Printer Friendly

The effects of advertising, social influences, and self-efficacy on adolescent tobacco use and alcohol consumption.

Exploring the simultaneous effects of key variables on the unhealthy consumption behavior of adolescents, two studies focused on the relative effects of advertising, parental and peer influence, and self-efficacy on adolescent tobacco use and alcohol consumption. The results suggest that (1) advertising effects are largely neutralized by parental and peer influence; (2) peer and parental influence strongly predict adolescent tobacco use and alcohol consumption; and (3) self-efficacy is a weak predictor of both adolescent risk behaviors.

**********

In response to the dire consequences of tobacco use and alcohol consumption in the adolescent population, recent research has sought to explore the causes or predictors of these unhealthy consumption behaviors (e.g., Andrews et al. 2004; Ellickson et al. 2005; Smith and Stutts 1999). In general, previous studies have revealed myriad environmental, social, and cognitive predictors related to the etiology of adolescent tobacco and alcohol use. For instance, one stream of research has found that advertising strongly influences adolescents' decisions to smoke cigarettes and/or drink alcohol (e.g., Grube and Wallack 1994; Pechmann and Knight 2002). A second body of literature, which focuses on either parental or peer substance use, found that each plays a major role in increasing the likelihood that children and adolescents will engage in similar substance use (e.g., Brown, Clasen, and Eicher 1986; Fawzy, Coombs, and Gerber 1983). A third group of studies investigated the effects of self-efficacy, a cognitive construct that pertains to the extent to which one believes she or he can meet personal goals. Generally, these studies found that self-efficacy is also a significant predictor of adolescent tobacco use and alcohol consumption (e.g., Dijkstra, Sweeney, and Gebhardt 2001).

While previous research has attempted to determine whether single predictors--specifically advertising, social influences (particularly, parental and peer influence), and self-efficacy--affect adolescent tobacco use and alcohol consumption, the extent to which each predictor affects consumption behavior when controlling for the possible effects of other predictors remains unknown. This is an important gap in the literature because, realistically, these predictors do not operate in isolation in an individual's life. Indeed, a prominent theory of human behavior, Bandura's (1986) social cognitive theory, states that human behavior is a result of the reciprocal, bidirectional interrelationship of a person's environment and cognitive processes. This theory implies that research aimed at understanding any type of behavior should view possible predictors as interwoven, rather than independent, sources of influence. Thus, to determine if advertising, parental and peer influence, and self-efficacy constitute significant predictors of adolescent tobacco use and alcohol consumption, the current study seeks to determine the relative explanatory power of each factor by examining all factors simultaneously. Further, social cognitive theory does not place equal emphasis on all associations among the variables. By assessing the unique and additive effects of advertising, social influences, and self-efficacy, we provide evidence as to the relative strength of each variable as a predictor of adolescent tobacco and alcohol use. The discovery of the relative strength of each predictor of adolescent risk behavior will assist parents, consumer protection agencies, and public policy makers in devising strategies to prevent risk behaviors from occurring and to achieve a higher level of success in reversing the risk behaviors if they have begun.

In addition to examining simultaneously the relative effect of each social cognitive variable on adolescent risk behavior, this study extends prior research in a number of ways. First, instead of determining the relative importance of predictor variables by assessing the size of the rotated structural correlations, we attempt to validate past research findings by using a hierarchical regression usefulness analysis. This approach allows for control variables to be entered as independent variables separately to determine the influence of each social cognitive predictor's contribution to unique variance in adolescent risk behavior beyond that of another predictor's contribution. Second, while other studies aimed to link current advertising exposure and social influence to adolescents' decision to initiate the use of risk-related products, our study focuses on assessing how each variable contributes to the current maintenance of multiple risk behaviors through the use of general or thirty-day recall measures of perceived social and cognitive influence, as well as actual estimates of current tobacco and alcohol use. This method should provide more accurate estimates in determining which variables have the greatest impact on current risk behavior. Finally, research pertaining to adolescent tobacco use typically addresses the differences between smokers and nonsmokers. While there may be a difference between the beliefs of these two groups, evidence suggests that what reinforces nonsmokers' decision not to smoke does little to change smokers' actual behavior (Wolburg 2006). Thus, we strictly focused our research efforts on adolescents currently engaged in tobacco use.

BACKGROUND AND HYPOTHESES

Advertising Effects

Advertisements depicting consumer behaviors that are widely regarded as risky or unhealthy typically use favorable stereotypes to imply that those who engage in such behaviors are attractive, successful, and healthy (Pechmann and Knight 2002; Pechmann and Shih 1999). Mere exposure effects suggest that such stereotypes enhance the appeal of smoking to many adolescents and contribute positively to their initiation and use of cigarettes (Andrews and Franke 1991; Botvin et al. 1993). Adolescents' recognition of tobacco advertisements has been correlated with tobacco use, suggesting that tobacco advertising may contribute to the maintenance of tobacco habits (Fischer 1989). Furthermore, tobacco advertising may provide positive reinforcement of peer and parental tobacco use.

Likewise, broadcast media are also a major source of alcohol-related information for youth (Grube and Wallack 1994). Research examining the effects of advertising on alcohol-related beliefs, attitudes, and consumption intentions have found correlations between exposure to alcohol-related advertising and positive attitudes toward alcoholic beverage ads, positive attitudes toward drinking outcomes, and future drinking behaviors (Atkin, Hocking, and Block 1984; Stacy et al. 2004). Grube and Wallack (1994) report that children who are aware of beer advertisements hold more favorable beliefs about drinking, intend to drink more in adulthood, and are more knowledgeable of beer brands and slogans. In summary, findings support the notion that alcohol awareness attributed to advertising may result in favorable drinking beliefs, knowledge, and drinking intentions among adolescents.

