Why Does Social Capital Matter in Health Communication Campaigns?

Abstract

The present study examined the psychological mechanisms underlying the indirect effects of antidrug-specific social capital on targeted parent-child communication about drugs. Moreover, it explored why campaign exposure and social capital exert interactive influences on parent-child communication. Using a three-round longitudinal panel data set from the National Survey of Parents and Youth (NSPY), we found that when taking into account behavioral attitude, subjective norm, and perceived behavioral control, parents’ attitude toward talking about drugs with their child mediated the effects of antidrug-specific social capital on targeted parent-child communication about drugs. Furthermore, behavioral attitude mediated the interactive effects of campaign exposure and social capital on parent-child communication. The implications of these findings for research on the connection between media exposure and conversation and for public health interventions were discussed.

Keywords

social capital health communication campaign targeted parent-child communication theory of planned behavior drugs

Social capital is one of the most widely used concepts in communication (Shah, Rojas, & Cho, 2009), and health communication scholars have recently recognized the role of social capital in determining the success of health communication campaigns (e.g., Lee, 2014; Viswanath, Steele, & Finnegan, 2006). Despite numerous explications of social capital, community members working together to solve their common problems lies at the heart of the meaning of social capital; thus, it has been commonly operationalized as group membership or local community ties (Hess, 2013; Kikuchi & Coleman, 2012). In the context of health campaigns, we conceptualize social capital as the extent to which one participates in campaign-relevant community activities, a definition which is consistent with those of the recent studies in this field (e.g., Southwell et al., 2010; Thorson & Beaudoin, 2004; Viswanath et al., 2006). Health communication scholars and public health practitioners have accentuated the importance of actual or potential resources embedded within a community, such as social capital, because health behaviors are as heavily influenced by social contextual factors as by exposure to health-related information through mediated and interpersonal sources (Biglan, Flay, Embry, & Sandler, 2012; Merzel & D’Afflitti, 2003). Although it has been well-established that social capital is good for health, we do not yet know the mechanisms by which social capital affects media audience’s health-related cognitions and behaviors (Kawachi, Subramanian, & Kim, 2008; Putnam, 2000). Hence, the following question has emerged: Why does campaign-specific social capital contribute to achieving campaign goals?

To answer this question, we examine how the National Youth Antidrug Media Campaign (antidrug campaign, hereafter) built and mobilized campaign-related social capital, thereby accelerating progress toward important campaign goals. The antidrug campaign was conducted in the United States between 1998 and 2004 with approximately one billion dollars appropriated by the U.S. Congress to prevent and discourage children and adolescents from using illegal drugs (Hornik, Jacobsohn, Orwin, Piesse, & Kalton, 2008; Orwin et al., 2006). Given that effective parenting practices are crucial to healthy youth development (Biglan et al., 2012), the campaign aimed to promote parent-child communication about drugs (Orwin et al., 2006). Whereas the summative evaluations of the campaign demonstrated that parents’ antidrug-specific community participation (i.e., antidrug-specific social capital) increased parent-child communication about drugs (Lee, 2014), the psychosocial processes that mediate such effect have been rarely described. To redress this oversight, we try to reveal the underlying mechanisms of antidrug-specific social capital’s effects on parent-child conversations about drugs.

More importantly, we aim to provide compelling explanations for the finding that campaign-specific social capital moderates the effects of campaign exposure on campaign-related conversations (Lee, 2014). Stated differently, exposure to the antidrug campaign messages induced parent-child communication about drugs only among parents who did not actively attend drug-prevention events in their local community. By contrast, this effect was not detected for those exhibiting high levels of such community engagement. How can we account for such linkages between campaign exposure and conversations about the campaign topic?

Although scholars have long wrestled with investigating the roles of conversation in media effects in general and health campaigns in particular, the mechanisms that explain the interplay between campaign exposure or media use more broadly and conversations have not been fully explicated (Shah et al., 2009; Southwell & Yzer, 2007; Thorson & Beaudoin, 2004). We attempt to contribute to this line of research in the following ways. First, we adopt theory of planned behavior (TPB), which will be explained in the next section, as a theoretical framework. Among a number of behavioral change theories, TPB has offered useful guidelines for designing effective campaign messages and evaluating campaign success by identifying major determinants of behavior (Fishbein & Ajzen, 2010). Second, most of the studies examining the connection between media exposure and conversation have adopted a rather simplistic conceptualization of conversation (Southwell & Yzer, 2007). To adequately incorporate conversation into studies on media effects, one should go beyond measuring the existence or absence of conversation and its frequency, and consider conversational partners, as well as actual content or topics of such conversations. To fill this gap, we employ the concept of targeted parent-child communication about drugs, which refers to direct and indirect, as well as one-time and ongoing, communication that parents have with their offspring specifically about alcohol, tobacco, marijuana, and other drugs (Kam & Middleton, 2013). Third, this study uses a three-round longitudinal panel survey, which enables researchers to detect changes over time and test mediational processes at an individual level (Eveland & Morey, 2011; Slater, 2006). Given that most studies on media, social capital, and health have relied on cross-sectional data, the present investigation has unique value.

To summarize, we explore the psychological mechanisms underlying the indirect effects of antidrug-specific social capital on targeted parent-child communication about drugs. Moreover, we examine why campaign exposure and social capital exert interactive influences on parent-child communication. By doing so, we respond to Southwell and Torres’s (2006) call for “the need to theorize about the mechanism by which mere exposure might encourage conversation when it otherwise would not have taken place” (p. 336). This study is theoretically and practically important because one cannot design and implement effective health promotion programs without understanding the conditions and processes of program effects, and because it adds new knowledge to the literature that connects media exposure and conversation.

The Roles of Social Capital in Health Communication Campaigns

Social capital was originally defined as “the aggregate of the actual or potential resources that are linked to possession of a durable network of more or less institutionalized relationships of mutual acquaintance and recognition” (Bourdieu, 1986, p. 248). Taking a slightly different approach, however, we situate the concept of social capital in the context of community-based drug-prevention efforts. This is advantageous in the following respects. First, unlike prior studies including both general social capital measures (e.g., civic participation) and campaign-relevant social capital measures, we define social capital as only antidrug-related community ties. Since “social capital . . . has a valence depending on the goal in question” (Sampson, Morenoff, & Earls, 1999, p. 636), researchers have begun to study differential effects of varying types of social group activities on health outcomes (e.g., Ellaway & Macintyre, 2007; Stephens, Rimal, & Flora, 2004). By focusing on antidrug-related activities in a local community, we ensure that social capital in this study captures only antidrug-related information which residents are exposed to through their contacts with community organizations or other residents in their communities.

Second, instead of asking how many groups respondents joined, the survey used in this study, the National Survey of Parents and Youth (NSPY), measured whether respondents participated in community antidrug-related events. Because “membership does not necessarily indicate action” (Kikuchi & Coleman, 2012, p. 196; see also Putnam, 2000, p. 157), the NSPY items are uniquely suitable to assess active relationships among community residents with regard to drug-prevention efforts. This is significant considering that the number of group memberships does not match one’s actual level of commitment to such groups and quality of relationships with community members (Kikuchi & Coleman, 2012).

