The Mass Media and Adolescent Socialization

Abstract

Despite the importance of the mass media to adolescent socialization, mass media effects on descriptive norms have, to a large degree, been neglected in the literature. Based on social comparison theory and with data from a national survey of U.S. adolescents (N = 1,436), this study examines how advertising exposure and attention predict descriptive norms specific to unhealthy food consumption. Advertising attention, as compared to exposure, had stronger positive associations with descriptive norms specific to three reference groups (i.e., family, close friends, and students at school). Advertising effects were stronger in two cases for older adolescents than for younger adolescents.

Keywords

advertising effects food advertising descriptive norms adolescents socialization

The mass media are a fundamental socializing agent during adolescence. Children aged eight to eighteen are inundated with media, experiencing (1) saturation or nearing saturation levels in access to television content, music and audio, computers, video games, print media, and movies and (2) increases in daily total media use from six hours nineteen minutes in 1999 to seven hours thirty-eight minutes in 2009.¹ Pertinent to socialization processes are social norms. Communication both (1) influences the development of people’s perceptions of social norms and (2) serves as a channel through which social norms can influence people’s behaviors.² A large body of research, including that on the focus theory of normative conduct, has investigated this latter linkage—from social norms to behavior—demonstrating the effects of descriptive norms and injunctive norms on health behaviors such as adolescents’ consumption of illicit substances.³ In contrast, the former linkage—from communication to social norms—has received much less attention in the literature. While social comparison theory has been used in the literature on media effects on body image,⁴ this research does not commonly extend specifically to descriptive norms, which are central to the present study. Moreover, only a few empirical studies focus on how health media campaigns influence descriptive norms or injunctive norms⁵—and none on the related normative influence of product advertising or, in particular, food advertising. Finally, there appear to be no studies that empirically assessed how the mass media influence on social norms varies by reference group and across adolescence.

The paucity of research in this area is surprising, given (1) that product advertising plays a critical role in assorted health contexts and (2) that empirical research on the latter linkage—from social norms to behavior—is often used to justify media campaigns that aim to influence social norms and, in turn, behavior.⁶ To address these gaps, the present study builds theory on and tests the influence of adolescents’ unhealthy food advertising exposure and attention on their descriptive norms about unhealthy food consumption, as well as how such influence varies by reference group and across adolescence.

Social Comparison Theory

People constantly rely on social comparisons to evaluate their own beliefs and behaviors relative to social reality and make appraisals related to self-evaluation, self-improvement, and self-enhancement.⁷ Such socialization processes are especially critical to adolescents, who frequently make comparisons between themselves and others⁸ to build an emerging self-concept.⁹ Comparisons can be downward or upward. People most commonly partake in downward comparison with less fortunate others because it leads to the development of increased self-esteem.¹⁰ In contrast, people who are motivated by self-improvement may partake in upward comparisons to more fortunate others, which can also lead to a decrease in self-esteem.¹¹ For instance, when women view thin models in the media, they tend to make upward comparisons, resulting in increased body dissatisfaction.¹²

Central to socialization are social norms, which afford people with perceived guidelines and standards for how they should behave.¹³ Research has commonly focused on descriptive and injunctive norms. Descriptive norms denote what behaviors are perceived to be typical or normal,¹⁴ while injunctive norms involve rules or beliefs about what constitutes morally approved and disproved behavior. Injunctive and descriptive norms are often consistent with one another.¹⁵ We focus on descriptive norms in the present study, positing that media effects may first influence descriptive norms, which have been conceived to precede injunctive norms in influencing behavior.¹⁶ The vital importance of normative development has been well documented in the literature. While there is a large body of research on how social norms influence adolescents’ use of illicit substances, such as cigarettes, drugs, and alcohol,¹⁷ only a few studies have tested normative influence on food consumption. Most of these studies do not distinguish between reference groups (e.g., peers vs. family), nor offer a consistent picture of how social norms influence food consumption.¹⁸ These inconclusive findings may result from the implementation of varied outcome measures, samples, and normative conceptualizations.