Based upon past research and the underlying tenets of mere exposure effects, advertising should significantly predict adolescent tobacco use and alcohol consumption after controlling for other sources of influence. Thus,

H1: After controlling for other sources of influence, exposure to tobacco- and alcohol-related advertisements with positive stereotypes is a significant predictor of adolescent tobacco use and alcohol consumption.

Parental Effects

In addition to the modeling and reinforcement effects of advertising, direct exposure to parental figures who engage in risky behavior further reduces an individual's inhibition to engage in such behavior (Bandura 1986). For example, normative standards of behavior set by parents for their children are based upon the parents' general attitude toward the behavior in question. Thus, an encouraging attitude toward smoking or drinking from the parent provides positive reinforcement for the child to engage in and/or continue patterns of similar risky behavior.

Parental modeling and reinforcement effects related to tobacco use and alcohol consumption support this assertion (Fawzy, Coombs, and Gerber 1983). Children are more likely to drink alcohol if one parent in the family drinks alcohol (Wilks, Callan, and Austin 1989) and, while parental smoking is less consistent in predicting smoking onset, an association between a parent's smoking behavior and his/her child's smoking behavior is shown to exist (Smith and Stutts 1999). Additionally, children with involved parents are less likely to engage in tobacco, alcohol, and other substance use during adolescence (Fletcher and Jefferies 1999).

Given these findings, parental influence should significantly predict adolescent tobacco use and alcohol consumption. In addition, there are several reasons to suggest parental influence is a stronger predictor of risk behavior than advertising. For instance, parental control over adolescent exposure to advertised media should preempt children's susceptibility to advertising's influence. Additionally, parental behavior provides a direct, rather than indirect, channel of influence. Thus, positive stereotypes and attitudes in advertisements related to risk behavior may be reinforced by parental figures displaying similar behavior. Thus,

H2: After controlling for other sources of influence, parental influence regarding tobacco use and alcohol consumption is a more significant predictor of adolescent tobacco use and alcohol consumption than advertising.

Peer Effects

Peers, as well as parents, have been shown to be influential modeling agents associated with adolescent risk behavior (Bachmann, John, and Rao 1993). For instance, peers can provide adolescents with cigarettes, alcohol, and other drugs; teach substance usage; and provide social reinforcement for early attempts at smoking, drinking, and/or drug use (Brown, Clasen, and Eicher 1986). Moreover, the smoking behavior of peers, as compared to that of parents, has a stronger effect on adolescent smoking (Flay et al. 1994).

Peers also indirectly influence adolescent risk behavior through perceived social norms. Because adolescence is a time of especially high motivation to conform to social norms, peers can create the perception that risk behavior increases social acceptance (Simons-Morton, et al. 2001). Indeed, increased levels of smoking are linked to friends' positive attitudes toward smoking and can reinforce the notion that smoking is an enjoyable activity that promotes popularity (Flay et al. 1994; McAlister, Krosnick and Milburn 1984).

Based on the empirical evidence related to tobacco use and alcohol consumption, peer influence should significantly predict adolescent tobacco use and alcohol consumption. Further, peer influence may be a stronger predictor of risk behavior than advertising and/or parental influence. Adolescence is associated with peer integration and heightened awareness of peer norms. Thus, adolescents are more likely to selectively adopt attitudes and behaviors of peer groups that overshadow messages in advertising media. Therefore,

H3: After controlling for other sources of influence, peer influence regarding tobacco use and alcohol consumption is a more significant predictor of adolescent tobacco use and alcohol consumption than advertising and/or parental influence.

Self-Efficacy Effects

Theories of social influence are particularly valuable when describing behavior under a person's control. However, in situations where individuals have a perceived or real lack of control over behavior, internal cognitive processes may preempt external influences (Godin and Kok 1996). Self-efficacy is a cognitive variable that refers to individuals' beliefs regarding control over events in their lives (Bandura 1986). In a risk-behavior context, low self-efficacy individuals are unlikely to resist engaging in behavior detrimental to their health, whereas high self-efficacy individuals are more likely to resist such behavior. Self-efficacy has been linked to consumer risk behavior, specifically to the use of tobacco products and problem drinking behaviors (Dijkstra, Sweeney, and Gebhardt 2001). For instance, there appears to be a consistent relationship between low self-efficacy expectations and high levels of alcohol consumption in adult, student, and adolescent populations (Aas et al. 1995; Skutle 1999). Furthermore, adolescents with low levels of self-efficacy and low levels of emotional adjustment are more likely to continue smoking cigarettes (Engels et al. 2005).

Based on these empirical findings, self-efficacy should significantly predict adolescent tobacco use and alcohol consumption. Moreover, beliefs about one's ability to resist social influence should impact the degree to which external environmental factors affect behavior. Individuals with a strong sense of self-efficacy should show weaker levels of influence from advertising and social forces. Thus, self-efficacy should prove to be a significantly stronger predictor of adolescent tobacco use and alcohol consumption than advertising, parental influence, and/or peer influence.

H4: After controlling for other sources of influence, general self-efficacy is a more significant predictor of adolescent tobacco use and alcohol consumption than advertising, parental influence, and/or peer influence.

STUDY 1

In this study, we investigated the relative importance of advertising, parental influence, peer influence, and self-efficacy beliefs as predictors of adolescent unhealthy consumption behavior. Specifically, the respondents of this initial study were adolescents exhibiting a single risk behavior: cigarette smoking.