Third, our narrow focus on campaign-specific community activities is in line with communication scholars’ recent usage of social capital (Southwell et al., 2010; Viswanath, Kosicki, Fredin, & Park, 2000). Scholars have contended that social capital includes the structural dimension (e.g., a tie or link among community residents) and the cognitive dimension (e.g., social trust and norms of reciprocity; see Harpham, Grant, & Thomas, 2002). In contrast, others (e.g., Edwards & Foley, 1998; Lin, 2001) have insisted that social trust and norms of reciprocity belong to cultural capital rather than social capital or that social trust and norms of reciprocity should be antecedent to social capital rather than social capital. Adopting the latter approach, communication scholars (e.g., Southwell et al., 2010; Viswanath et al., 2000) have recently argued for employing only local community ties as a valid indictor of social capital.

Empirical evidence has confirmed that social capital functions as one of the most valuable resources that health campaigns can create or mobilize in a community (Kawachi et al., 2008; Thorson & Beaudoin, 2004; Viswanath et al., 2006). As outlined at the outset of this study, for instance, in the context of antidrug campaign, parents who worked with their community members for the purpose of reducing illegal drug use among youth were found to be more likely to talk about drugs with their offspring (Lee, 2014). Therefore, media health campaigns increasingly try to include extensive public-relations and community-outreach activities in order to mobilize campaign-specific social capital by convincing community residents to participate in planning, designing, and diffusing campaign messages, and to build and maintain health-promoting community environments (Biglan et al., 2012; Merzel & D’Afflitti, 2003).

Notably, we limit our attention only to individual-level social capital because empirical research (e.g., Lee, 2014; Thorson & Beaudoin, 2004) has provided more compelling evidence for the critical roles of social capital in media health campaigns at an individual level than at a community level although social capital can be validly captured and analyzed at both individual and macro-social levels (Kawachi et al., 2008).

Applying TPB to Campaign-Specific Social Capital

To understand the processes underlying the effects of campaign-specific social capital on conversations about campaigns or campaign topics, we use one of the most widely used behavioral change theories, TPB, for the following reasons. First, given that campaign-related conversations do not just serve as a channel for information delivery among individuals but are a “relationally and socially consequential behavior” (Southwell & Yzer, 2007, p. 422), TPB is applicable to targeted parent-child communication as “a general theory that could be used to predict, explain, and influence behavior in any domain” (Fishbein & Ajzen, 2010, p. 17). Assuming that “human social behavior is really not that complicated, that people approach different kinds of behavior in much the same way, and that the same limited set of constructs can be applied to predict and understand any behavior of interest,” Fishbein and Ajzen (2010, p. 2) argued for the valid application of TPB to any type of human behaviors, including parent-child communication about drugs.

Second, theoretical discussions of social capital’s effects on health outcomes have paid attention to the following roles played by social capital: (a) increasing knowledge about health issues, (b) clarifying health-related social norms, and (c) building skills and abilities necessary to engage in pro-health behaviors (Kawachi et al., 2008; Viswanath et al., 2006). These mechanisms exactly match three major components of TPB: behavioral attitude, subjective norm, and perceived behavioral control (PBC), which will be fully elaborated below.

Fishbein and Ajzen (2010) developed theory of reasoned action (TRA), the original formulation of TPB, to explain why a person does or does not perform a given behavior (for an overview, see Fishbein & Ajzen, 2010). TRA posits that the most important proximal predictor of behavior is an intention to engage in that behavior. Behavioral intention, in turn, is a function of attitude toward performing the behavior and subjective norm to engage (or not to engage in) the behavior. Attitude toward the behavior is the overall favorable or unfavorable predispositions toward performing the behavior. Subjective norm is the perception of the degree to which relevant significant others approve or disapprove of performing the behavior. TRA was later extended to TPB by adding PBC with respect to performing the behavior to TRA variables (behavioral attitude and subjective norm) as an additional cognitive variable influencing behavior and behavioral intention. PBC refers to the extent to which a person believes that performing the behavior is (or is not) under his or her control.

According to TRA and TPB, health promotion programs should target behavioral, normative, or control beliefs that are the bases of behavioral attitude, subjective norm, and PBC, respectively, to shape, change, or reinforce individuals’ health-related behaviors (Fishbein & Yzer, 2003). The antidrug campaign to some extent employed TRA and TPB as its program theory (Hornik & Yanovitzky, 2003; Orwin et al., 2006). Many antidrug-related community events emphasized that parents should frequently talk about drug-related issues with their offspring (Hornik & Yanovitzky, 2003). Thus, we expect that attending such events promotes parent-child communication about drugs by cultivating antidrug attitude, subjective norm, and PBC, which in turn leads to targeted parent-child communication about drugs. The above considerations lead us to propose the following hypothesis:

Hypothesis 1: Behavioral attitude, subjective norm, and PBC will mediate the effects of antidrug-specific social capital (i.e., antidrug-related community participation) on targeted parent-child communication about drugs.

The proposed model focuses only on behavior itself (i.e., parent-child communication), not behavioral intention for the following considerations. First, behavioral intention is quite useful for health professionals and health communication researchers, especially when it is very hard to directly observe actual behavior (e.g., condom use, risky sexual behaviors). However, parent-child communication is a behavior variable that can be easily assessed in health surveys. Second, contrary to the previously held belief that behavioral intention is a strong predictor of behavior, there have been mixed findings on the relationship between behavioral intention and behavior (for an overview, see Sheeran, 2002). Because our model includes actual behavior measures, adding behavioral intention was not necessary. It is an article of faith among behavioral and social scientists that a theoretical model should be parsimonious by including the most necessary and important components required to describe and explain social phenomena (Jaccard & Jacoby, 2009; Shoemaker, Tankard, & Lasorsa, 2003).

Explaining the Interactive Effects of Social Capital and Campaign Exposure

Based on the findings that connections among individuals within a community (an essence of social capital) either amplify or dampen campaign effects (e.g., Hornik & Yanovitzky, 2003; Southwell et al., 2010), we anticipate that social capital may moderate the indirect effects of campaign exposure on campaign outcomes. To be more specific, effects of health campaigns may be larger for those with low (vs. high) levels of social capital because individuals can rely on either their network members within a community (e.g., family, friends, neighbors) or campaign messages as a source of pro-health information and because media campaign messages and information that can be obtained from community members tend to overlap (Lee, 2009; Valente & Saba, 1998). Antidrug campaign messages might be so short and superficial that they are less likely to provide audiences with information that can complement the health information diffused within a community (Lee, 2014).

However, the opposite pattern of interaction may be the case. That is, one cannot completely rule out the possibility that campaign exposure and social capital have synergistic effects, so that campaign exposure may exert stronger influences on those with high levels of social capital than those with low levels of social capital. This is because social capital facilitates information diffusion (Coleman, 1988).

