Back to homepage

Related articles from STATS

CAMY's last gasp
Why the end of the Center for Alcohol Marketing and Youth is an opportunity for better research on alcohol advertising and underage drinking - July 2008.

 

 

Alcohol and Advertising
December 2010
Does alcohol advertising make teenagers more likely to drink? It may seem like a rhetorical question, but the answer is far from a slam dunk.

Anti-alcohol advocacy groups such as the Center on Alcohol Marketing and Youth (CAMY), the National Center on Addiction and Substance Abuse (CASA), and the Center on Alcohol Advertising (CAA) regularly assert in reports and studies that alcohol advertising leads to underage drinking. The cumulative implication of this research is that such advertising should be restricted. But the problem is that when the overall body of research on the relationship between advertising and underage drinking is analyzed the evidence of a direct link disappears.

Two reviews by the National Institute on Alcoholism and Alcohol Abuse (NIAAA, 1995 and 2000)  did not find the evidence for a link compelling. As the NIAAA reported in its 10th Special Report to Congress on Alcohol and Health, (2000) there is “very little consistent evidence that alcohol advertising influences per capita consumption, sales or problems.”

The real world would appear to back the NIAAA. The 2009 National Instituteof Health (NIH) report Monitoring the Future found that that the annual prevalence for both "any use of alcohol" and "been drunk" declined slightly for 8th, 10th, and 12th graders between 2002 and 2007 – a period when alcohol advertising increased. (1) Furthermore, The National Survey on Drug Use and Health reported that during the same period, current use of alcohol by 12- to 17-year-olds decreased significantly and did not change for 18- to 20-year-olds. (2)

The reason for the gap between advocacy research and weight-of-evidence reviews of the full research is that the studies purporting to show a link between advertising and underage drinking rely on confusing, often contradictory data. Plus, the many factors involved in trying to measure the effect of advertising on alcohol consumption make this a complex and difficult issue to study.

Advertising theory tells us that while advertising is useful in getting consumers to switch brands or to purchase a more expensive brand it does a poor job of increasing total demand within an industry. In fact, the total advertising elasticity of demand for beer has been calculated at 0.0 – meaning advertising has no impact on total demand. Companies advertise for a specific product, and when they do see an increase in demand, it is at the expense of a competitor. This is why companies spend billions of dollars each year on advertising: they want a larger share of the pie; they are not making a bigger pie.

The divergent position of advocacy research means we need to understand why certain kinds of research provide more reliable evidence than others.

Experimental Research
Experimental research provides controlled testing of causal processes. In the case of underage drinking following exposure to alcohol advertising, researchers must control for variables such as education and after school supervision in order to eliminate their effects.

Researchers have found either no effect or a small, short-lasting effect on attitudes towards drinking following alcohol advertising but have not found a casual link between alcohol advertising and underage drinking. (3) It is important to note that attitudes about drinking rather than actual drinking behavior are measured.

Survey Research
Survey research uses questionnaires and/or statistical surveys to gather data about specific behaviors and attitudes.

The survey research shows that parents and peers are the overwhelming causal factor in underage drinking, along with a young person’s tolerance for risk taking. Study results show a relatively strong association between drinking or drinking initiation and a weak-to-nonexistent relationship to advertising. When non-advertising variables are analyzed, peer drinking is always one of the stronger variables and a young person’s appetite for risky behavior is also strong.

Econometric Research
Economic Research uses regression analysis to look at the relative impact of potential causal variables; in this case, researchers analyze consumption in relation to advertising spending, advertising restrictions or alcohol policies.

A survey of other econometric studies found alcohol advertising exposure to have an extremely small impact on alcohol consumption, particularly when compared to the effects of peer drinking. (4) Price was shown to have a greater influence on youth consumption than advertising.

Reviewing the Research
The 2009 Roper Youth Report, (5) which annually conducts 1,000 face-to-face interviews, asked students what influences them most about their decision to drink or not drink alcoho:. 68 percent  responded “parents” while 11 percent responded “best friend.”  Only 2 percent  said that “what they saw in the media” (radio, TV, magazines, etc.) influenced them and another 2 percent said that “advertising” most influenced them. Just as important, for those students who were employed, the “parents” figure dropped to 45 percent  while the “best friend” percentage increased to 26 percent. 

Researchers have also examined the effect of alcohol branded merchandise and underage drinking. A 2006 article analyzing the relation between ownership of an alcohol branded item by middle school students and initiation of alcohol use found that students who own an alcohol branded item are 1.5 times more likely to initiate drinking relative to students who do not own alcohol branded merchandise but that peer drinking has the largest impact on a student’s decision to drink. (6) Among students whose friends drink alcohol, 50 percent themselves initiate alcohol use. Such a student is 18.6 times more likely to begin drinking than a student whose friends do not drink.

Table 1 shows the odds ratios in different situations from a 2007 study, (7) published in the Journal of Alcohol and Alcoholism. While retail ad exposure is positive (1.5) and owning alcohol branded merchandise is also positive (3.3), having at least one friend who drinks is just as high (3.3) as is risk taking (3.3). 