Method

Procedure

After IRB approval for questioning adolescents was obtained, survey administration took place during evening hours inside and outside of restaurants and cafes across three cities--Chicago, Illinois; San Antonio, Texas; and Washington, DC. A systematic sampling technique was primarily used; however, toward the end of data collection, a quota technique was used to minimize selection bias.

Once intercepted, respondents were screened for engagement in cigarette usage and the extent to which they felt competent in recalling and reporting the manner in which certain factors had impacted their smoking behavior. Once an individual was screened, the purpose and nature of the study was explained and $2 was offered in exchange for participation. In an effort to encourage respondents to report accurate information, counter-biasing information was provided during survey administration. For example, respondents were told "It's very common to smoke cigarettes," "In fact, a considerable number of people smoke cigarettes," etc. Counter-biasing methods aim to decrease the ego-defensive tendency to respond in a socially normative manner (Sudman and Bradburn 1974) and is shown to lower social desirability bias in health-related research (Raghubir and Menon 1996). To determine how each variable contributes to the current maintenance of multiple risk behaviors, we assessed recall estimates of perceived social and cognitive influence, as well as actual estimates of current tobacco and alcohol use. Because this study focuses on actual behavior rather than artificial experimental manipulation, the accuracy of recall estimates is critical to the validity of the results. Research guidelines suggest employing a short time frame to ensure accurate responses (Sudman and Schwarz 1989). Thus, this study used a reference time period of thirty days for advertising exposure and substance use. However, with regard to the parental and peer influence and self-efficacy factors, respondents were asked about the past in general. Each respondent who agreed to participate in the study was asked for contact information in case additional information was needed in the future. Of those who gave the contact information, a random sample was chosen for test-retest reliability checks.

Sample

A total of 101 qualified adolescents agreed to participate in the study. The mean age of the sample was 15.9 years and, on average, each respondent smokes over forty-eight cigarettes a month. Chi-square analysis revealed an over-representation of males in the sample. However, all other demographic parameters of the sample represented that of the population from which the sample was selected.

Independent Measures

Advertising influence ([alpha] = .87) was assessed using recalled exposure to advertised messages, as well as cognitive and affective processes elicited by their presence that may encourage individuals to engage in risk behavior. The current study adapted measures constructed by Schooler, Feighery and Flora (1996) to estimate advertising exposure through both broadcast and print media. To assess exposure, we asked respondents to provide the approximate value for the following questions: "In the past month, about how many cigarette ads do you recall seeing in print (i.e., magazines, newspapers, billboards, retail displays) media?" and "In the past month, about how many cigarette ads do you recall seeing in broadcast (i.e., radio, television, Internet) media?" To measure respondent affect toward ad stereotypes, we asked, "What is your general attitude toward those depicted or shown in cigarette ads?" (1 = Unfavorable; 5 = Favorable).

All items used to examine parental/peer influence ([alpha]par = .80; [[alpha].sub.Peer] = .85) were adapted from measures used by Grube and Wallack (1994). First, we assessed perceived parental and peer behavior by asking, "How often do your parents (peers) smoke cigarettes?" (1 = Never; 5 = Almost every day). Then we assessed attitude and subjective norms related to the behavior by asking the following questions: "What is your parents' (peers') general attitude toward smoking cigarettes?" (1 = Unfavorable; 5 = Favorable) and "How often do you feel you need to comply with your parents' (peers') attitude toward smoking cigarettes?" (1 = Never; 5 = All the time).

Self-efficacy ([alpha] = .94) was measured as a general construct that reflects an individual's ability to cope with a broad range of stressful or challenging demands. Four items were used from the general self-efficacy scale developed by Sherer et al. (1982). A typical question from the scale is "I never give up on things before completing them" (1 = Strongly disagree; 5 = Strongly agree).

Dependent Measures

Questions used to assess consumer risk behaviors are similar to those commonly used in substance abuse surveys (Simons-Morton et al. 2001). Specifically, actual behavior was assessed by asking "How many cigarettes have you smoked in the past thirty days?"

Demographic Measures

This study controlled for possible extraneous influences of socioeconomic status, gender, and ethnicity. We measured respondent gender and ethnicity using traditional nominal scales. As is common in youth research, adolescent SES was determined by parental occupation and educational attainment.

To check for scale reliability, test-retest average correlation coefficients were computed. Based on an overall score for each scale, the test-retest correlation was .91, n = 31, sixteen-day interval.

Analysis

Hierarchical regression analysis allows partitioning of the total variance explained by a set of predictor variables into unique proportions explained by discrete subsets of the antecedents. One approach, called usefulness analysis, is to vary the order of entry of different predictor variables so that a predictor subset's "contribution to unique variance in a criterion beyond another predictor's contribution may be explained" (Organ and Konovsky 1989, p. 161). We conducted hierarchical regression analysis in five major steps. First, demographic control variables were regressed on the risk behavior. Second, each predictor was regressed separately on the risk behavior, with the inclusion of significant demographic control variables from the first step. Third, each of the four predictor variables was added separately to the equations of the second step, and the change in [R.sup.2] was investigated for significance (p [less than or equal to] .05).

Fourth, each of the four predictor variables was added separately to subsets of every possible pair of predictor variables from the third step, and the change in [R.sup.2] was investigated for significance. Finally, each of the four predictor variables was added separately to subsets of every possible three-way combination of predictor variables from the fourth step, and the change in [R.sup.2] was again investigated for significance. An alpha value of .05 was used for determining statistical significance for all results.

