U.S. flag

An official website of the United States government

NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

National Research Council (US) and Institute of Medicine (US) Committee on Developing a Strategy to Reduce and Prevent Underage Drinking; Bonnie RJ, O'Connell ME, editors. Reducing Underage Drinking: A Collective Responsibility. Washington (DC): National Academies Press (US); 2004.

Cover of Reducing Underage Drinking

Reducing Underage Drinking: A Collective Responsibility.

Show details

6National Media Campaign

The committee was asked, particularly, to consider the role of a national media campaign in preventing and reducing underage drinking. As indicated in Chapter 1, we considered two mass media campaign approaches to affect youth alcohol consumption: a campaign directed primarily at youth, to affect their drinking decisions, and a campaign directed to parents and other adults who influence youth. On the basis of our review of the current evidence, we believe that an adult-oriented campaign holds more promise than a campaign directed at youth. Although there is limited direct evidence of effectiveness for either type of campaign, there is a clear and strong logical foundation for an adult campaign given the facilitative role of adult behavior in underage drinking and the potential preventive effect of parental monitoring. This logical argument is also strongly bolstered by the evidence of effectiveness of media campaigns in related public health areas. In the committee's judgment, this body of evidence provides reason for optimism regarding the potential effectiveness of an adult-oriented campaign on underage drinking and is sufficient to support a presumptive commitment to launch such a campaign after a carefully designed developmental phase. Our assessment of the possible effectiveness of a youth-focused campaign is presented in Chapter 10. In this chapter we address the adult-oriented campaign.

Recommendation 6-1: The federal government should fund and actively support the development of a national media effort, as a major component of an adult-oriented campaign to reduce underage drinking.

The goals of the national media campaign would be to instill a broad societal commitment to reduce underage drinking, to increase specific actions by adults that are meant to discourage or inhibit underage drinking, and to decrease adult conduct that tends to facilitate underage drinking.

Such a campaign would be undertaken in the context of a comprehensive, society-wide effort to address underage drinking through the other mechanisms described in this volume. Those efforts could and should be undertaken while the media campaign is being developed. In this chapter we describe the underlying logic for the approach, what the campaign would look like, evidence concerning its promise, and its possible strengths and weaknesses. We conclude by outlining the developmental approach we propose. We believe that the development of an adult-oriented campaign warrants a substantial investment if its promise of effectiveness is borne out during the developmental period.



Many groups concerned with underage drinking claim that societal acceptance of, or at least ambivalence toward such drinking, reflected both in expressed attitudes and in the actions of many adults in facilitating underage drinking, is a substantial cause for such drinking (see National Center on Addiction and Substance Abuse [CASA], 2002). Yet, when asked, the great majority of adults express concern about underage drinking. For example 84 percent of respondents to the 2001 survey undertaken by CASA said that underage drinking was either “a big problem” or “somewhat of a problem” in their community, and 92 percent said they were personally “somewhat” or “very much concerned” with underage alcohol use. That survey also shows widespread support for many public policy actions, with 83 percent favoring regulation of location of alcohol outlets near school, and 78 percent, 71 percent, and 69 percent supporting undercover compliance checks, “cops in shops” checks, and “shoulder tap” checks, respectively, to reduce alcohol outlet sales to underage youth.

Despite the high level of expressed adult concern, most underage drinking requires involvement by some adults—in selling the alcohol to youth, in purchasing it on behalf of youth, or in permitting minors to have parties in their homes where alcohol is served. Obviously, many adults are not taking proper precautions to restrict underage drinking, and often facilitate it in violation of the law. And even if adults do not supply alcohol to minors directly, underage drinking is tacitly supported by many adults—such as parents who view youth drinking as an inevitable part of adolescence and do not respond to it, or who do not make sure that parties their children attend are alcohol free and properly supervised by adults. Even though adults tend to favor some strong measures to prevent underage drinking, the CASA study suggests that they are least enthusiastic about regulations that would affect them directly: only 60 percent favor restriction on the number of alcohol outlets, and 51 percent favor limitations on outlet days or hours. The reluctance of policy makers to enact alcohol restrictions that might reduce youth use but also affects adult use (such as controlling outlets) presumably reflects, in some part, perception of the lack of public support for such restrictions.