The Role of the Mass Media

Normative influence is not only limited to exposure to one’s proximate world (e.g., friends and family) but also entails one’s exposure to the distal world (e.g., other acquaintances, strangers, the media). In constructing a sense of social reality and making judgments of the prevalence of events and behaviors and comparing them with reality, people often rely on communication and the media in particular to assess discrepancies between their own behaviors and the behaviors of other people.¹⁹ Across adolescence, the influence of family on adolescent socialization diminishes, while the influence of other factors, such as peers and the mass media, increases.²⁰ Socialization is the process through which children learn values, normative beliefs, knowledge, and behaviors.²¹ According to primary socialization theory, adolescents learn social behaviors primarily through family, school, and peers, and secondarily through the mass media, religion, community, and culture. The primary socialization process influences media effects through selection, selective perception, and exposure norms.²² Because humans have limited cognitive capacity, they cannot attend to or process all present information and, instead, partake in selective exposure, as well as selective attention, to attain and sustain perceived control of specific stimulus events.²³ Selective perception comes into play in terms of the manner in which primary socialization factors influence how media messages are interpreted and processed.²⁴ Finally, in terms of exposure norming, interactions with sources of primary socialization—such as family, school, and peers—can influence an adolescent’s interpretation of previously seen media messages.

The Case of Unhealthy Food Advertising

Theorizing on and empirical testing of mass media influence on normative perceptions is especially prevalent in the area of body image, where research has found that men and women alike feel pressured to conform to the norms set forth by same-sex media characters.²⁵ In the context of lifestyle behaviors, while there is acknowledgment of the role of the mass media, empirical assessments are few. Such assessments have demonstrated that news coverage of binge drinking is associated with perceived peer disapproval of binge drinking.²⁶ In addition, a media campaign led to a reduction in college students’ descriptive norms about alcohol consumption,²⁷ and an experimental study documented the influence of alcohol messaging on college students’ normative perceptions.²⁸

In lieu of such empirical research on product advertising in general or food advertising in particular, we turn to research on food media portrayals. Adolescents are exposed to food advertising in different forms, including in-school marketing, product placements, and Internet content, but television is its largest source.²⁹ Advertising is rife with depictions of unhealthy foods. Children are inundated with television commercials, including those for unhealthy food products. For instance, 51% of food ads in television programming from 2 p.m. to 10 p.m. targeted children, with most of these ads for pizza/fast food, breakfast foods, and sweets.³⁰ In prime-time television hours, food advertisements made up 18% of the advertising total.³¹ In order, the top five such advertisers were McDonalds, Kentucky Fried Chicken, Wendy’s, Pizza Hut, and Burger King. In each thirty minutes of programming, an average of 4.7 food advertisements ran. Moreover, advertising targeted toward school-age children often associates food and beverages with good times (75%), as well as with being hip or cool (43.2%) and feeling happy (43.2%).³² More broadly, use of television content has increased among children aged 8-18, from three hours and forty-seven minutes in 1999 to four hours and twenty-nine minutes in 2009.³³ As compared with 8-10-year-olds (3:41), daily use of TV content is higher for 11-14-year-olds (5:03) and 15-18-year-olds (4:22).

This inundation of unhealthy food advertising is expected to spur adolescents’ perceptions of how often people consume unhealthy foods (i.e., descriptive norms), whether with families, close friends, or students at school. Advertising use is measured in terms of exposure and attention. These concepts are common in media effects research³⁴ and share common ground with awareness and attention, which are critical to advertising hierarchical models, including the AIDA (Attention-Interest-Desire-Action) model and information processing model.³⁵ In the present study, advertising exposure is a behavior, one encompassing an adolescent’s viewing of unhealthy food advertising. In contrast, advertising attention is cognitive, representing an increased mental effort given to unhealthy food advertising.

H1: Adolescents’ unhealthy food advertising exposure and attention are positively associated with their descriptive norms about unhealthy food consumption.