Given the theoretical plausibility of both scenarios, we base our hypothesis on empirical findings in this area. Several prior studies of media health campaigns have supported our expectation. Valente and Saba (1998) demonstrated that a reproductive health communication campaign in Bolivia was more successful in promoting adoption of modern contraceptives among people whose personal networks (individuals who people discuss important matters with) rarely used these methods than people whose network members consisted of a majority of contraceptive users. These results, however, might have come from the fact that most respondents with personal networks composed of a majority of contraceptive users already adopted family planning methods; thus, campaign exposure’s effects on contraceptive adoption was detected only among those with personal networks composed of few contraceptive users.

More recent studies provide the strongest evidence in favor of our expectation. For instance, Lee (2014) demonstrated that the effect of parents’ antidrug campaign exposure on their conversation about drugs with their offspring was enhanced among parents with low levels of social capital rather than those with high levels of social capital. To be more specific, campaign exposure did not increase targeted parent-child communication about drugs among parents with high levels of social capital (from 3.68 to 3.87) as much as it did for parents with low levels of social capital (from 3.01 to 3.44). These interaction patterns were not because respondents with high levels of social capital had reached the upper bound of the measure of targeted parent-child communication about drugs (a score of “5”). Lee (2009), in a slightly different context, showed that obtaining health information from both TV and the Internet had stronger associations with an index of healthy lifestyle behaviors, which includes no smoking, no binge drinking, fruit and vegetable consumption, and exercising, among respondents with low levels of conversations about health issues with family and friends than those with high levels of such conversations. Among those who had high levels of conversations about health issues, the variance in healthy lifestyle behaviors was not restricted, and the mean (M = .56, SD = 2.79) was not even close to the maximum for the healthy-lifestyle-behaviors measure (7.88). Thus, Lee’s findings do not result from the measurement ceiling either. In light of these studies, we expect that media’s transmission of health information (either media health campaign or news media’s coverage) and audience’s social capital tend to make up for the absence of the other rather than reinforce the effect of the other.

By considering both the previously detected interaction effect between campaign exposure and campaign-specific social capital and the mediation model proposed in the present study, we expect that our mediation model may vary according to levels of social capital. That is, social capital will likely substitute rather than compliment the indirect effects of antidrug campaign exposure on targeted parent-child communication about drugs through behavioral attitude, subjective norm, and PBC. As such, we test a mediated moderation model (Muller, Judd, & Yzerbyt, 2005), which will yield new insights into why campaign exposure and campaign-specific social capital interact and jointly influence campaign-related conversations, a finding reported by Lee (2014). We thus propose:

Hypothesis 2: The effects of campaign exposure on behavioral attitude, subjective norm, and PBC will be larger for those with low levels of campaign-specific social capital (i.e., antidrug-related community participation) than for those with high levels of social capital. In turn, behavioral attitude, subjective norm, and PBC will mediate the interactive effects of campaign exposure and campaign-specific social capital on targeted parent-child communication about drugs.

Method

Sample

This study uses a three-round longitudinal panel data set collected as part of NSPY, an in-home survey designed to be representative of children aged 9 to 18 years and their parents.¹ The purpose of NSPY was to assess the effectiveness of the antidrug campaign, which targeted youth aged 9 to 18 years, their parents, and other influential adults. NSPY questionnaires were administered in respondents’ homes using touch-screen laptop computers (for details, see Orwin et al., 2006). Round 1 (Waves 1 through 3), Round 2 (Waves 4 and 5), and Round 3 (Waves 6 and 7) interviews were administered between November 1999 and June 2001, between July 2001 and June 2002, and between July 2002 and June 2003, respectively.

NSPY achieved a response rate (the screener response rate × roster response rate × interview response rate) of 64.8% for youth and 62.7% for parents at Round 1. For the Round 2 data collection, 92.1% of the youth who were interviewed at Round 1 were eligible (i.e., 18 years or younger, not deceased, not institutionalized, etc.) to participate in the follow-up surveys. Among these youth, 93.6% successfully completed the follow-up surveys. Therefore, the final Round 2 response rate for youth was 55.9%, which is the product of the Round 1 response rate, the percentage of eligible youth from the Round 1 interviews, and the Round 2 interview response rate (64.8 × 92.1 × 93.6). In addition, 92% of the parents who were interviewed at Round 1 were eligible to participate in the follow-up surveys. Among these parents, 92% successfully completed the follow-up surveys. Thus, the final Round 2 response rate for parents was 53% (62.7 × 92 × 92).

For Round 3 data collection, 93.6% of the youth who were interviewed at Round 2 were eligible to participate in the follow-up surveys. Among these youth, 96% successfully completed the follow-up surveys. Therefore, the final Round 3 response rate for youth was 50.2%, which is the product of the Round 2 response rate, the percentage of eligible youths from the Round 2 interviews, and the Round 3 interview response rate (55.9 × 93.6 × 96). In addition, 94.1% of the parents who were interviewed at Round 2 were eligible to participate in the follow-up surveys. Among these parents, 95.4% successfully completed the follow-up surveys. Thus, the final Round 3 response rate for parents was 47.6% (53 × 94.1 × 95.4).

It should be noted that 5,504 parents participated in the NSPY study at Round 1; however, the present investigation’s sample size is 7,620 (sample size of youth at Round 1) for the following reasons. More than one child was able to participate in the study and the dependent variable, targeted parent-child communication about drugs, was specific to each child; thus, the same parent could provide different answers for each child. Since the answers from the same parent may be more similar than could occur by chance, we accounted for clustering effects in the model (see the analyses section).

Measures

Among the measures, higher scores indicate greater general exposure to antidrug campaign ads, active participation in antidrug-related community activities, attitude in favor of talking to one’s child about drugs, believing that important others favor talking to one’s child about drugs, feeling efficacious about talking to one’s child about drugs, and engaging in targeted parent-child communication about drugs. See Table 1 to view the means, standard deviations, bivariate correlations, and Cronbach’s alphas.

Table 1.

Descriptive Statistics and Bivariate Correlations.

Variable	M (SD)	1	2	3	4	5	6
1. Campaign exposure	7.03 (9.00)	—
2. Antidrug-specific social capital	1.22 (1.59)	.17***	—
3. Behavioral attitude	6.25 (0.82)	.10***	.09***	.73
4. Subjective norm	4.51 (0.71)	.09***	.12***	.42***	—
5. Perceived behavioral control	4.54 (0.54)	.05***	.10***	.29***	.24***	.65
6. Targeted parent-child communication about drugs	3.92 (1.35)	.14***	.17***	.24***	.21***	.13***	—

Note. Cronbach’s alpha coefficients are included in the diagonal for multi-item scales. However, because campaign exposure, social capital, and targeted communication are an index rather than a scale, we do not report their reliability coefficients. For all other scales with ordinal measures, we included the means and standard deviations.

***

p < .001, two-tailed.