Adults themselves do not think that parents are doing all that they can do to prevent underage drinking. There is substantial recognition by adults that parents are the most important channel of influence on their children's underage drinking. When asked to choose what was most responsible for “preventing us from effectively reducing underage drinking” more than half of the adult respondents said “lack of or limited parental involvement in teens' lives” (see Table 6-1). This response indicates that there is another way that adult behavior may support underage drinking in addition to explicit or tacit support for alcohol use, which involves ineffective parenting more generally.

TABLE 6-1. Adults' Reports of Barriers to Reducing Underage Drinking (in percent).


Adults' Reports of Barriers to Reducing Underage Drinking (in percent).

The parenting literature argues that effective parenting includes monitoring and supervising youth behavior. For younger adolescents, this parenting includes such things as: knowing who a child's friends are, making sure that children are always supervised by adults, knowing what a child's plans are for the coming day, knowing what children are doing when away from home, and enforcing evening curfews. It also includes engagement with children's lives, including doing projects and activities with them, and the use of appropriate punishments for misbehavior and rewards for positive behavior. There is good evidence that such parenting is associated with a reduced risk of using all substances, including alcohol (see Chapter 4). A recent national longitudinal survey of parents and their 9- to 18-year-old children supports this claim. Children of parents who were relatively high on a supervision and monitoring scale were compared with children of parents who were low on the monitoring scale. Less monitored youth were more likely, subsequently, to progress to alcohol consumption in the next 12 to 18 months, and if they were already drinkers, were more likely to continue drinking.1


There is significant evidence that parents are unaware of the extent and riskiness of youth drinking. The national longitudinal survey mentioned above compared reports of alcohol use among youth aged 12 to 18 with their parents' perception of their alcohol use. Both parents and their children were asked about whether or not the child had used alcohol, more than a few sips, in the previous 12 months. Overall, parents moderately underestimated what their children reported as use: for 12- to 13-year-olds, the parents thought that 7 percent had used alcohol, but 11 percent of their children said they did; for 14- to 15-year-olds, the comparable numbers were 21 percent and 33 percent; and for 16- to 18-year-olds, the numbers were 44 percent and 56 percent.

An even more telling way to look at these data is to turn them around and ask how often parents knew when their own child was drinking. Including all of the children from 12 to 18 years old, 44 percent of all the youth who had had drinks in the past year were described by their parents as nondrinkers. Moreover, 31 percent of the youth who said they had been drunk in the past year were said by their parent to be nondrinkers, and 27 percent of those who said they had had five or more drinks in the past month were said by their parents to be nondrinkers. While a majority of parents may know whether or not their children drink, there is a substantial fraction who do not, even when their children admit to recent heavy drinking (see also Sieving, 1997; Beck et al., 1995).

A 1998 study by Bogenschneider and colleagues found that less than one-third of parents (29 percent of mothers and 31 percent of fathers) were aware of their adolescents' drinking even though all of the adolescents reported using alcohol at least once in the past month.2 Parents were more likely to report use by their adolescents' close friends than by their own children. More than one-half of parents (56 percent of both mothers and fathers) reported that they were not sure or thought it was likely that their children's close friends use alcohol.

Bogenschneider and colleagues also found that adolescent children of aware parents were more likely to drink and drive than adolescent children of unaware parents. In response to these unanticipated results, the researchers postulated “episodes of drinking and driving serve to alert parents to the possibility that their offspring use alcohol” (1998:369). If this hypothesis is true, parents become aware of their children's drinking after they have engaged in a risky behavior.