Variance by Reference Group

Normative influence operates in the context of different reference groups, determining the perceived prevalence and morality of behaviors and, in turn, the likelihood that group members will adopt certain behaviors.³⁶ For adolescents, reference groups vary from the proximate or close to the distal or distant. Proximate group norms tend to have more influence on adolescent behaviors than distal group norms.³⁷ For example, research on alcohol consumption among college students has indicated the greater normative influence of proximate reference groups than distal reference groups.³⁸

It can also be hypothesized that advertising effects on social norms are greater for distal groups than proximate groups because adolescents have different levels of exposure to and experience with these reference groups. An adolescent knows others well in close or proximate groups and interacts commonly with them, thus gaining an entrenched and accurate view that is largely based in direct observation and direct experience. While adolescents would have accurate perceptions of the normative behaviors of such proximate groups, they would know much less about the normative behaviors of distal groups, who are not close and, thus, permit much less direct observation and direct experience.³⁹ In terms of descriptive norms, we expect that advertising will play a stronger role in predicting adolescents’ normative estimates of distal reference groups than of proximate reference groups. Providing general support for this postulation, research has found that a media campaign on colorectal screening did not influence perceptions of how many friends were screened, but did affect an increase in perceptions of how many local community members were screened.⁴⁰

H2: Adolescents’ unhealthy food advertising exposure and attention are more positively associated with their descriptive norms about unhealthy food consumption for distal reference groups (i.e., friends at school) than proximate reference groups (i.e., close friends and family members).

Variance across Adolescence

Advertising effects on descriptive norms may also vary across adolescence, including from early to middle adolescence, which is the present study’s focus.⁴¹ As adolescents mature through these stages, their self-concept is in great flux.⁴² Whereas parental influence on a child is strong during early ages,⁴³ adolescents, as they grow older, spend less time with parents and family and more time with friends.⁴⁴ Thus, familial influence on adolescent behaviors decreases through adolescence, whereas peer influence increases. One pertinent empirical study demonstrated the increased effects of descriptive peer norms on adolescent smoking over the course of high school.⁴⁵

This pattern of normative influence has intuitive appeal in the context of adolescent consumption of unhealthy foods. As adolescents age, they enjoy greater independence and increased freedom in eating habits.⁴⁶ In spending less time at home, they eat dinner more frequently away from their families⁴⁷ and are exposed to more external dietary influences.⁴⁸ Thus, as they become less influenced by family, adolescents are more impacted by other factors, such as peers and the media.⁴⁹ Importantly, with maturation from early to middle adolescence, adolescents become more able to make their own choices about what media messaging they expose themselves to and how they process it.⁵⁰

H3: Adolescents’ unhealthy food advertising exposure and attention are more positively associated with their descriptive norms about unhealthy food consumption for older adolescents than younger adolescents.

Method

Data are from an online survey of adolescents aged 13-17 in the United States (N = 1,436) that was conducted from May 27 to June 8, 2011. Survey Sampling International (SSI) provided the survey sample.⁵¹ SSI’s selection of survey respondents for the present study involved stratified random sampling to approximate the U.S. census in terms of this study’s population, but there were some noteworthy differences between the resulting sample and the population parameters.⁵² There was a completion rate of 92.4%. The survey was hosted online. Regarding readability, the survey instrument had a Flesch score of 62.7 and a Flesch-Kincaid Grade Level of 7.5.

Measurement

The measures of descriptive norms, advertising exposure and attention, and unhealthy food consumption were for each of the following four unhealthy food contexts: sugar-sweetened drinks, potato and tortilla chips, food from fast food restaurants, and chocolate bars and pieces.⁵³ Measures of descriptive norms were adapted from previous research.⁵⁴ In each food context, there were measures for family, close friends, and students at school.⁵⁵ Factor analysis (principal components with oblique rotation) indicated three dimensions: (1) student norms (eigenvalue = 3.23, variance explained = 27.0%; α = .77), (2) close friends norms (eigenvalue = 2.40, variance explained = 20.0%; α = .76), and (3) family norms (eigenvalue = 2.29, variance explained = 19.1%; α = .75).

There were eight advertising items, one on exposure and one on attention for each of the four aforementioned unhealthy food types.⁵⁶ Factor analysis (principal components with oblique rotation) indicated two dimensions: (1) advertising attention (eigenvalue = 3.26, variance explained = 40.7%, α = .92) and (2) advertising exposure (eigenvalue = 3.09, variance explained = 38.7%, α = .90).