Results

Table 1 shows the results of the five steps of the hierarchical regression analysis and the standardized beta estimates for the full regression model. Before addressing the ultimate testing of the hypotheses (Step 5), a brief overview of the results of the preliminary steps of the model is appropriate. Significant demographic control variables based on t-values greater than or equal to [+ or -] 1.645 (p [less than or equal to] .05) were retained during the first hierarchical regression step. Based upon this assessment, socioeconomic status was negatively related to smoking behavior. In addition, smoking behavior was associated more with males than females and was higher among Caucasians than African Americans or Hispanics.

Step 2 of the hierarchical regression analysis determined if each social cognitive predictor variable explained the unique variance in smoking behavior when controlling for the demographic variables. Based upon the significant change in [R.sup.2] associated with each social cognitive variable in Step 2, advertising, parental influence, peer influence, and self-efficacy were all shown to significantly predict smoking behavior among adolescents. Step 3 determined the predictive ability of each social cognitive variable after controlling for the effects of one other social cognitive variable. The findings show that advertising is a significant predictor of smoking beyond self-efficacy, but not beyond parental and peer influence. Indeed, both parental and peer influence remain significant beyond each of the three other variables. The opposite was true for self-efficacy: after controlling for one of the three other variables (advertising, parental influence, and peer influence), self-efficacy did not significantly predict smoking behavior. These results are largely mirrored in Step 4. Finally, Step 5 determined the predictive ability of each social cognitive variable after controlling for the three other cognitive predictor and demographic variables. When controlling for all other subsets of predictors in Step 5, the resulting change in [R.sup.2] values and standardized beta coefficient revealed that advertising is nonsignificant for smoking behavior, parental influence and peer influence are significant for smoking behavior, and self-efficacy is nonsignificant for smoking behavior.

The results show that advertising is not a significant predictor of adolescents' smoking behavior when accounting for parental influence, peer influence, self-efficacy, and demographic variables; thus, H1 is rejected. In contrast, the results support H2 and H3 in that peer influence and parental influence are significant, positive predictors of smoking behavior among adolescents after removing the possible effects of advertising, self-efficacy, and demographic controls. On the other hand, self-efficacy does not significantly predict adolescent smoking behavior after controlling for other social cognitive predictors and demographic variables. Thus, no support was found for H4. The beta coefficients in Table 1 reveal that peer influence is the strongest predictor of smoking behavior, followed by parental influence. Advertising and self-efficacy do not significantly predict smoking behavior.

STUDY 2

Study 1 examined the relative predictive ability of each social cognitive variable as it relates to adolescent smoking behavior. In an effort to enhance the external validity of the findings, Study 2 focuses on adolescents exhibiting multiple risk behaviors: the utilization of tobacco in any form and the consumption of alcohol.

Method

The method employed in the first study was largely followed in Study 2. Only the methodological elements that are unique to Study 2 are mentioned here.

Sample

A total of eighty-nine qualified adolescents agreed to complete the survey. The respondent mean age was 16.3 years, and, on average, each respondent uses over thirty-five tobacco products and consumes over sixty-three alcoholic beverages a month. Chi-square analysis revealed upper-income households were under-represented in the sample. However, all other demographic parameters of the sample represented those of the population from which the sample was drawn.

After individuals were systematically intercepted, they were screened for tobacco use and alcohol consumption and the extent to which they felt competent in recalling and reporting the manner in which certain factors had impacted their tobacco use and alcohol consumption. Again, counter-biasing information was provided to the subjects. For example, respondents were told "It's very common to use tobacco and drink" and "In fact, a considerable number of people use tobacco and drink alcohol."

Independent and Dependent Measures

Subjects were administered the set of independent and dependent items as those present in Study 1 with the exception of the term "cigarettes" replaced with "tobacco, beer, wine, and/or alcohol". Again, test-retest correlations indicated scale reliability: r=92, n=26, seventeen-day interval.

Results

The results of the hierarchical regression analysis and the standardized beta estimates for the full regression model are shown in Table 2. Step 1 of the hierarchical regression analysis revealed significant relationships between demographic control variables and both tobacco use and alcohol consumption. Specifically, detailed analyses of the findings reveal that socioeconomic status is negatively associated with tobacco use and alcohol consumption, and males are more risk prone than females.

Step 2 of the hierarchical regression analysis determined if each social cognitive predictor variable explains the unique variance in tobacco use and alcohol consumption behavior when controlling for the demographic variables. Based upon the significant change in [R.sup.2] associated with each social cognitive variable in Step 2, advertising, parental influence, peer influence, and self-efficacy were all shown to significantly predict alcohol consumption. Although advertising, parental influence, and peer influence surfaced as significant predictors of tobacco use, the predictive ability of self-efficacy failed to reach significance.

Step 3 determined the predictive ability of each social cognitive variable after controlling for the effects of one other social cognitive variable. The findings show that advertising is only a significant predictor of tobacco use and alcohol consumption when it is coupled with parental influence and self-efficacy, respectively. Otherwise, its predictive ability is nonsignificant. Both parental and peer influence remain significant beyond each of the three other variables for tobacco use and alcohol consumption. Identical to the findings for smoking behavior, self-efficacy does not significantly predict tobacco use or alcohol consumption after controlling for one of the three other variables (advertising, parental influence, and peer influence). The results of Step 4 show that after two other social cognitive variables are controlled, only parental and peer influence surface as significant predictors of tobacco use and alcohol consumption. Finally, Step 5 determined the predictive ability of each social cognitive variable after controlling for the three other cognitive predictor and demographic variables. Again, these findings lead to the rejection of Hl and H4 and to the support of H2 and H3.

The beta coefficients in Table 2 are similar to those found in Study 1. Peer influence is the strongest predictor of both tobacco use and alcohol consumption behavior, followed by parental influence. Advertising and self-efficacy do not significantly predict adolescent tobacco use and alcohol consumption behavior.