Additional evidence that parents are not aware of teenage drinking comes from two surveys conducted by the state of Maine: one of parents of teenagers and the other of teenagers.3 Although more than half of the parents surveyed (55.6 percent) reported that they are more concerned about teenage use of alcohol than tobacco, marijuana, other illegal drugs, or prescription drugs, these parents greatly underestimate the extent to which eighth to twelfth graders drink alcohol (Strategic Marketing Services, 2002). Table 6-2 presents the differences between the reports of the parents and the teenagers. The discrepancy between parent and youth reports is particularly notable for heavy drinking: although nearly all parents (99 percent) reported that they did not think their children had had five or more drinks in the past 2 weeks, one in five youth (20 percent) reported having done so at least once.

TABLE 6-2. Parents' and Teenagers' Reports of Alcohol Use (in percent).


Parents' and Teenagers' Reports of Alcohol Use (in percent).

While the majority of parents (56 percent) in the Maine Survey reported having serious talks about alcohol with their child several times a year, only slightly more than a third (34.2 percent) said that they had these discussions once a month or more. Not surprisingly, drinking and driving was the most common topic discussed, with 71 percent of parents reporting that they discussed this issue. Other primary topics, though less common, included the effects of alcohol on judgment or decision making (50.7 percent), peer pressure (48.9 percent), negative medical effects of alcohol (34.9 percent), and parental feelings about underage drinking (34.1 percent).

Clearly, many parents do not know when their children are drinking. However, even if the parents know their children are drinking, there is a question of whether they see underage alcohol consumption as risky. The evidence concerning whether or not parents perceive risks in underage drinking comes from the study by CASA (2002). In Table 6-3, we present the proportion of adult respondents who indicated that each of the named consequences of underage drinking was a concern.

TABLE 6-3. Adult Reports of Concern for Potential Consequences of Underage Drinking (in percent) (N = 900).


Adult Reports of Concern for Potential Consequences of Underage Drinking (in percent) (N = 900).

There are two ways to read Table 6-3. One perspective notes that a majority of adults recognize every potential risk as a matter of concern. This is reinforced by the finding that, at most, only 14 percent of the respondents indicated that any of the potential consequences was not a concern at all. This view emphasizes that people recognize the risks. An alternative perspective notes the minimal discrimination among the various consequences. This lack of discrimination among consequences suggests that the respondents have not thought about the consequences seriously and suggests that adults only recognize the risks in a rote fashion, when primed.


The primary role of a societal, adult-oriented media campaign would be to convince parents and other adults not merely that there is a general problem with underage drinking in their communities, an idea that they appear to accept already, but also that it is very likely a problem for their own children and their children's friends, that there are important negative consequences of such alcohol use besides those risks associated with drinking and driving, and that they have an obligation to their children and the community to do something about it. The campaign would argue that by taking specific personal actions to prevent underage alcohol use, by increasing recommended parenting behaviors, and by support of community-level policies, parents and other adults can affect underage drinking and reduce its bad consequences.

The campaign rests on five assumptions:

  • Many parents do not recognize either the prevalence of or the many risks associated with underage drinking for their own children.
  • Many parents effectively facilitate their underage children's drinking by giving youth access to alcohol, by not responding to known incidents of children's drinking, and by not adequately monitoring and supervising their children's lives, generally.
  • If parents changed their beliefs about the nature of underage alcohol use and its consequences, they would increase monitoring and other actions to limit their children's use.
  • Because many underage drinkers obtain their alcohol from adult acquaintances or even strangers, if adults' willingness to buy alcohol for young people or to facilitate their drinking decreases, it would be more difficult for underage youths to obtain or use alcohol.
  • If parents and other adults increased monitoring and other actions aiming to limit use, there would be a reduction in underage drinking, particularly heavy drinking.

What are the arguments in favor of such an adult-oriented campaign? One argument is that campaigns that offer new information are more promising than campaigns that revisit information that already has been widely distributed. Youth have often heard anti-alcohol messages addressed to them, but they have shown little change in recent years in most drinking behavior, although they have been somewhat responsive to the drinking and driving message (see below). A new youth-focused campaign would be seen as old hat and redundant with what they are already hearing. In contrast, an adult-oriented campaign would present new messages. It would target parents and other adults, who are now facilitating youth alcohol use because they are not sufficiently aware of the problem for their children; not aware of the many harmful consequences of youth alcohol use; not aware that actions they take can affect the risks; and not aware that buying alcohol for underage persons or giving it to them, is socially irresponsible and usually illegal. For adults, in contrast with youth, there is the possibility of a communication program diffusing new information, which suggests greater effectiveness. In addition, it is possible that an adult-focused campaign may work to make what parents already believe more salient to them as they consider their actions to restrict their children's alcohol access. Effectively, a campaign can give them permission to act on the concerns they already have.