Control variables included demographics, media use, body mass index (BMI), unhealthy food consumption, number of parents living at home, and parental education.⁵⁷ Demographics were age, education (i.e., grade level), grade-point average (GPA), gender, and ethnicity. Media use was measured (hours per day) for television, video games, and the Internet. Unhealthy food consumption was specific to each of the four aforementioned unhealthy food types.⁵⁸ Factor analysis (principal components) supported the grouping of these food items (eigenvalue = 2.25, variance explained = 56.4%, α = .72).

Statistical Analysis

Stata 13 was used for statistical analysis. Data were initially screened and examined for normality and outliers. There was an ordinary least squares (OLS) regression model for each of the three normative reference groups. Variables were entered in three blocks, with control variables in Block 1 and advertising variables in Block 2. Blocks 3a-b each added a two-way interaction term between age and advertising attention or exposure. Given the possibility of multicollinearity,⁵⁹ each interaction term was inserted separately in Blocks 3a or 3b. In the testing of interaction terms, predictors were centered around their mean, predictor and criterion variables were standardized, and unstandardized regression coefficients were interpreted.⁶⁰ Significant interactions were plotted at three points: one standard deviation below the mean, the mean, and one standard deviation above the mean.⁶¹ Using Stata’s svy function, the regression models implemented probability weights, which were computed according to age, ethnicity, and gender.

Results

Descriptive statistics are presented in Tables 1 to 2. In addition, there were significant differences between family norms and both close friends norms, t(1434) = 24.84, p < .001, and student norms, t(1,434) = 22.97, p < .001, but not between close friends norms and student norms, t(1434) = 1.69, p = .09. The effects of control variables are depicted in Block 1 of Table 3.

Table 1.

Descriptive Statistics for Control Variables (N = 1,436).

	M (%)	SD
Age
13	16.3
14	17.9
15	21.4
16	21.3
17	23.1
Ethnicity
White	65.9
Black	13.3
Hispanic	13.9
Asian	4.0
Other ethnic group	2.9
Education
5th	0.3
6th	2.0
7th	11.1
8th	15.0
9th	21.0
10th	19.6
11th	19.8
12th	11.4
Gender (male = 1)	49.2
Grade point average	3.17	0.73
Body mass index	22.02	5.08
TV viewing	3.22	1.84
Video game playing	1.84	1.84
Internet	2.83	2.10
Parents at home	1.63	0.53
Parental education
Kindergarten	0.1
Elementary school	0.3
Middle school/Junior high school	4.9
High school	31.6
Junior college/two-year degree	24.6
Four-year college degree	24.6
Master’s degree	10.2
Doctoral degree	2.1

Table 2.

Descriptive Statistics for Unhealthy Food Consumption, Advertising, and Normative Variables (N = 1,436).

	M	SD
Unhealthy food consumption
Sweetened drink ads	4.58	1.67
Potato and tortilla chip ads	3.85	1.10
Fast food restaurant ads	3.08	0.98
Chocolate bar/pieces ads	3.30	1.14
Unhealthy food advertising exposure
Sweetened drink ads	5.17	3.47
Potato and tortilla chip ads	4.34	3.12
Fast food restaurant ads	6.02	3.56
Chocolate bar/pieces ads	5.42	3.62
Unhealthy food advertising attention
Sweetened drink ads	2.78	1.05
Potato and tortilla chip ads	2.87	1.08
Fast food restaurant ads	3.07	1.11
Chocolate bar/pieces ads	2.98	1.15
Descriptive family norms
Sweetened drink consumption	2.97	0.93
Potato and tortilla chip consumption	3.28	0.74
Fast food restaurant consumption	2.92	0.90
Chocolate bar/pieces consumption	2.97	0.84
Descriptive close friend norms
Sweetened drink consumption	3.39	0.70
Potato and tortilla chip consumption	3.49	0.59
Fast food restaurant consumption	3.39	0.64
Chocolate bar/pieces consumption	3.30	0.65
Descriptive student norms
Sweetened drink consumption	3.41	0.67
Potato and tortilla chip consumption	3.53	0.58
Fast food restaurant consumption	3.47	0.58
Chocolate bar/pieces consumption	3.39	0.60

Table 3.