General exposure to antidrug campaign ads (R1)

General exposure to the campaign ads was measured using the following three questions, based on wording from the Monitoring the Future Survey (see Hornik et al., 2008; Orwin et al., 2006): “In recent months, about how often have you seen such antidrug ads . . . on TV, or heard them on the radio?,” “In recent months, about how often have you seen such antidrug ads in newspapers or magazines?,” and “In recent months, about how often have you seen any antidrug billboards or other public antidrug ads such as on buses, in malls, or at sports events?” Respondents provided their answers on a six-point scale (1 = not at all, 2 = less than 1 time a month, 3 = 1 to 3 times a month, 4 = 1 to 3 times a week, 5 = daily or almost daily, 6 = more than 1 times a day). Although it is common to treat ordinal variables as interval variables in structural equation modeling (SEM), we tried to meet the assumptions of SEM by converting this ordinal measure into a ratio-level measure (Hayes, 2005). That is, we recoded the responses in a way that represents the total number of ads viewed by the respondents, assuming that respondents interpret the recent month as the last month (for details, see Hornik et al., 2008; Orwin et al., 2006; 1 = 0, 2 = 0.5, 3 = 2.0, 4 = 8.0, 5 = 30.0, 6 = 45.0). Then, we averaged these scores and created an index of general exposure to antidrug campaign ads.

Antidrug-specific social capital (R1)

Parent’s antidrug-specific social capital was operationalized as an additive index of nine dichotomous items (0 = no, 1 = yes), asking respondents the following questions: (a) “Did you attend a meeting or rally in support of drug prevention?”; (b) “Did you join an active group in support of drug prevention?”; (c) “Did you attend a drug-prevention activity sponsored by PTA?”; (d) “Did you attend a drug-prevention activity sponsored by church?”; (e) “Did you attend a drug-prevention activity sponsored by a neighborhood group?”; (f) “Did you attend a drug-prevention activity sponsored by the local police?”; (g) “Did you attend drug-prevention activity sponsored by community group?”; (h) “Did you attend a drug-prevention activity sponsored by other organizations?”; and (i) “Did you attend a parent effectiveness workshop?”²

TPB variables (R2)

Based on TPB, this study included three possible mediators of the relationship between exposure to antidrug campaign ads and targeted parent-child communication about drugs: attitude toward talking about drugs, subjective norm in favor of talking about drugs, and PBC over talking about drugs.

First, attitude toward talking about drugs with one’s child was operationalized using the following three questions: “How good do you think it is to discuss drug use with your child in the next 12 months?” (1 = extremely bad, 2 = bad, 3 = somewhat bad, 4 = neither good nor bad, 5 = somewhat good, 6 = good, 7 = extremely good); “How pleasant do you think it is to discuss drug use with your child in the next 12 months?” (1 = extremely unpleasant, 2 = unpleasant, 3 = somewhat unpleasant, 4 = neither pleasant nor unpleasant, 5 = somewhat pleasant, 6 = pleasant, 7 = extremely pleasant); “How important do you think it is to discuss drug use with your child in the next 12 month?” (1 = extremely unimportant, 2 = unimportant, 3 = somewhat unimportant, 4 = neither important nor unimportant, 5 = somewhat important, 6 = important, 7 = extremely important). Cronbach’s alpha was .73.

Second, subjective norm about talking about drugs with one’s child was measured by asking respondents to what extent they thought that important people think they should talk with their child about drugs in the next 12 months (1 = definitely should not, 2 = should not, 3 = neither should nor should not, 4 = should, 5 = definitely should).

Third, PBC over talking about drugs with one’s child was measured with three questions on the five-point scale (1 = very unsure, 2 = unsure, 3 = neither sure nor unsure, 4 = sure, 5 = very sure): “How sure are you that you would be able to talk with your child about drug use in general?” “How sure are you that you would be able to talk with your child about how to avoid drugs?” and “How sure are you that you would be able to talk with your child about drugs if your relationship is tense?” Cronbach’s alpha was .65.³

Targeted parent-child communication about drugs (R2 and R3)

This index was measured using five questions, the first being “How often did you talk to your child about drugs in the last 6 months?” (1 = never, 2 = once, 3 = 2 to 3 times, 4 = 4 to 5 times, 5 = 6 to 10 times, 6 = more than 10 times). Parents were then asked, “Did you talk with your child in the last 6 months about . . . ?” (a) “ . . . family rules about drugs?” (b) “ . . . how to avoid drugs?” (c) . . . others who have gotten in trouble because of drugs?” and (d) “ . . . drug use in movies, music, and on TV?” (0 = no, 1 = yes). After recoding the first item as dichotomous one (never = 0, more than once = 1), we composed an additive index.

Control variables (R1)

We controlled for parent’s age (M = 40.69, SD = 7.15), gender (65.7% females), education (median: some college), household income (median: between $35,000 and $49,999), race/ethnicity (67.2% White, 14.9% African American, 14.1% Hispanic, 3.8% Other), and past drug use at R1. Past drug use behavior was measured by asking parents whether they had ever used marijuana before, whether they had ever used inhalants, and whether they had ever used an illegal drug other than marijuana or inhalants. Then, we summed these three dichotomous items (M = .83, SD = .88). Other types of parent’s exposure to drug-related information at R1 were also controlled. First, parent’s antidrug-related mass media use was measured by asking respondents whether they recalled stories that dealt with drug use among young people in recent months in a variety of media sources (e.g., news on TV or radio, magazines, and newspapers; 0 = no, 1 = yes; 64% said yes). Second, parent’s antidrug-related Internet use was constructed by adding the following two yes/no questions: “In the last 6 months, have you ever visited any website that talked about drug use?” (7.3% yes) and “In the last 6 months, have you ever visited any website that talked about parenting skills?” (8.6% yes; r = .82, p < .001).

Moreover, child’s gender (51.9% female), age (M = 12.61, SD = 2.55), and drug use at R1 were included as a control variable.⁴ Child’s drug use was measured by asking whether he or she had ever used cigarettes, drunk alcohol, or smoked marijuana before. A summed scale was then created from these dichotomous items (M = .58, SD = .97).

In addition, the SEM model controlled for prior targeted parent-child communication about drugs (measured at R2).

Statistical analyses

In Mplus (Muthén & Muthén, 2007), we used SEM to validate our measures and to test our hypotheses. Model fit can be described as excellent or acceptable based on the following criteria. An excellent-fitting model should have a root mean square error of approximation (RMSEA) of ≤.06 and a comparative fit index (CFI) of ≥.95 (Hu & Bentler, 1999). An acceptably fitting model should have a RMSEA < .08 and a CFI can be ≥.90 (Beaudoin & Thorson, 2006; Holbert & Stephenson, 2008; Kam & Middleton, 2013). In additioin, a standardized root mean square residual (SRMR) should be < .08 (Hu & Bentler, 1999). To handle the missing data across rounds, we used the full information maximum likelihood (FIML), which is better than listwise deletion because the latter often results in a substantial loss of power and can produce biased parameter estimates (Graham, 2009). As outlined earlier, because parent’s answers to targeted parent-child communication about drugs were nested within household, we used Type = Complex in Mplus to handle the multilevel-structured data. Type = Complex calculates the standard errors and a chi-square test while considering the nonindependence of observations (Muthén & Muthén, 2007).