Disseminating Facts

An important question is what evidence is available that a large-scale communication campaign could affect awareness of extent and riskiness, and (assuming that such knowledge leads to motivation to act) teach effective actions (for review, see Atkin, 2004).

There have been a small number of campaigns that tried to affect adult awareness of the extent and perceived riskiness of underage drinking and parental actions to reduce underage drinking, but they do not provide a solid foundation for estimating the promise for this approach. For example, the Australian National Alcohol Campaign in 2000 and 2001 primarily addressed youth, with some magazine advertising and brochure distribution to parents. However, the parent component was probably too small to expect much effect, and the evaluation information for parents does not provide an adequate basis for determining its behavioral effect (Ball et al., 2002). We then turn to less direct evidence that such a campaign will influence parent knowledge and behavior, relying on a reasonable generalization of evidence from other programs. In some sense, it is possible to separate two aspects of such a campaign and ask about the availability of evidence for each.

Insofar as an adult-focused campaign is only presenting facts, not previously known by parents, there is good evidence that this can be accomplished readily. Diffusing facts is what communication programs do well.

There are many examples of diffusion of a new idea or set of facts through mass media. And there are some examples that suggest circumstances when simple diffusion of new facts was sufficient to produce behavior change. In one clear case, campaigns to encourage parents to put their infants to sleep on their backs to avoid sudden infant death syndrome have had fairly quick and widespread success (Willinger et al., 2000; Engelberts et al., 1991). This behavior required a small change by parents and the changed sleeping position promised to avoid a dreaded consequence. Of most relevance for this discussion, the value of the back sleeping position represented new information for most parents.

Closer to the alcohol area, the idea of the “designated driver” diffused rapidly in the United States. The first mention of the term in the Lexis-Nexis electronic news major papers database is the fall of 1982, which is presumably when it was introduced to public discussion. By 1987, 91 percent of the respondents to a Gallup Poll indicated they approved of the idea (Roper Center at the University of Connecticut, 1987). By 1988, the term could be used in questions without a parenthetical definition, suggesting that it was well known. The proportion of adults who reported using designated drivers also grew rapidly (Roper Center at the University of Connecticut, 1988; see also Winsten, 1994).

In terms of diffusing facts, the major issue is not whether it can be done, but whether a comprehensible and credible message can be transmitted with sufficient reach and frequency so that most people become aware of it and whether the particular facts encourage behavior change.4

Changing Behavior?

The more difficult claim to support is not whether a mass media campaign can diffuse facts but whether it can effectively teach and, most important, influence specific new behaviors by adults. We do not know of any studies of a specific intervention of this sort relating to alcohol. The closest published example comes from Project Northland (Perry et al., 2002), which used only community-level media, along with other community activities, to try and affect parenting norms and behaviors. That limited intervention showed little effect on parental attitudes or behavior (although, as we discuss in Chapter 10, the initiative overall provides evidence for positive effects on youth). However, like the Australian campaign described above, Project Northland is a much more limited program than we propose, which would involve a national focus, heavy use of mass media, and involvement of the alcohol industry and other important institutions.

There is evidence that mass media campaigns, usually combining publicity and law enforcement, have succeeded in influencing drinking and driving (Murray, 1991; Bierness et al., 2000; Voas et al., 1997; Nienstedt, 1990; Hurst and Wright, 1981; Cameron and Newstead, 1996; Tay, 2000). There is also evidence that campaigns have influenced the use of designated drivers (Winsten, 1994; Dejong and Hingson, 1998; Boots and Midford, 1999; Hingson et al., 1996).