Ordinary Least Squares Regression Predictors of Unhealthy Food Consumption Descriptive Norms.

	Descriptive family norms	Descriptive close friends norms	Descriptive student norms
Block 1
Age	−.13**	−.09	−.08
Gender (male = 1)	−.01	−.03	−.06
Education	.12**	.10	.10
Grade-point average	.04	.00	.02
White	−.04	−.06*	−.10***
Parental education	−.06**	−.01	.03
Parents at home	.04	.02	.01
Unhealthy food consumption	.57***	.38***	.30***
Body mass index	.00	−.01	−.03
TV viewing	.11***	.04	.02
Video game playing	.00	−.02	.01
Internet use	−.01	.01	.01
R²	.37***	.16***	.11***
Block 2
Unhealthy food advertising exposure (EXP)	−.03	.04	.08**
Unhealthy food advertising attention (ATTN)	.20***	.22***	.18***
ΔR²	.03***	.04***	.04***
Block 3a
Age × EXP	.01	.07*	.03
ΔR²	.00	.01*	.00
Block 3b
Age × ATTN	.06*	.03	.01
ΔR²	.01*	.00	.00

p < .05. **p < .01. ***p < .001.

H1 was supported in four of six cases (see Block 2, which brought about a significant increase in R²). Advertising exposure was associated with descriptive student norms but not descriptive close friends norms or descriptive family norms. Advertising attention was associated with descriptive family norms, descriptive close friends norms, and descriptive student norms. These effects, as well as the effects of Block 2 as a whole, were significant after accounting for the control variables.

Support for H2 was limited to advertising exposure (see Block 2). Advertising exposure effects on social norms for the distal group (descriptive student norms) were greater than those for the two close or proximate groups (descriptive family norms and descriptive close friends norms). Post hoc analysis with Stata’s suest command indicated that these differences were significant, with advertising exposure’s effect on student norms significantly stronger than its effects on family norms, F(1, 1435) = 13.85, p < .001, and close friends norms, F(1, 1435) = 6.45, p < .05. Counter to the hypothesis, the effects of advertising attention were greatest on descriptive norms specific to close friends, followed by family, and students. However, post hoc analyses using Stata’s suest command demonstrated that advertising attention’s effects did not vary significantly among the three normative measures.

H3 was supported in two of six cases (see Blocks 3a and 3b). The association between advertising exposure and descriptive close friends norms was significantly moderated by age. This effect of advertising attention became more positive as adolescent age increased, narrowing the normative gap (see Figure 1). The association between advertising exposure and descriptive close friends norms was significantly moderated by age. This effect of advertising exposure grew more positive as adolescent age increased, diminishing the normative gap (see Figure 2). In both cases, the change in R² was significant for Block 3.

Figure 1.

Relationship between unhealthy food advertising attention and descriptive family norms as moderated by age.

Figure 2.

Relationship between unhealthy food advertising exposure and descriptive close friends norms as moderated by age.

Discussion

The present study relied on social comparison theory to postulate advertising effects on descriptive norms in the context of unhealthy food consumption. Advertising was generally a significant predictor of descriptive norms, a finding congruent with previous research on media campaign effects.⁶² This predictive role was theorized to result from both dynamic social processes and media content. Children aged 8-18 are inundated with media, including television,⁶³ which is commonly embedded with advertising that depicts food consumption norms.⁶⁴ Such media depictions are secondary factors in primary socialization theory, determining how adolescents see the world around them⁶⁵ and make appraisals related to self-evaluation, self-improvement, and self-enhancement.⁶⁶ Less clear, however, is whether adolescents tend to process unhealthy food advertisements via upward or downward comparison.⁶⁷ Depending on their specific evaluation of unhealthy foods, adolescents could view mediated unhealthy food consumers to be more fortunate or less fortunate than themselves.