To assess indirect effects (Hypothesis 1), direct effects were modeled from social capital to behavioral attitude, subjective norm, and PBC, as well as to targeted parent-child communication about drugs. Direct effects also were modeled from attitude, norm, and efficacy to targeted parent-child communication. RMediation was used to obtain the asymmetric 95% confidence intervals (CIs; Tofighi & MacKinnon, 2011). RMediation handles the non-normality in the product of coefficients’ distribution and computes asymmetric 95% CIs. Indirect effects were considered significant when zero was not within the 95% asymmetric CI. To accurately employ RMediation, we multiplied the mediating and dependent variables by 10 (MacKinnon, personal communication, June 24, 2013).

To test Hypothesis 2, we conducted a mediated moderation analysis. We first created an interaction term between campaign exposure and social capital after centering each variable. Then, we entered this interaction term into the model as an independent variable along with two main-effect variables, campaign exposure and social capital (see Figure 2).

Figure 1.

Social capital’s indirect effect on targeted parent-child communication about drugs.

Figure 2.

A mediated moderation model: The interactive effect of campaign exposure and social capital on targeted parent-child communication about drugs.

Results

Prior to examining the full structural model, we analyzed a measurement model that included latent factors of behavioral attitude and PBC with their corresponding indicators. These latent factors were correlated. Subjective norm was excluded from this model because it was operationalized with one observed variable. We treated campaign exposure, social capital, and targeted parent-child communication about drugs as an index defined by causal indicators and not a scale defined by effects indicators (Bollen & Lennox, 1991; Streiner, 2003). With behavioral attitude and PBC, the full measurement model yielded an adequate fit, χ²(8) = 242.456, p < .001; RMSEA = .062, 90% CI = [.055, .069]; CFI = .934; SRMR = .051, with standardized factor loadings greater than .49.

After examining the measurement model, we inspected the full structural model. As seen in Figure 1, this model fit the data acceptably, χ²(80) = 825.014, p < .001; RMSEA = .035, 90% CI = [.033, .037]; CFI = .921; SRMR = .024. When taking into account the control variables, the model explained 15.3% of the variance in behavioral attitude, 10.3% of the variance in subjective norm, 5.9% of the variance in PBC, and 24.2% of the variance in targeted parent-child communication about drugs.

A deeper inspection into the results revealed that antidrug-specific social capital was positively related to behavioral attitude (b = .20, SE = .07, β = .04, p < .01), subjective norm (b = .26, SE = .07, β = .06, p < .001), perceived behavioral control (b = .18, SE = .04, β = .07, p < .001), and targeted parent-child communication about drugs (b = .71, SE = .12, β = .08, p < .001). Yet, only behavioral attitude was significantly related to targeted parent-child communication, and in the positive direction (b = .19, SE = .04, β = .11, p < .001). The asymmetric 95% CIs revealed that social capital exhibited a significant indirect effect on targeted parent-child communication through parents’ pro-attitude toward talking to their child about drugs (95% CI = [.001, .007]). In contrast, social capital did not exhibit significant indirect effects on parent-child communication through PBC (95% CI = [−.001, .004]) or subjective norm (95% CI = [0.00, .004]). Thus, only behavioral attitude functioned as a significant mediator.

Figure 2 presents the results of our test of H2. Overall, the mediated moderation model fit the data acceptably, χ²(84) = 853.981, p < .001; RMSEA = .035, 90% CI = [.033, .037]; CFI = .920; SRMR = .023. The effect of campaign exposure on behavioral attitude was moderated by social capital (b = −.02, SE = .01, β = −.07, p < .05), partially supporting Hypothesis 2. This finding demonstrates that the effect of campaign exposure on behavioral attitude was larger for parents with low levels of social capital.⁵ To gain a clearer understanding of the interaction, model parameter estimates (i.e., unstandardized coefficients after controlling for all confounders) were used to predict behavioral attitude by social capital and campaign exposure (see Figure 3). According to the recommendations by Hayes (2005), we computed the effects of Round 1 campaign exposure on Round 2 behavioral attitude at one standard deviation below the mean levels of social capital, at the mean levels of social capital, and at one standard deviation above the mean levels of social capital. A positive relationship between campaign exposure at Round 1 and behavioral attitude at Round 2 was apparent only among respondents with the average or lower than the average level of social capital. Among respondents with above-average levels of social capital, the relationship between campaign exposure and behavioral attitude was generally flat. No other paths were statistically significantly moderated by social capital. The interactive effect of campaign exposure and social capital on targeted parent-child communication about drugs can be explained by behavioral attitude (95% CI = [−.008, −.001]).

Figure 3.

Relations between campaign exposure and social capital at Round 1 and attitude at Round 2.

Discussion

This study yields useful insights into why antidrug-specific social capital moderates the effect of antidrug campaign exposure on targeted parent-child communication about drugs. In our examination of the mediated moderation model proposed in Hypothesis 2, we found that the antidrug campaign manifested its influences more among parents with low levels of social capital than parents with high levels of social capital because the campaign increased parents’ attitude toward talking about drugs with their child more among the former than the latter. Notably, there is no evidence that this result occurred because respondents with high levels of social capital had reached the higher bound of our measure of behavioral attitude (a score of “7” indicating strong favorable attitude). As Figure 3 shows, it is clear that there still remained room for those with high levels of social capital to improve their attitude by gaining more antidrug-related information from the campaign.

Using Health Campaigns to Promote Targeted Parent-Child Communication About Drugs

This study offers theory-based, practical guidelines for future campaigns to promote parent-child communication about drugs. Although there have been numerous media intervention efforts to encourage parents to talk about drugs with their offspring, these campaigns have not been firmly based on theory and research (e.g., Mahabee-Gittens, Huang, Slap, & Gordon, 2007; Surkan, DeJong, Herr-Zaya, Rodriguez-Howard, & Fay, 2003). By employing TPB as a theoretical framework, we found that behavioral attitude rather than subjective norm and PBC predicts targeted parent-child communication about drugs. Though not focusing on parent-child communication, previous meta-analyses of TRA and TPB have reported similar findings (e.g., Albarracín, Johnson, Fishbein, & Muellerleile, 2001; Armitage & Conner, 2001). Our results suggest that there is still much room for improvement in media health campaigns to promote parents’ initiation and continuation of drug-related conversations with their offspring. Previous campaigns either conveyed medical facts to increase parents’ knowledge about dangers of and harmful effects of youth drug use or tried to supply parents with skills and abilities necessary to engage in parent-child communication (e.g., Mahabee-Gittens et al., 2007; Surkan et al., 2003). Although some of these campaigns were found to be successful, the present study leads us to contend that future intervention efforts may achieve more by putting greater emphasis on specific benefits and costs of drug-related conversations (attitudinal beliefs), thereby persuading parents to form or reinforce positive attitudes toward parent-child communication.

