼http://search.proquest.com/docview/621980353?accountid¼11620. Lipscomb, T. J., Bregman, N. J., & McAllister, H. A. (1983). The effect of words and actions on American children’s prosocial behavior. Journal of Psychology: Interdisciplinary and Applied, 114(2), 193–198. Retrieved from EBSCO host. Maccoby, E. E., & Martin, J. A. (1983). Socialization in the context of the family: Parent–child interaction. In E. M. Hetherington (Ed.), Socialization, personality, and social development: Vol. 4. Handbook of child psychology (pp. 1–101). New York: Wiley. McLellan, J. A., & Youniss, J. (2003). Two systems of youth service: determinants of voluntary and required youth community service. Journal of Youth and Adolescence, 32(1), 47–58. http://dx.doi.org/10.1023/A:1021032407300. Musick, M. A., & Wilson, J. (2008). Volunteers: A social profile. Bloomington, IN: Indiana University Press. Mustillo, S., Wilson, J., & Lynch, S. M. (2004). Legacy volunteering: a test of two theories of intergenerational transmission. Journal of Marriage and Family, 66(2), 530–541. http://dx.doi.org/10.1111/j.1741-3737.2004.00036.x. Nesbit, R. (2010). A comparison of volunteering data in the panel study of income dynamics and the current population survey. Nonprofit and Voluntary Sector Quarterly, 39(24), 753–761. http://dx.doi.org/10.1177/0899764009339075. Oliner, S. P., & Oliner, P. M. (1988). The altruistic personality: Rescuers of Jews in Nazi Europe. New York: The Free Press. Padilla-Walker, L. M., Carlo, G., Christensen, K. J., & Yorgason, J. B. (2012). Bidirectional relations between authoritative parenting and adolescents’ prosocial behaviors. Journal of Research on Adolescence, 22(3), 400–408. http://dx.doi.org/10.1111/j.1532-7795.2012.00807.x. Preston, S. D., & de Waal, F. M. (2002). Empathy: its ultimate and proximate bases. Behavioral and Brain Sciences, 25(1), 1–20. http://dx.doi.org/10.1017/ S0140525X02000018. Ribar, D. C., & Wilhelm, M. O. (2006). Exchange, role modeling and the intergenerational transmission of elder support attitudes: evidence from three generations of Mexican-Americans. Journal of Socio-Economics, 35(3), 514–531. http://dx.doi.org/10.1016/j.socec.2005.11.014. Schaefer, E. S. (1965). Children’s reports of parental behavior: an inventory. Child Development, 6(2), 417–424. Soenens, B., Duriez, B., Vansteenkiste, M., & Goossens, L. (2007). The intergenerational transmission of empathy-related responding in adolescence: the role of maternal support. Personality and Social Psychology Bulletin, 33(3), 299–311. http://dx.doi.org/10.1177/0146167206296300. Steinberg, L., Lamborn, S. D., Dornbusch, S. M., & Darling, N. (1992). Impact of parenting on adolescent achievement: authoritative parenting, school involvement, and encouragement to succeed. Child Development, 63(5), 1266–1281. http://dx.doi.org/10.1111/j.1467-8624.1992.tb01694.x. Stukas, A. R., Switzer, G. E., Dew, M., Goycoolea, J. M., & Simmons, R. G. (1999). Parental helping models, gender, and service-learning. Journal of Prevention & Intervention in the Community, 18(1–2), 5–18. http://dx.doi.org/10.1300/J005v18n01_02. Tourangeau, R., & Smith, T. W. (1998). Collecting sensitive information with different modes of data collection. In M. Couper, R. P. Baker, J. Bethlehem, C. Clark, J. Martin, W. L. Nicholls II, & J. M. O’Reilly (Eds.), Computer assisted survey information collection (pp. 431–453). New York: John Wiley & Sons. U.S. Department of Education, & National Center for Education Statistics. (1997). Student participation in community service activity, NCES 97-331, by Mary Jo Nolin, Bradford Chaney, and Chris Chapman. Project Officer, Kathryn Chandler. Washington, DC. Retrieved from http://nces.ed.gov/pubs97/97331.pdf.