DISCUSSION

Research Implications

Decreasing the prevalence of adolescent smoking and drinking would lower personal and societal costs that accrue from increased health care expenditures, absenteeism from school and work due to poor health, and family leave time to care for loved ones. While several factors have been shown to contribute to the maintenance of tobacco and alcohol use, advertising has been singled out as a potentially powerful agent of influence on decisions by adolescents to use such products. Yet, our results challenge the efficacy of advertising media as a significant predictor of adolescent tobacco use and alcohol consumption when factoring in other sources of influence. In fact, public policy directed at eliminating all tobacco and/or alcoholic advertising may not prove to be effective in reducing the smoking prevalence among adolescents and, consequently, may not serve its intended purpose.

Instead, the aforementioned results provide support for directions for risk-behavior prevention and cessation programs that rely on interventions that account for an adolescent's entire external environment. This includes factoring in risks related to the physical environment, associations with peer and family members who reinforce risk behavior, and the availability of role models who engage in risk behavior (Capella 2005).

For instance, cafes and restaurants are difficult settings for smokers and drinkers who are trying to quit. Smokers are routinely surrounded by other smokers in such settings, making it even more difficult to maintain cessation behavior. In fact, there is such a close association with drinking and smoking that just entering a setting where alcohol is consumed can trigger a smoker to desire a cigarette (Wolburg 2008). So while more than half of the nation's restaurants remain smoke-free by law, those owners who continue to oppose bans for fear of lost revenue only help to encourage the maintenance of tobacco and smoking habits (Wolburg 2008). Our findings from this study underscore the importance of tobacco-free environments that are open to adolescents. Moreover, the strong effects of peer influence on tobacco and alcohol use suggest strict enforcement of underage smoking and drinking laws.

Our findings also suggest that mechanisms intended to increase adolescent self-efficacy refusal skills for alcohol and tobacco products, such as properly designed antismoking public service announcements and educational programs, are ineffective. Wolburg (2009) recently noted that while public service announcements point out the severity of the consequences of risk behavior and personal vulnerability to those consequences, such announcements fail to deliver a message that builds confidence in adolescent self-efficacy beliefs. Developers of public service announcements are advised to provide help lines, web sites, cessation aides, and local support group information within the message to increase self-efficacy beliefs (Wolburg 2006).

CONCLUSIONS

Unlike previous research efforts that focused primarily on a single factor of influence, the overall purpose of this research was to determine if four social cognitive factors provide enhanced insight into the development of risk behavior above and beyond that offered by a single source of influence. Specifically, we examined the relative explanatory power of advertising, parental influence, peer influence, and self-efficacy as predictors of adolescent tobacco use and alcohol consumption.

Because government agencies and consumer groups are aggressively pursuing consumer protection for children and adolescents, a fundamental goal herein was to untangle the equivocal findings related to advertising's influence on adolescent risk behavior. Critics of advertising media argue unhealthy product advertisements are responsible for the prevalence of tobacco use and alcohol use in today's adolescent population. Yet, these two studies find limited support for this contention, indicating that advertising explains only about 1% of unique variance across both adolescent risk behaviors.

When controlling for other social cognitive predictors, peer influence is found to be the strongest predictor of adolescent tobacco use and alcohol consumption. Parental influence is shown to be the second most influential variable for both adolescent risk behaviors. Specifically, parental and peer influence combined to account for 26% of unique variance in adolescent smoking behavior, 31% of unique variance in adolescent tobacco use, and 35% of unique variance in adolescent alcohol consumption. These results suggest that peer and parental influence are particularly strong during adolescence. Interestingly, the predictive ability of self-efficacy fails to reach significance once other factors (advertising, parental influence, and peer influence) are considered and controlled. Self-efficacy and advertising, in and of themselves, are indeed important correlates of tobacco and alcohol consumption. However, the predictive power of these two factors sharply declines when factoring in the relative explanatory power of peer and parental influence.

Perhaps the lack of statistical significance of self-efficacy once other factors were controlled stems from the challenge of removing positive associations with particular adolescent risk behaviors. Because smoking cigarettes and drinking alcohol are likely perceived by adolescents as a means to gain acceptance from peer groups, initiatives aimed to improve self-efficacy refusal skills that discourage such behavior are subject to denial, disregard, and boomerang effects. While this may suggest self-efficacy is a weak predictor of risk behaviors viewed positively by peer groups, risk behaviors that are viewed less positively by peer groups (i.e., overeating unhealthy foods, drinking and driving, drug use, etc.) may be more susceptible to self-efficacy influence. Considering our findings for two specific risk behaviors regarded favorably by adolescents, future research on self-efficacy should explore less popular risk behaviors such as drug use, unsafe sexual practices, overeating, and refusal to use seat belts. Another possible explanation for the nonsignificant findings is the use of a general measure of self-efficacy, rather than a context-specific measure (i.e., for smoking or alcohol consumption) as was used in some previous studies (e.g., Dijkstra, De Vries, and Roijackers 1998). This possibility introduces the need for future research to compare the validity of general and context-specific measures of self-efficacy.