While TPB emphasizes the importance of behavioral attitude, subjective norm, and PBC, behavioral attitude was the only significant mediator. Some explanations may exist for the primary role that behavioral attitude plays in predicting targeted parent-child communication about drugs. First, as Ybarra and Trafimow (1998) suggested, in such an individualistic society as the United States, parents seem to behave as an independent agent, so that they can be persuaded to have drug-related conversations with their offspring only after realizing the positive outcomes that they can expect from targeted parent-child communication, regardless of social pressure from their significant others or confidence in the ability to engage in that behavior. Second, our findings may be because behavioral attitude serves as a complete mediator between PBC, subjective norm, and targeted parent-child communication about drugs. In other words, the effects of subjective norm and PBC on targeted parent-child communication might be indirect and can only be exerted through behavioral attitude. These explanations, while speculative, invite future studies to further investigate the relations among behavioral attitude, subjective norm, and PBC in the context of antidrug campaign and targeted parent-child communication about drugs.

Social Capital Matters in Health Campaigns

Before discussing the roles of social capital in campaign processes and effects, it would be useful to mention the trend of public health interventions that has continued since the 1970s (Merzel & D’Afflitti, 2003; Viswanath & Emmons, 2006). In a very general sense, media health campaigns have over time become community-level interventions extending beyond merely transmitting public service announcements through mass media. Against this backdrop, it should be highlighted that, even after controlling for campaign exposure and other confounding factors, antidrug-related community participation increased behavioral attitude, subjective norm, and PBC over time, which in turn promoted targeted parent-child communication about drugs. Thus, we are encouraging a health promotion program to increase its efforts to create and strengthen local community ties and to mobilize community members as an efficient means for enhancing the program impact (Stephens et al., 2004; Viswanath et al., 2006).

The effects of antidrug-specific social capital at Round 1 on attitude, subjective norm, and PBC at Round 2, however, should be interpreted with caution because, as Eveland and Morey (2011, p. 23) observed, “The length of time necessary for a cause to produce an effect may be the most central concern in the design of a panel study” (see also Slater, 2006). We expected that cumulative attendance at multiple community events may be required for antidrug-specific social capital to produce any influences on antidrug-related cognitions. In other words, we did not assume that antidrug-specific social capital may produce immediate responses among target audience because “individuals will take repeated convincing before they are ready to change” (Hornik & Yanovitzky, 2003, p. 210). If so, the time lag between rounds adopted in this study (i.e., 1 year) may be reasonable. In contrast, one may conjecture that social capital’s effects on behavioral attitude, subjective norm, and PBC are short-lived and brief, instead of cumulative or sustained in its effect. Then, it may be that the significant relationships between Round 1 antidrug-specific social capital and Round 2 behavioral attitude, subjective norm, and PBC result from stability in social capital over time. That is, Round 1 social capital may serve as an error-ridden proxy for Round 2 social capital, thereby affecting attitude, subjective norm, and PBC at Round 2. If this is the case, our causal claims might be somewhat compromised. To effectively capture such immediate, concurrent (or “ongoing but transient”) effects of social capital, future research should employ time-series or tracking design (Slater, 2006, p. 153).

More interestingly, the effect of campaign exposure on attitude toward parent-child communication about drugs was found to be larger for parents with low levels of social capital than for those with high levels of social capital. This indicates that media health campaigns can be a very effective tool for cultivating positive attitude toward parent-child communication about drugs especially among parents who do not actively participate in antidrug-related community events. Since behavioral attitude was positively related to targeted parent-child communication about drugs, health educators and public health practitioners should consider employing media campaigns to reduce the existing gap in effective parenting strategies, such as targeted parent-child communication about drugs, by social capital.

However, we should also acknowledge that according to a group of political communication scholars, the effects of mass media use to obtain public-affairs information on political participation were larger for people who frequently talked about politics with others than for those who rarely discussed politics (e.g., Scheufele, 2002). This also was the case in regard to Internet use and online interactions (e.g., Hardy & Scheufele, 2005). Since political participation inherently needs substantial knowledge or deep understanding of political issues, which requires much “internal reflective integration” and “external reflective integration” of media content (i.e., interpersonal discussion of politics; see Sotirovic & McLeod, 2001), even those who often talk about politics still have more room to increase their participatory behaviors by gaining more political information from media channels (Hardy & Scheufele, 2005; Scheufele, 2002). In contrast, there may be a great deal of overlap between antidrug-related information conveyed through media campaigns and the information diffused through local community ties, so that the effects of one source may substitute for those of the other (for details, see Lee, 2009, 2014). These differential patterns of interaction suggest that scholars should theorize the interaction between media exposure and conversation depending on the type of behavior and mediated information (e.g., political vs. health).

Theorizing the Connection Between Media Exposure and Conversation

Given that mass media are quite useful in raising issue awareness and knowledge, but conversation is important for actual behavior change to occur (Rogers, 2003; Valente & Saba, 1998), many media campaigns intend to increase conversations about campaigns or campaign topics among target audience. However, little theoretically derived research has been conducted on the psychological mechanisms that may underlie the effect of campaign exposure on conversation. Even a handful of studies addressing this issue have concentrated only on one or two psychological predictors of conversation, while omitting other, potentially equally important or even stronger ones. Such an omission prevents researchers from rigorously evaluating the importance of each determinant of conversation, thereby impeding theoretical advancement in this area of research. That is, though having improved our understanding of the connection between media exposure and conversation, previous studies are limited in terms of being synthesized into a coherent and systematic theory.

To redress the aforementioned problems, we used TPB to account for the effects of antidrug campaign exposure on campaign-relevant targeted parent-child communication for parents exhibiting low levels of campaign-specific social capital. The present study joins a recent group of projects that have taken a social cognitive approach and identified audience’s psychological variables, such as conversation norms, conversational competency, and perceived understanding of the topic, as a mediator of media effects on conversations about media content (e.g., Southwell & Torres, 2006). In so doing, this study contributes to existing research on how media exposure is connected to conversation.

In addition, we note that this article contributes to TPB by identifying an important distal variable, campaign-specific social capital. Under the banner of distal variables, Fishbein and his colleagues listed a variety of factors that have significant influences on important, proximal predictors of behavioral intention and behavior (Fishbein & Yzer, 2003; Yzer et al., 2004). These include macro-social variables, such as culture, as well as individual-level variables, such as demographic variables; attitudes toward targets; personality, mood, and emotions; other individual difference variables; and exposure to media and other interventions (Fishbein & Yzer, 2003; Yzer et al., 2004). Although not being explicitly mentioned, we argue that social capital can be considered as one of the distal variables because of positive influences of social capital on health-related cognitions and behaviors (for an overview, see Kawachi et al., 2008).