The effects of advertising attention were sizable and significant in predicting each normative outcome, while the effects of advertising exposure were significant in only one case (i.e., descriptive student norms), and that effect was significantly weaker than the analogous effect of advertising attention. These relative effects are generally consistent with research on news and political participation,⁶⁸ with such variance justified in terms of conceptual differences, with attention being a cognitive measure and exposure being a behavioral measure.⁶⁹

In addition, because adolescents tend to base their normative construction on primary factors of socialization for close or proximate groups, but on secondary factors of socialization for distant groups, we postulated that advertising effects on normative development would be greater for distal reference groups than for proximate or close reference groups. There was empirical support for this hypothesis in terms of advertising exposure, but not advertising attention. These findings are consistent with one previous empirical study on a colorectal screening media campaign.⁷⁰ The lack in variance in advertising attention effects may result from its reaching a ceiling (with βs from .18 to .22), while the significantly lower effects of advertising exposure (with βs from −.03 to .08) enjoyed greater room for variance.

In regards to the question of whether advertising effects on normative development may vary across adolescence, analysis provided modest support. Both significant cases are consistent with the notion that, as adolescents age, they are more prone to the influence of unhealthy food advertising in the development of descriptive norms. These findings generally support the idea that, as children spend more time away from family and home, they are further exposed to external influences on diet,⁷¹ tend to eat away from home more often,⁷² and become more independent in what media they watch and how they cognitively process media messaging.⁷³

The findings also have implications for practice, specifically providing support for counteradvertising campaigns and food advertising restrictions. Given the obesity epidemic in the United States,⁷⁴ challenging the norms set forth by unhealthy food advertising in counteradvertising campaigns may be an effective strategy to reduce adolescent consumption of unhealthy foods. In addition, decreasing the prevalence of unhealthy food advertising to adolescents through regulation could be a means to altering normative development and helping turn back the trend in obesity. Efforts to restrict food advertising have been fairly common,⁷⁵ but are limited because they are voluntary and, at times, specific to children and not adolescents. The exclusion of adolescents is especially critical considering the present study’s documentation of how the relationship between advertising exposure and descriptive norms is exacerbated across adolescence.

There are six notable limitations. First, this study relies on cross-sectional data, which disallows the demonstration of causation. While social norms may influence advertising exposure and attention, the posited direction from advertising exposure and attention to social norms is consistent with our theorizing and with prior research that has used experimental methods,⁷⁶ as well as time-lagged data⁷⁷ and longitudinal data.⁷⁸ Second, the effect size of the two significant interaction terms was small. However, these interaction effects were significant after accounting for twelve control variables and two advertising variables—and brought about significant increases in the respective R² values. In addition, our sample size of 1,436 is in line with recommendations of using a sample of size of at least 1,045 for detecting small effects in regression models with fifteen independent variables.⁷⁹

Third, our study focused exclusively on descriptive norms, disallowing comparisons with other types of norms (e.g., injunctive). However, injunctive and descriptive norms are often consistent with one another,⁸⁰ and this study’s survey instrument permitted comparisons across reference groups for descriptive norms. Fourth, the survey instrument measured advertising exposure and attention in a media-general manner, one that does not enable comparisons across media (e.g., television vs. print). However, television is known to be the largest source of food marketing to adolescents.⁸¹

Fifth, the advertising measures are based in self-report. This approach, however, is common in survey literature that makes assessments of advertising, news, and health campaign effects.⁸² Sixth, there are possible alternative explanations for demonstrated relationships. To help address such scenarios, twelve control variables were implemented, including demographics, media use, BMI, unhealthy food consumption, parental education, and number of parents living at home.

While the present study focused on the predictive role of advertising, it is recommended that future research also consider the influence of interpersonal communication and media campaigns that discourage unhealthy food consumption. With regard to interpersonal communication, it could be a case of relative influence (e.g., advertising effects vs. interpersonal communication effects), step-like influence (i.e., advertising precipitating interpersonal communication), or synergistic influence (i.e., advertising and interpersonal communication having interactive effects). Finally, it is recommended that future research examine similar advertising processes while including injunctive norms and using panel survey data to make causal inferences.

Footnotes

Declaration of Conflicting Interests

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

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.