This contributes to the existing literature in two key ways. First, the total unique variance explained by advertising, parental influence, peer influence, and self-efficacy beliefs for smoking, tobacco use, and alcohol consumption was 26%, 32%, and 37%, respectively. Past research indicates that attitudes, subjective norms, and perceived behavioral control yield, on average, 34% of explained variance in many health-related behaviors (Godin and Kok 1996). Thus, these results are comparable to the standards of established models in applied psychology and represent a clear improvement over research models that recognize a single factor of influence. These new findings also suggest that social cognitive variables are interwoven, rather than independent, sources of influence. Such findings support Bandura's (1986) social cognitive framework and Bauer's (1968) "limited effects" model, suggesting environmental media is a secondary source of influence and such influence is largely neutralized by interactive interpersonal communications with parents and peers. Secondly, the findings reinforce past research that shows risk behavior differs by gender, social class, and ethnicity, as evidenced by the greater propensity of males, individuals from lower socioeconomic groups, and Caucasians to engage in tobacco use and alcohol consumption.

This provides valuable insights into the nature of the social cognitive influences on adolescent risk behavior, with implications for consumers, health care providers, and public policy officials, all of whom are affected by the consequences of adolescents who engage in risky, unhealthy consumption behaviors.

Additional research is still needed in the area of adolescent risk behavior in order to better understand the developmental process and the impact of influences at different life stages. For instance, one might expect parental influence to wane when adolescents transition into adulthood and out of the home and away from parental control. Peer influence could shift if adolescents surround themselves with different peer groups after graduating high school. Self-efficacy may be more influential if individuals attain more sophisticated cognitive defense mechanisms to manage an environmental stimulus that promotes risk behavior. In addition, certain risk behaviors may become more or less socially acceptable over time. For instance, while smoking cigarettes was once associated with intelligence and sophistication, the current perception of smoking in the United States is generally negative and has forced changes to how tobacco products can be advertised and packaged, how tobacco products are sold, and where tobacco products can be used. Over time, such changes to the environmental landscape can significantly impact how advertising and social variables influence adolescent risk behavior.

A modified replication of this research would provide additional insight if conducted with a more direct method of measuring the social cognitive variables. Although we used the self-report measures with great care to enhance scale validity to the extent possible, we would be remiss if we did not note that previous studies have found that perceptions of others' risk behavior tend to reflect more the respondent's own behavior than the actual behavior of peers or parents (Sussman et al. 1988).

REFERENCES

Aas, H., K.I. Klepp, J.C. Laberg, and A.L. Aaro. 1995. Predicting Adolescents Intentions to Drink Alcohol: Alcohol Expectations and Self-Efficacy. Journal of Studies on Alcohol, 56 (May): 293-299.

Andrews, Rick L. and George R. Franke. 1991. The Determinants of Cigarette Consumption: A Meta-Analysis. Journal of Public Policy and Marketing, 10 (Spring): 81-100.

Andrews, J. Craig, Richard G. Netemeyer, Scot Burton, D. Paul Moberg, and Ann Christiansen. 2004. Understanding Adolescent Intentions to Smoke: An Examination of Relationships among Social Influence, Prior Trial Behavior, and Anti-tobacco Campaign Advertising. Journal of Marketing, 68 (July): 110-123.

Atkin, Charles, J. Hocking, and Martin Block. 1984. Teenage Drinking: Does Alcohol Advertising Make a Difference? Journal of Communication, 34 (Spring): 157-167.

Bachmann, Gwen R., Deborah Roedder John, and Akshay R. Rao. 1993. Children's Susceptibility to Peer Group Purchase Influence: An Exploratory Investigation. In Advances in Consumer Research, edited by Leigh McAlister and Michael L. Rothschild, vol. 20 (463-468). Provo, UT: Association for Consumer Research.

Bandura, Albert. 1986. Social Foundations of Thought and Action: A Social Cognitive Theory. Englewood Cliffs, NJ: Prentice Hall.

Bauer, Raymond and S. Greyer. 1968. Advertising in America: The Consumer View. Boston: Harvard University Press.

Botvin, Gilbert J., Catherine J. Goldberg, Elizabeth M. Botvin, and Linda Dusenbury. 1993. Smoking Behavior of Adolescents Exposed to Cigarette Advertising. Public Health Reports, 108 (March-April): 217-224.

Brown, Bradford, Donna Clasen, and Sue Eicher. 1986. Perceptions of Peer Pressure, Peer Conformity Dispositions, and Self-reported Behavior among Adolescents. Developmental Psychology, 22 (July): 521-530.

Capella, Michael L. 2005. A Review of the Effect of Advertising on Cigarette Initiation, Continuation, and Brand Behavior: A Mixed Method Approach. (Doctoral Dissertation, Mississippi State University).

Dijkstra, Arie, Hein De Vries, and Jolanda Roijackers. 1998. Long-Term Effectiveness of Computer-Generated Tailored Feedback in Smoking Cessation. Health Education Research, 13 (June): 207 -214.

Dijkstra, Arie, Laura Sweeney, and Winnie Gebhardt. 2001. Social Cognitive Determinants of Drinking in Young Adults: Beyond the Alcohol Expectancies Paradigm. Addictive Behaviors, 26 (September-October): 689-706.

Ellickson, Phyllis, Rebecca Collins, Katrin Hambarsoomians, and Daniel McCaffrey. 2005. Does Alcohol Advertising Promote Adolescent Drinking? Results from a Longitudinal Assessment. Addiction, 100 (February): 245-246.

Engels, Rutger, William Hale, Marc Noom, and Hein DeVries. 2005. Self-Efficacy and Emotional Adjustment as Precursors of Smoking in Early Adolescence. Substance Use and Misuse, 40 (December): 1883-1893.

Fawzy, FI, Robert Coombs, and Barry Gerber. 1983. Generational Continuity in the Use of Substances: The Impact of Parental Substance Use on Adolescent Substance Use. Addictive Behaviors, 8 (February): 109-114.

Fischer, P. 1989. Recall and Eye Tracking Study of Adolescents Viewing Tobacco Advertising. Journal of the American Medical Association, 261 (January): 84-94.