66


U.S. Department of Education, & National Center for Education Statistics. (2003). Volunteer service by young people from high school through early adulthood, NCES 2004-365, by Mike Planty, and Michael Regnier. Washington, DC. Retrieved from http://nces.ed.gov/pubs2004/2004365.pdf. Vesterlund, L. (2006). Why do people give? In R. Steinberg, & W. W. Powell (Eds.), The nonprofit sector: A research handbook (2nd ed.) (pp. 568–590). New Haven, CT: Yale University Press. Wilhelm, M. O. (2006a). New data on charitable giving in the PSID. Economics Letters, 92(1), 26–31. http://dx.doi.org/10.1016/j.econlet.2006.01.009. Wilhelm, M. O. (2006b). The 2001 Center on Philanthropy Panel Study user’s guide. Indianapolis, IN: IUPUI. https://resources.oncourse.iu.edu/access/content/ user/mowilhel/Web_page/data.htm. Wilhelm, M. O., Brown, E., Rooney, P. M., & Steinberg, R. (2008). The intergenerational transmission of generosity. Journal of Public Economics, 92(10–11), 2146–2156. http://dx.doi.org/10.1016/j.jpubeco.2008.04.004. Wuthnow, R. (1995). Learning to care: Elementary kindness in an age of indifference. New York: Oxford University Press. Youniss, J., McLellan, J. A., Su, Y., & Yates, M. (1999). The role of community service in identity development: normative, unconventional, and deviant orientations. Journal of Adolescent Research, 14(2), 248–261. http://dx.doi.org/10.1177/0743558499142006. Zaff, J. F., Moore, K. A., Papillo, A. R., & Williams, S. (2003). Implications of extracurricular activity participation during adolescence on positive outcomes. Journal of Adolescent Research, 18(2), 599–630. http://dx.doi.org/10.1177/0743558403254779. Zahn-Waxler, C. (2000). The early development of empathy, guilt, and internalization of responsibility: implications for gender differences in internalizing and externalizing problems. In R. Davidson (Ed.), Anxiety, depression, & emotion: Vol. 1. Wisconsin symposium on emotion (pp. 222–265). Oxford, UK: Oxford University Press.

Do Market Incentives Crowd Out Charitable Giving?

Greater Emotional Gain from Giving in Older Adults: Age-Related Positivity Bias in Charitable Giving.

Demotivating incentives and motivation crowding out in charitable giving.

Commentary: Greater Emotional Gain from Giving in Older Adults: Age-Related Positivity Bias in Charitable Giving.

Framing charitable donations as exceptional expenses increases giving.

Giving more information about drug treatment.

Nursing and philanthropy: both about giving and joy.

Charitable giving and lay morality: understanding sympathy, moral evaluations and social positions.

Charitable giving and reflexive individuals: How personal reflexivity mediates between structure and agency.

The heart of the story: peripheral physiology during narrative exposure predicts charitable giving.

On Reminder Effects, Drop-Outs and Dominance: Evidence from an Online Experiment on Charitable Giving.

The evaluability bias in charitable giving: Saving administration costs or saving lives?

Giving and Receiving Awards.

Mental Imagery, Impact, and Affect: A Mediation Model for Charitable Giving.

Charitable giving for HIV and AIDS: results from a Canadian national survey.

Impact of the 'Giving Cigarettes is Giving Harm' campaign on knowledge and attitudes of Chinese smokers.

How Research on Charitable Giving Can Inform Strategies to Promote Human Milk Donations to Milk Banks.

Giving back or giving up: Native American student experiences in science and engineering.

Physician giving consent.

Thinkers and feelers: Emotion and giving.

Giving a health talk.

A time for giving.

Editorial: Giving up smoking.

Giving patients a choice.