Limitations and Future Research

This study is not without limitations. First, our measure of subjective norm is a single-item variable, which does not allow us to control for potential measurement error. However, this is a tradeoff of greater statistical power for inflexibility in choosing measures to test our hypotheses. In exchange for our inability to use several-item measures, which can be obtained in smaller, investigator-designed surveys, we were able to gain substantial statistical power to detect even small effects by using a large-scale nationally representative data set collected for other purposes.

Second, our measure of general campaign exposure is somewhat limited. The campaign exposure measure is quite comprehensive because it taps one’s exposure to campaign messages through diverse channels available in the information environment rather than just via mass media. Nevertheless, this measure might be confounded with one’s involvement in drug-related issues. Although we tried to mitigate this concern by controlling for factors that might be associated with issue interest or issue involvement, we might not have been able to exhaustively eliminate third-variable threats.

Third, our study does not include any beliefs as predictors of behavioral attitude, subjective norm, and PBC. Although the present study suggests that antidrug campaign ads should target behavioral attitude to increase targeted parent-child communication about drugs especially for those who are socially isolated and thus fail to receive social support, it provides just a starting point for designing effective campaign messages since one needs to further identify determinants of behavioral attitude. Campaign planners and designers should figure out what attitudinal beliefs we should use in campaign ads.

Finally, the sizes for the social capital’s effects on behavioral attitude and for the interaction of campaign exposure and social capital (measured in terms of standardized coefficients) are small. However, we contend that small effects can be very meaningful in several respects. First, small effects sizes should not be ignored if researchers are interested in audiences’ overtime, cumulative exposure to media content, when a particularly study is only able to track these effects over a relatively short period of time (in this case, two years; see Hornik & Yanovitzky, 2003). Second, even small effects deserve serious attention if they have significant public health implications (Fishbein, 1996). Third, in estimating media effects with the potential to reach a large proportion of the population, one should consider both the magnitude of the effect per exposure (as captured by standardized coefficients) and the magnitude of that exposure (as captured by the proportion of the population being exposed to antidrug campaign messages). In each of these criteria, effect sizes for exposure to antidrug campaign ads and social capital would appear practically significant even if small. Thus, in judging the overall significance of any findings, one should consider practical implications of the findings and its contribution to the literatures above and beyond statistical effect size and variance explained.

Conclusion

The antidrug campaign offered an invaluable opportunity to examine the nexus of media campaign, campaign-related social capital, and conversations. Campaign-specific social capital was found to facilitate campaign-related conversations by inducing positive attitudes toward those conversations. Moreover, among parents with low levels of social capital, the effects of campaign exposure on campaign-related conversations were mediated by attitudes. These findings have wide-ranging implications for the ongoing debates regarding the roles of social capital in media effects in general and health campaigns in particular. Our theoretical framework also contributes to the endeavors to theorize and test connections between media exposure and conversation in such disciplines as communication, sociology, and social psychology.

Footnotes

Authors’ Note

The data reported herein were collected under the auspices of NIDA contract N01DA-8-5063. The data were subsequently obtained via the National Survey of Parents and Youth Center (NIDA contract N01DA-5-5532).

Declaration of Conflicting Interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Notes

Author Biographies

Chul-joo Lee (PhD, University of Pennsylvania, 2009) is Assistant Professor in the Department of Communication at Seoul National University. His research interests include exploring the interplay of mass media health information and social capital or social networks, and investigating the roles that the Internet may play in the evolving health care system.

Jennifer A. Kam (PhD, The Pennsylvania State University, 2009) is Assistant Professor in the Department of Communication at University of California, Santa Barbara. Her research interests include examining the effects of cultural stressors on adolescents’ substance use and identifying the protective nature of interpersonal communication.

References

Albarracín

Johnson

B. T.

Fishbein

Muellerleile

P. A.

(2001). Theories of reasoned action and planned behavior as models of condom use: A meta-analysis. Psychological Bulletin, 127, 142-161.

Armitage

C. J.

Conner

(2001). Efficacy of the theory of planned behavior: A meta-analytic review. British Journal of Social Psychology, 40, 471-499.

Beaudoin

C. E.

Thorson

(2006). The social capital of blacks and whites: Differing effects of the mass media in the United States. Human Communication Research, 32, 157-177.

Biglan

Flay

B. R.

Embry

D. D.

Sandler

I. N.

(2012). The critical role of nurturing environments for promoting human well-being. American Psychologist, 67, 257-271.

Bollen

Lennox

(1991). Conventional wisdom on measurement: A structural equation perspective. Psychological Bulletin, 110, 305-314.

Bourdieu

(1986). The forms of capital. In Richardson

J. G.

(Ed.), Handbook of theory and research for the sociology of education (pp. 241-258). New York, NY: Greenwood Press.

Coleman

(1988). Social capital in the creation of human capital. American Journal of Sociology, 94, 95-120.

Edwards

Foley

(1998). Civil society and social capital beyond Putnam. American Behavioral Scientist, 42, 124-139.

Ellaway

Macintyre

(2007). Is social participation associated with cardiovascular disease risk factors? Social Science & Medicine, 64, 1384-1391.

10.

Eveland

W. P.

Jr. Morey

A. C.

(2011). Challenges and opportunities of panel designs. In Bucy

E. P.

Holbert

R. L.

(Eds.), Sourcebook for political communication research: Methods, measures, and analytical techniques (pp. 19-33). New York, NY: Routledge.

11.

Fishbein

(1996). Editorial: Great expectations, or do we ask too much from community-level interventions? American Journal of Public Health, 86, 1075-1076.

12.

Fishbein

Ajzen

(2010). Predicting and changing behavior: The reasoned action approach. New York, NY: Taylor & Francis.

13.

Fishbein

Yzer

M. C.

(2003). Using theory to design effective health behavior interventions. Communication Theory, 13, 164-183.

14.

Graham

J. W.

(2009). Missing data analysis: Making it work in the real world. Annual Review of Psychology, 60, 549-576.

15.

Hardy

B. W.

Scheufele

D. A.

(2005). Examining differential gains from Internet use: Comparing the moderating role of talk and online interactions. Journal of Communication, 55, 71-84.

16.

Harpham

Grant

Thomas

(2002). Measuring social capital within health surveys: Key issues. Health Policy and Planning, 17, 106-111.

17.

Hayes

A. F.

(2005). Statistical methods for communication science. Mahwah, NJ: Lawrence Erlbaum.

18.

Hess

(2013). Tertius Tactics: “Mediated social capital” as a resource of power for traditional commercial news media. Communication Theory, 23, 112-130.

19.

Holbert

R. L.

Stephenson

M. T.

(2008). Commentary on the uses and misuses of structural equation modeling in communication research. In Hayes

A. F.

Slater

M. D.

Snyder

L. B.

(Eds.), The SAGE sourcebook of advanced data analysis methods for communication research (pp. 185-218). Thousand Oaks, CA: Sage.

20.

Hornik

R. C.

Jacobsohn

Orwin

Piesse

Kalton

(2008). Effects of the National Youth Anti-Drug Media Campaign on youths. American Journal of Public Health, 98, 2229-2236.