Flay, Brian, Frank B. Hu, Ohidul Siddiqui, L.E. Day, Donald Hedeker, John Petraitis, Jean Richardson, and Steve Sussman. 1994. Differential Influence of Parental Smoking and Friends' Smoking on Adolescent Initiation and Escalation of Smoking. Journal of Health and Social Behavior, 35 (September): 248-265.

Fletcher, Anne and Brandi Jefferies. 1999. Parental Mediators of Associations between Perceived Authoritative Parenting and Early Adolescent Substance Use. Journal of Early Adolescence, 19 (November): 465-87.

Godin, Gaston and Gerjo Kok. 1996. The Theory of Planned Behavior: A Review of Its Applications to Health-Related Behaviors. American Journal of Health Promotion, 11 (November-December): 87-98.

Grube, Joel W. and Lawrence Wallack. 1994. Television Beer Advertising and Drinking Knowledge, Beliefs, and Intentions among Schoolchildren. Journal of Public Health, 2 (February): 254-259

McAlister, Alfred, Jon Krosnick, and Michael Milburn. 1984. Causes of Adolescent Cigarette Smoking: Tests of a Structural Equation Model. Social Psychology Quarterly, 47 (March): 24-36.

Organ, Dennis and Mary Konovsky. 1989. Cognitive Versus Affective Determinants of Organizational Citizenship Behavior. Journal of Applied Psychology, 74 (February): 157-164.

Pechmann, Cornelia and Susan J. Knight. 2002. An Experimental Investigation of the Joint Effects of Advertising and Peers on Adolescents' Beliefs and Intentions about Cigarette Consumption. Journal of Consumer Research, 29 (June): 5-19.

Pechmann, Cornelia and Chuan-Fong Shih. 1999. Smoking Scenes in Movies and Antismoking Advertisements Before Movies: Effects of Youth. Journal of Marketing, 63 (July): 1-13.

Raghubir, Priya and Geeta Menon. 1996. Asking Sensitive Questions: The Effects of Type of Referent and Frequency Wording in Counterbiasing Methods. Psychology and Marketing, 13 (October): 633-653.

Schooler, C., E. Feighery, and J.A. Flora. 1996. Seventh Graders' Self-reported Exposure to Cigarette Marketing and its Relationship to Their Smoking Behavior. American Journal of Public Health, 86 (September): 1216-1221.

Sherer, Mark, James E. Maddux, Blaise Mercandante, Steven Prentice-Dunn, Beth Jacobs, and Ronald W. Rogers. 1982. The Self-Efficacy Scale: Construction and Validation. Psychological Reports, 51 (October): 663-671.

Simons-Morton, Bruce, Denise Haynie, Aria Crump, Patricia Eitel, and Keith Saylor. 2001. Peer and Parental Influences on Smoking and Drinking Behavior among Early Adolescents. Health Education and Behavior, 28 (February): 95-107.

Skutle, Arvid. 1999. The Relationship among Self-Efficacy Expectancies, Severity of Alcohol Abuse, and Psychological Benefits from Drinking. Addictive Behaviors, 24 (January-February): 87-98.

Smith, Karen and Mary Ann Stutts. 1999. Factors that Influence Adolescents to Smoke. Journal of Consumer Affairs, 33 (Winter): 231-268.

Stacy, Alan, Jennifer Zogg, Jennifer Unger, and Clyde Dent. 2004. Exposure to Televised Alcohol Ads and Subsequent Adolescent Alcohol Use. American Journal of Health Behavior, 28 (November-December): 498-509.

Sudman, Seymour and Norman Bradburn. 1974. Response Effects in Surveys: A Review and Synthesis. Chicago, IL: Aldine.

Sudman, Seymour and Norbert Schwarz. 1989. Contributions of Cognitive Psychology to Advertising Research. Journal of Advertising Research, 29 (June-July): 43-53.

Sussman, Steve, Clyde Dent, Jill Mestel-Rauch, and C. Anderson Johnson. 1988. Adolescent Nonsmokers, Triers, and Regular Smokers' Estimates of Cigarette Smoking Prevalence: When Do Overestimations Occur and by Whom? Journal of Applied Social Psychology, 18 (June): 537-551.

Wilks, Jeffrey, Victor Callan, and Derek Austin. 1989. Parent, Peer and Personal Determinants of Adolescent Drinking. British Journal of Addiction, 84 (June): 619-630.

Wolburg, Joyce M. 2006. College Students' Responses to Antismoking Messages: Denial Defiance, and Other Boomerang Effects. Journal of Consumer Affairs, 40 (Winter): 294-323.

Wolburg, Joyce M. 2008. Smoking Cessation: Why Do Smokers Fail? Journal of Consumer Marketing, 25 (2): 72-72.

Wolburg, Joyce M. 2009. Misguided Optimism Among College Student Smokers: Leveraging their Quit-Smoking Strategies for Smoking Cessation Campaign. The Journal of Consumer Affairs, 43 (Summer): 305-331.