Although there have been many campaigns to affect adult drinking behavior (other than driving), they have generally been understudied. For example, the National Institute on Alcohol Abuse and Alcoholism sponsored campaigns in 1971-1972, and in 1980-1982, and there is some published information about regular campaigns about drinking in Denmark from 1990 to 1996 (Strunge, 1998) and about single campaigns in other places. However the evaluations of those campaigns are weaker than those addressing drunk driving, and they often do not measure behavior or have credible comparison groups.5 In any case, these results, even the favorable ones about drunk driving, are not the same as evidence that parenting behaviors concerning their children's alcohol use can be affected. We recognize that in contrast to these other campaigns, the recommended campaign does not focus on a single behavior, nor does it focus on specific changes in adults' own behavior with regard to alcohol use. It urges parents and other adults to accept and act on a broad social norm, and it therefore has a longer time horizon than many of these more focused campaigns.

Although the available evidence bearing on the effectiveness of an adult-oriented campaign is modest, the committee is reasonably optimistic about the potential value of such a campaign, for two reasons. First, there is associational evidence that periods of broad national campaigns incorporating a variety of channels and institutional change efforts have been matched by periods of reductions in risky behavior. Those examples include the National High Blood Pressure Campaign from 1972 to 1984 (Roccella, 2002), the anti-tobacco efforts of the late 1960s-early 1970s (Warner, 1981) and the late 1990s-early 2000s (Siegel, 2002), and the antidrug campaigns of the middle 1980s (Institute of Medicine, 2002).

Second, there is particularly strong evidence for the positive effects of media campaigns, which are able to link communication efforts with enforcement. The drinking and driving efforts described above were successful when they were able to link their messages to a specific expectation of enforcement. Similarly, seat belt use has climbed quickly when media publicity is linked to enforcement (Williams et al., 1996). Indeed one metaanalysis of the literature concludes that the largest effects evident in the mass media campaign literature come from campaigns which link media publicity with enforcement (Snyder and Hamilton, 2002).

Also pertinent to this aspect of the proposed campaign, both for parents and other adults, is the literature on compliance with the law. A central theme in the proposed campaign is that facilitating underage drinking is not only socially irresponsible but also illegal in most situations. Deterrence—through the threat of criminal prosecution or of civil liability for any injuries to third parties—can be part of the message, but the more powerful mechanism may be through the “expressive” or “declarative” function of the law, the mechanism through which the law registers social disapproval, teaches that the behavior is perhaps more dangerous than may have been appreciated, and thereby instills or reinforces the desired social norm. The mechanism here is not fear, but rather a powerful form of instruction drawing on the general desire to comply with legal rules (see Bonnie, 1985; Tyler, 1992; Tyler and Huo, 2002).

This discussion has focused on intervening with parents and other adults in ways that encourage their active prevention of underage drinking specifically. However, there is a second route of intervention with parents that might well be incorporated into the proposed campaign. Evidence shows that the extent of supervision and monitoring by parents of youth, in general, affects youth initiation of risky behaviors and that intensive individual or group parent counseling can affect parenting behavior and, in turn, youth risk behaviors, including alcohol consumption (Taylor and Biglan, 1998; Dishion and Andrews, 1995; Dishion et al., 2002). So a reasonable argument has been made that it would be worthwhile to attempt to communicate about new parenting behaviors, particularly through the use of advertising.

The Office of National Drug Control Policy (ONDCP) took this task on as a complement to its youth-focused campaign, described in Chapter 10. It dedicated nearly an equal amount of its ad campaign to parenting skills, particularly encouraging the close monitoring of children by parents. The expectation was that success in encouraging closer monitoring would in turn affect youth use of drug. Thus far the evaluation of this component of the anti-drug campaign suggests mixed results (Hornik et al., 2002). There is some evidence of an effect on the extent that parents talk with their children about drugs and for an effect on parental beliefs about the value of monitoring. The evidence for effects on actual monitoring and supervision is less strong, however. The ONDCP campaign is still ongoing, so definitive results are not yet available.