21.

Hornik

R. C.

Yanovitzky

(2003). Using theory to design evaluations of communication campaigns: The case of the National Youth Anti-Drug Media Campaign. Communication Theory, 13, 204-224.

22.

Hoyle

R. H.

Kenny

D. A.

(1999). Sample size, reliability, and tests of statistical mediation. In Hoyle

R. H.

(Ed.), Statistical strategies for small sample research (pp. 195-222). Thousand Oaks, CA: Sage.

23.

L. T.

Bentler

P. M.

(1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling, 6, 1-55.

24.

Jaccard

Jacoby

(2009). Theory construction and model-building skills: A practical guide for social scientists. New York, NY: Guilford Press.

25.

Kam

J. A.

Middleton

A. V.

(2013). The associations between parents’ references to their own past substance use and youth’s substance-use beliefs and behaviors: A comparison of Latino and European American youth. Human Communication Research, 39, 208-229.

26.

Kawachi

Subramanian

S. V.

Kim

(2008). Social capital and health. New York, NY: Springer.

27.

Kikuchi

Coleman

C.-L.

(2012). Explicating and measuring social relationships in social capital research. Communication Theory, 22, 187-203.

28.

Lee

C.-J.

(2009). The interplay between media use and interpersonal communication in the context of healthy lifestyle behaviors: Reinforcing or substituting? Mass Communication and Society, 13, 48-66.

29.

Lee

C.-J.

(2014). The role of social capital in health communication campaigns: The case of the National Youth Anti-drug Media Campaign. Communication Research, 41, 208-235.

30.

Lin

(2001). Building a network theory of social capital. In Lin

Cook

Burt

(Eds.), Social capital: Theory and research (pp. 3-29). New York, NY: Aldine De Gruyter.

31.

Mahabee-Gittens

E. M.

Huang

Slap

G. B.

Gordon

J. S.

(2007). An emergency department intervention to increase parent-child tobacco communication: A pilot study. Journal of Child & Adolescent Substance Abuse, 17, 71-83.

32.

Merzel

D’Afflitti

(2003). Reconsidering community-based health promotion: Promise, performance, and potential. American Journal of Public Health, 93, 557-574.

33.

Muller

Judd

C. M.

Yzerbyt

V. Y.

(2005). When moderation is mediated and mediation is moderated. Journal of Personality and Social Psychology, 89, 852-863.

34.

Muthén

L. K.

Muthén

B. O.

(2007). Mplus user’s guide (5th ed.). Los Angeles, CA: Author.

35.

Orwin

Cadell

Chu

Kalton

Maklan

Morin

. . .Tracy

(2006). Evaluation of the National Youth Anti-Drug Media Campaign: 2004 report of findings. Washington, DC: Westat.

36.

Putnam

R. D.

(2000). Bowling alone: The collapse and revival of American community. New York, NY: Simon & Schuster.

37.

Rogers

E. M.

(2003). Diffusion of innovations (5th ed.). New York, NY: The Free Press.

38.

Sampson

R. J.

Morenoff

J. D.

Earls

(1999). Beyond social capital: Spatial dynamics of collective efficacy for children. American Sociological Review, 64, 633-660.

39.

Scheufele

D. A.

(2002). Examining differential gains from mass media and their implications for participatory behavior. Communication Research, 29, 46-65.

40.

Shah

D. V.

Rojas

Cho

(2009). Media and civic participation: On understanding and misunderstanding communication effects. In Bryant

Oliver

M. B.

(Eds.), Media effects: Advances in theory and research (3rd ed., pp. 207-227). New York, NY: Routledge.

41.

Sheeran

(2002). Intention-behavior relations: A conceptual and empirical review. European Review of Social Psychology, 12, 1-36.

42.

Shoemaker

P. J.

Tankard

J. W.

Lasorsa

D. L.

(2003). How to build social science theories. Thousand Oaks, CA: Sage.

43.

Slater

M. D.

(2006). Specification and misspecification of theoretical foundations and logic models for health communication campaigns. Health Communication, 20, 149-157.

44.

Sotirovic

McLeod

J. M.

(2001). Values, communication behavior, and political participation. Political Communication, 18, 273-300.

45.

Southwell

B. G.

Slater

J. S.

Rothman

A. J.

Friedenberg

L. M.

Allison

T. R.

Nelson

C. L.

(2010). The availability of community ties predicts likelihood of peer referral for mammography: Geographic constraints on viral marketing. Social Science & Medicine, 71, 1627-1635.

46.

Southwell

B. G.

Torres

(2006). Connecting interpersonal and mass communication: Science news exposure, perceived ability to understand science, and conversation. Communication Monographs, 73, 334-350.

47.

Southwell

B. G.

Yzer

M. C.

(2007). The roles of interpersonal communication in mass media campaigns. Communication Yearbook, 31, 419-462.

48.

Stephens

K. K.

Rimal

R. N.

Flora

J. A.

(2004). Expanding the reach of health campaigns: Community organizations as meta-channel for the dissemination of health information. Journal of Health Communication, 9, 97-111.

49.

Streiner

(2003). Being inconsistent about consistency: When coefficient alpha does and doesn’t matter. Journal of Personality Assessment, 80, 217-222.

50.

Surkan

P. J.

DeJong

Herr-Zaya

K. M.

Rodriguez-Howard

Fay

(2003). A paid radio advertising campaign to promote parent-child communication about alcohol. Journal of Health Communication, 8, 489-495.

51.

Thorson

Beaudoin

(2004). The impact of a health campaign on health social capital. Journal of Health Communication, 9, 167-194.

52.

Tofighi

MacKinnon

D. P.

(2011). RMediation: An R package for mediation analysis confidence intervals. Behavior Research Methods, 43, 692-700.

53.

Valente

T. W.

Saba

W. P.

(1998). Mass media and interpersonal influence in a reproductive health communication campaign in Bolivia. Communication Research, 25, 96-124.

54.

Viswanath

Emmons

K. M.

(2006). Message effects and social determinants of health: Its application to cancer disparities. Journal of Communication, 56, S238-S264.

55.

Viswanath

Kosicki

G. M.

Fredin

E. S.

Park

(2000). Local community ties, community-boundedness, and local political affairs knowledge gaps. Communication Research, 27, 27-50.

56.

Viswanath

Steele

W. R.

Finnegan

J. R.

(2006). Social capital and health: Civic engagement, community size, and recall of health messages. American Journal of Public Health, 96, 1456-1461.

57.

Ybarra

Trafimow

(1998). How priming the private self or collective self affects the relative weights of attitudes and subjective norms. Personality and Social Psychology Bulletin, 4, 362-370.

58.

Yzer

M. C.

Cappella

J. N.

Fishbein

Hornik

Sayeed

Ahern

R. K.

(2004). The role of distal variables in behavior change: Effects of adolescents’ risk for marijuana use on intention to use marijuana. Journal of Applied Social Psychology, 34, 1229-1250.