Brian R. Kinard (kinardb@uncw.edu) is an Assistant Professor of Marketing at the University of North Carolina--Wilmington, and Cynthia Webster (cwebster@cobilan.msstate.edu) is Professor of Marketing at Mississippi State University.
TABLE 1
Adolescent Smoking Behavior

                                                     [DELTA][R.sup.2]
Regression Steps                                     Smoking Behavior

Step 1
Demographic control variables only                       .04 (a)

Step 2
Advertising                                              .08 (b)
Parental influence                                       .29 (b)
Peer influence                                           .38 (b)
Self-efficacy                                            .03 (a)

Step 3
Advertising beyond parental influence                    .02
Advertising beyond peer influence                        .01
Advertising beyond self-efficacy                         .08 (b)
Parental influence beyond advertising                    .27 (b)
Parental influence beyond peer influence                 .10 (b)
Parental influence beyond self-efficacy                  .27 (b)
Peer influence beyond advertising                        .33 (b)
Peer influence beyond parental influence                 .17 (b)
Peer influence beyond self-efficacy                      .34 (b)
Self-efficacy beyond advertising                         .02
Self-efficacy beyond parental influence                  .00
Self-efficacy beyond peer influence                      .00

Step 4
Advertising beyond parental influence and peer
  influence                                              .00
Advertising beyond parental influence and
  self-efficacy                                          .05 (a)
Advertising beyond peer influence and
  self-efficacy                                          .01
Parental influence beyond advertising and peer
  influence                                              .15 (b)
Parental influence beyond advertising and
  self-efficacy                                          .25 (b)
Parental influence beyond peer influence and
  self-efficacy                                          .19 (b)
Peer influence beyond advertising and parental
  influence                                              .16 (b)
Peer influence beyond advertising and
  self-efficacy                                          .31 (b)
Peer influence beyond parental influence and
  self-efficacy                                          .16 (b)
Self-efficacy beyond advertising and parental
  influence                                              .00
Self-efficacy beyond advertising and peer
  influence                                              .00
Self-efficacy beyond parental influence and peer
  influence                                              .00

Step 5
Advertising beyond parental influence, peer
  influence, and self-efficacy                           .00
Parental influence beyond advertising, peer
  influence, and self-efficacy                           .11 (b)
Peer influence beyond advertising, parental
  influence, and self-efficacy                           .15 (b)
Self-efficacy beyond advertising, parental
  influence, and peer influence                          .00
Advertising [beta]                                       .008
Parental influence [beta]                                .166 (b)
Peer influence [beta]                                    .266 (b)
Self-efficacy [beta]                                     .003

(a) p [less than or equal to] .05, (b) p [less than or equal to] 01.

TABLE 2
Adolescent Tobacco Use and Alcohol Consumption

                                              [DELTA][R.sup.2]

                                            Tobacco      Alcohol
Regression Steps                              Use      Consumption

Step 1
Demographic control variables only          .03 (a)       .05 (a)

Step 2
Advertising                                 .05 (a)       .06 (a)
Parental influence                          .23 (b)       .22 (b)
Peer influence                              .28 (b)       .26 (b)
Self-efficacy                               .01           .04 (a)

Step 3
Advertising beyond parental influence       .03 (a)       .02
Advertising beyond peer influence           .02           .02
Advertising beyond self-efficacy            .01           .03 (a)
Parental influence beyond advertising       .20 (b)       .19 (b)
Parental influence beyond peer influence    .09 (b)       .08 (a)
Parental influence beyond self-efficacy     .21 (b)       .19 (b)
Peer influence beyond advertising           .22 (b)       .20 (b)
Peer influence beyond parental influence    .15 (b)       .13 (b)
Peer influence beyond self-efficacy         .23 (b)       .11 (b)
Self-efficacy beyond advertising            .01           .04 (a)
Self-efficacy beyond parental influence     .00           .02
Self-efficacy beyond peer influence         .00           .02

Step 4
Advertising beyond parental influence
  and peer influence                        .02           .01
Advertising beyond parental influence
  and self-efficacy                         .02           .01
Advertising beyond peer influence and
  self-efficacy                             .01           .01
Parental influence beyond advertising
  and peer influence                        .17 (b)       .07 (a)
Parental influence beyond advertising
  and self-efficacy                         .07 (a)       .17 (b)
Parental influence beyond peer influence
  and self-efficacy                         .19 (b)       .05
Peer influence beyond advertising and
  parental influence                        .19 (b)       .18 (b)
Peer influence beyond advertising and
  self-efficacy                             .21 (b)       .09 (a)
Peer influence beyond parental influence
  and self-efficacy                         .20 (b)       .07 (a)
Self-efficacy beyond advertising and
  parental influence                        .00           .02
Self-efficacy beyond advertising and
  peer influence                            .00           .01
Self-efficacy beyond parental influence
  and peer influence                        .00           .01

Step 5
Advertising beyond parental influence,      .01           .01
  peer influence, and self-efficacy
Parental influence beyond advertising,      .13 (b)       .15 (b)
  peer influence, and self-efficacy
Peer influence beyond advertising,          .18 (b)       .20 (b)
  parental influence, and self-efficacy
Self-efficacy beyond advertising,            .00          .01
  parental influence, and peer influence
Advertising [beta]                           .012         .01
Parental Influence [beta]                    .206 (b)     .208 (b)
Peer influence [beta]                        .284 (b)     .273 (b)
Self-efficacy [beta]                         .005         .019

(a) p [less than or equal to] .05. (b) p [less than or equal to] .01.
COPYRIGHT 2010 American Council on Consumer Interests
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2010 Gale, Cengage Learning. All rights reserved.

 Reader Opinion

Title:

Comment:



 

Article Details
Printer friendly Cite/link Email Feedback
Author:Kinard, Brian R.; Webster, Cynthia
Publication:Journal of Consumer Affairs
Article Type:Essay
Geographic Code:1USA
Date:Mar 22, 2010
Words:7187
Previous Article:Are consumers disadvantaged or vulnerable? An examination of consumer complaints to the Better Business Bureau.
Next Article:Credit counseling to help debtors regain footing.
Topics:

Terms of use | Copyright © 2014 Farlex, Inc. | Feedback | For webmasters