It may be possible that some actions to prevent youth drinking are easier to affect than general parent monitoring. For example, it may be easier to convince parents and other adults to stop facilitating drinking parties by their underage children than it is to encourage them to systematically monitor their children. Although this may be a sensible argument, there is not yet evidence to support it.


In sum, there are several arguments in favor of a campaign aimed at parents and other adults: they often do not know about their children's drinking behavior; they probably do not have a well-developed understanding of the specific risks of drinking; communication campaigns are often quite good at diffusing new knowledge; such campaigns have been successful in promoting specific protective behaviors by parents and in changing attitudes and behaviors relating to drunk driving; and parental monitoring of their children is prospectively related to their likelihood of using alcohol. At the same time, the evidence is not clear as to whether such knowledge about their children's risk of alcohol use and abuse, and of the negative consequences of such use, affects parents' behavior with regard to alcohol. In addition, there is as yet no evidence of any campaign effects on relevant parental behavior, particularly on monitoring and supervision behaviors. Yet the campaign proposed here is sufficiently different than previous efforts that we ought not be constrained by the lack of clear prior success. It is different because it intends a more comprehensive approach, combining media attention with community efforts; it is different because it addresses both general parenting skills and specific adult behaviors that may facilitate underage drinking; and it is different because it is taking the long view, recognizing that there is a need for transformation in that community social norms underlie underage drinking.

A complex additional concern has to do with the indirectness of this approach. A campaign directed toward youth will be successful if it affects youth behavior, its immediate target. A campaign directed toward adults is only successful if it first affects adults' behaviors and if those behaviors then affect youths' behavior. Such a two-stage approach may appear to be less promising than a direct one if the goal is to reduce youth drinking. However, from a longer-term perspective, the goal is to change or strengthen the social norm against facilitation of underage drinking, and any effects on youth are secondary to this change in the normative climate. The goal is to convince parents and other adults to embrace the idea that they need to take their own obligation seriously and that they need to hold other parents and adults to a high standard as well. The success of this effort cannot be judged, in the short term, by whether it has an effect on underage drinking. A reasonable goal, in the very short term, would be to show an effect on adult knowledge and attitudes, and, in the intermediate term, to show an effect on adult behavior. If the campaign has these effects, the committee is reasonably confident that it would ultimately have an impact on youth drinking over the long term.


It is premature for the committee to propose a particular structure or content for the media campaign before a program of formative research has been carried out.6 However, to give Congress and other interested audiences a more concrete idea of what we envision, we briefly describe a possible structure for such a campaign.

An adult-focused campaign might have three major themes. At the start, much of the emphasis might on broad social norms. As detailed above, most parents consider youth alcohol use to be problematic, but do not accurately perceive the risk to their own children and fail to recognize their own role in influencing their children's alcohol use. The earliest phase of the campaign might work on these perceptions, readying parents and other adults for the need to change their own behavior insofar as it facilitates or condones underage alcohol use. If the legal messages are determined in the developmental phase to be significant with the target audience, this phase might also include information about the law governing underage drinking and the legal duties of parents and other adults.

The preparatory phase might serve as the foundation for the next phase a more action-oriented phase that might include two complementary approaches. One approach might focus on specific beliefs related to youth alcohol use that has been shown by preliminary research to motivate adult action—such as perception that youth drinking is harmful in specific ways other than drinking and driving or that parents expect other parents to take action to prevent youth drinking. The accompanying message might recommend specific actions to prevent youth alcohol use, such as not hosting teenage parties that provide alcohol and not purchasing liquor on behalf of underage youth. The second, complementary approach might focus on parental monitoring—including supervision, engaging with children, and imposing consistent and appropriate discipline—on the grounds that those skills can have a major influence on underage drinking, as well as other risky behavior.

These themes might be delivered through a variety of channels. Paid media would probably have to play a substantial role in diffusing specific messages and would probably represent the most costly portion of the campaign. However, the messages conveyed through paid media would probably be insufficient to produce a change in parental norms and behavior unless messages supporting such changes were reinforced through other channels that reach parents and adults. Assuring that multiple sources of information and a wide variety of institutions are presenting reinforcing messages would be a crucial objective for the campaign.

Possible Models

One model for this approach can be found in one of the first and most successful national campaigns, the National High Blood Pressure Education Campaign. Although this campaign did some advertising of its messages, it relied heavily on actions by other institutions: campaign planners worked with physicians' organizations to encourage physicians to provide advice about high blood pressure consistent with national guidelines; they proposed stories to newspapers and television and radio that conveyed their priority messages; and they developed affiliations with, and provided materials to, grassroots organizations interested in hypertension (Roccella, 2002).

The California anti-tobacco campaign provides another possible model for this approach. That effort was able to mobilize statewide support for a dedicated tobacco tax: those revenues were then used to purchase media time for anti-tobacco advertising, to support school and community-level anti-tobacco programs, and to promote local policies limiting exposure to environmental tobacco smoke. The media campaign was expected to directly influence smokers or potential smokers, but it also served to energize the population around the state and to promote policy changes designed to further reduce opportunities for smoking. While the increased tax on tobacco surely had independent effects on consumption, there is evidence that the observed reductions in smoking in California reflected the interaction of higher cost, direct advertising, and complementary local activities (Pierce, 2002; Hu et al., 1995).

As we envision it, an adult-focused campaign to reduce underage drinking would work similarly to the California campaign and other successful campaigns. It would make use of a range of channels to reach its audience, continuously adjusting its message strategy, its mix of channels, and its links with grassroots organizations and national leadership groups and policy makers. It would both encourage policy changes and grassroots support for local and national policy changes and use policy changes as a basis for disseminating messages to individual parents and other adults and for encouraging specific behavioral actions.

A Developmental Approach

The committee believes that there is a sound, though limited, evidentiary and logical foundation for an adult-focused campaign. Given the limited knowledge, however, some people may be reluctant to support a national campaign and suggest, instead, pilot programs to test the strategy. The committee disagrees with this view because underage drinking is so well embedded in U.S. culture that small-scale prototype tests may miss the point or at least find it difficult to make much headway.

Because the acceptability of underage drinking is widespread, it will take more than an isolated campaign to affect it, no matter how well designed such a campaign might be. It will take a multipronged effort over a long period, making use of as many routes of influence as possible. We think the evidence is sufficient to support a presumptive commitment to the idea of launching such a campaign after a carefully designed, step-by-step developmental phase. Thus, while we do not favor a small-scale pilot test of the campaign as a whole, we do recommend pilot testing of specific features of the campaign during the developmental period, as discussed below.

The Challenge

The history of campaigns aimed at tobacco use and drunk driving are instructive in indicating why we think that the presumptive commitment is warranted. The anti-tobacco efforts have been successful because they changed the culture about smoking. That cultural change has permitted many other things about smoking to change (social and legal norms about where people can smoke, etc.). Drunk driving has been subject to a similar cultural shift. Grassroots lobbying from Mothers Against Drunk Driving and others, along with specific government-sponsored interventions and new regulations and enforcement policies, transformed a behavior that was once perceived as risky, but not strongly condemned, into conduct that is universally regarded as socially unacceptable.

If one assumes that the culture around underage drinking needs to change in the same way that the cultures around drunk driving and tobacco use have changed, then that history becomes the model. National campaigns are part of a broad effort to affect the culture. It is instructive that both of these exemplars reflect interactions between private and public entities. They are better described as social movements, in which the government's role is to respond to, support, and stimulate private action.

In this approach, a government-sponsored communication campaign has as its goal the support of changes in social norms that relate to adult behavior insofar as it enables and facilitates underage drinking. Such efforts are not to be evaluated by their short-term effects on youth drinking, but by their effects on broad social norms about that drinking, by the willingness of adults to take actions to reduce it, and by changes in public policies known to affect underage drinking rates.

An adult-focused mass communication campaign is also meant to support local efforts to reduce drinking. It is important not only because of what it does on its own, but also because its effects provide leverage for local efforts—and vice versa. It would link its activities to the broadest group of adult stakeholders—industry, colleges and universities, the military, and community organizations. Wherever possible, a national campaign would coordinate activities with local needs and provide for the tailoring of its messages for different communities. It cannot be tested as a prototype on a very local scale because its effectiveness depends on the involvement of a wide range of constituencies and, ideally, the engagement of the entire nation's attention.

Our Approach

In the end, the committee is faced with a conundrum in formulating its recommendation for developing and implementing an adult-focused media campaign. On one side, we find the idea to be highly promising, but lacking the kind of direct evidence needed for unequivocal endorsement of the high costs of such a campaign. On the other side, testing the campaign in a very limited way in order to gather more evidence is not a viable option because a small-scale effort (relying, for example, only on paid media messages delivered in one locality alone without any reinforcement by national or even regional media and other institutional partners) would almost certainly produce unimpressive results (see Institute of Medicine, 2002). In short, implementing an adult-oriented campaign in the absence of an opportunity to build the social movement around the idea would not be a fair test of the approach. Is there a middle course? We suggest a substantial effort to develop the program, with the aim of erecting circumscribed conditions of demonstrated efficacy that would have to be satisfied before the decision is made to implement a national campaign.

The proposed campaign would be developed over time on the basis of an intensive agenda of formative research designed to define the promising messages, the appropriate channels to reach target audiences, and the timing and weight given to various messages and channels over time. Campaign strategy would reflect initial research with target audiences and continual monitoring of how they respond to the campaign as it evolves. A central issue would be how best to link the campaign with local, state, and national policy changes about youth alcohol use, if they are adopted.

During the first phase, we suggest funding the basic formative work for the campaign, including intensive and multifaceted developmental research. This phase might involve controlled tests of whether high exposure to messages leads parents and other adults to be willing to change their behavior, particularly if this can be done in the context of actions by local organizations concerned about youth alcohol use or of local policy changes. While we would be reluctant to expect too much influence from such a highly localized campaign, given the lack of nonlocal sources of reinforcing information, a good deal might be learned about what is promising and what falls completely flat.

A second preliminary phase might build on that local effort and develop larger-scale efforts in one or two states where there was substantial interest among government or grassroots organizations. While the national media and national policy would not be engaged, there would still be opportunities for building a statewide social movement, with analogies to the state-level anti-tobacco campaigns in California, Florida, and Massachusetts. If these statewide efforts show enough promise to suggest that a nationwide effort would be worthwhile, then the decision to build the broad national effort would rest on a strong scientific foundation and would be fully justified.

If results were positive at each stage, a projected schedule might require 3 years for developmental and preliminary testing. This would be followed by 5 years of full operation. In subsequent years, booster efforts might be implemented, as monitoring evidence established a need.



Based on analysis by Robert Hornik of unpublished data collected by Westat for the National Institute on Drug Abuse for the evaluation of the Office of National Drug Control Policy's national youth antidrug media campaign.


Awareness was defined as being unsure or believing that their adolescents' alcohol use was likely.


The parents included in the parent survey were not necessarily the parents of the teens included in the teen survey.


We recognize that any message encouraging parents to recognize that many children drink has the risk of containing an additional implied message: that all kids drink so maybe it is not a big concern. It will be important to test messages for parents to make sure they do not hear such a mixed message.


The best evaluated of these, the winners campaign in California (Wallack and Barrows, 1982), showed some differences in levels of exposure between control and mass media communities and some hints about attitude difference, but none suggesting behavior change. However, the authors of the evaluation caution against generalizing too much from the results since they indicate how far the realized program departed from the proposed intervention.


During this period of formative research, consideration must be given to the growing diversity of the U.S. population. For example, the percentage of Hispanic children increased from 9 to 16 percent between 1980 and 2000 and is projected to increase to 22 percent by 2020. Cultural and linguistic factors will need to be taken in to account for specific racial or ethnic groups.

Copyright © 2004, National Academy of Sciences.
Bookshelf ID: NBK37590


  • PubReader
  • Print View
  • Cite this Page
  • PDF version of this title (14M)

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...