Table 1
Adjusted Odds Ratio For Ever Drinking Alcohol
Variable Odds Ratio
Retail ad exposure
1.5
Owns alcohol promotional item
3.3
Exposure to parental alcohol use
1.8
At least one friend drinks alcohol
3.3
Risk-taking
3.3

Likewise, Table 2 shows results from a 2006 article, published in the Archives of Pediatric and Adolescent Medicine. The odds ratio for owning alcohol branded merchandise is 1.5. But, peer drinking is 6.1 for those who have “Some” friends that drink and an astonishing 18.6 for “Most/all” friends drinking. Additionally the odds ratio for those who had ever tried smoking, a very risky behavior, was 2.4.

Table 2
Adjusted Odds Ratio For Initiation of AlcoholUse
Variable Odds Ratio
Owns alcohol branded merchandise
1.5
Peer drinking  

Some

6.1

Most/all

18.6
Ever tried smoking
2.4

A 2007 Journal of Adolescent Health article (8) found that while alcohol branded merchandise is associated with both drinking and the intention to drink, the advertising variables are weak. Peer drinking, parental approval, friend approval, deviance, impulsivity, low religiosity and sports activity were all shown to be stronger indicators than any of the advertising variables. With this in mind, the authors concluded: “Although causal effects are uncertain, policy makers should consider limiting a variety of marketing practices that could contribute to drinking in early adolescence.”

A 2006 article calling for a ban on alcohol advertising through sports found that alcohol advertising had increased in the United States by 50.8 percent between 2001 and 2007, and youth exposure to alcohol advertising on television had increased by 38 percent during the same time interval. (9)

However, over the same time period, The National Survey on Drug Use and Health reported that between 2002 and 2007, current use of alcohol by 12- to 17-year-olds decreased significantly and did not change for 18- to 20-year-olds. In addition, there had been no appreciable change in past-month binge or heavy alcohol use among any of the underage groups. Driving under the influence for persons 12 years or older had also decreased significantly. Similarly, Monitoring the Future reported that the annual prevalence for both “any use of alcohol” and “been drunk” declined slightly for 8th, 10th, and 12th graders from 2002 to 2007 (continuing trends observed since 1993). If increased alcohol advertising were linked to more underage drinking, these trends could not have been observed.

While the survey research consistently shows that advertising has, at most, a small effect on teen drinking and a marginally greater effect on teen attitudes about drinking, it cannot show cause and effect. As a result, it is hard to know what is really happening. For instance, while advertising may make alcohol attractive to some teens, they could just as easily enjoy or be attracted to alcohol advertising because they already hold positive attitudes towards drinking.

In other words, the fact that the teens who like alcohol ads are more likely to want to drink may simply reflect their pre-existing attitudes, rather than show an effect of the commercials. The fact that the effect on behavior is much smaller than that on attitudes again shows that one cannot rely on attitudes alone to measure the real world effects of ads.

The price of alcohol has been shown to have a greater influence on youth consumption than advertising. (10) Adolescents may be especially sensitive to price because they often have little money of their own, and those who drink heavily may not yet be addicted or may not be so addicted that they become less responsive to price changes.

In 2003, researchers looked at the relationship between local alcohol marketing rates (such as the number of bars and liquor stores located near campus) and promotions (like “all you can drink” for one price) and college binge drinking rates. They found that the greater the number of these establishments, the more college students drank. Promotions like special prices at certain times and sales of high volume containers (kegs and “party balls”) were also strongly linked with increased heavy drinking.

Interestingly, this study also found that binge drinking rates were elevated when college students were more likely to be “carded” or “proofed” to verify their ages to drink at bars. While the researchers thought this might be explained by greater enforcement efforts in areas with bigger problems, it could support the argument of those who believe that a higher drinking age is likely to increase bingeing. In this scenario, under-age drinkers, fearful of being caught, buy alcohol in larger quantities when they can get it and drink it faster so they can hide the evidence.

Studies using longitudinal surveys have not established that advertising is a causal factor for youth drinking. A 2010 review of longitudinal studies found that many “ignore statistical problems and solutions that are well-known in econometrics, including issues of specification bias, measurement error, endogeneity (a variable is said to be endogenous when there is a correlation between the variable and the error term) and sample selection.” (11)

This review also found that advertising studies are not particularly robust; in other words,when these studies are fully examined there are negative, null and positive results for advertising variables, sometimes in the same study. For example, among the 63 estimates of the effects of advertising and promotion on adolescent drinking reviewed, only 21 of 63 estimates (33 percent) are statistically significant. For drinking onset, only 5 of 14 estimates for mass media are statistically significant, but 4 of these are from the same study. Furthermore, the review found little replication across studies, making assessment more difficult.

A 2003 working paper for the National Bureau of Economic Research concluded that eliminating alcohol advertising completely would reduce the proportion of adolescents who drink each month from 25 percent  to 21 percent. More significantly, the study claimed that a total ad ban would reduce the population of teen binge drinkers from 12 percent of adolescents to 5 percent. (12) But the very same year, a World Health Organization sponsored study concluded that advertising bans and other marketing regulations were among the least effective policy strategies. (13)

While a total ban on alcohol advertising might be effective in cutting teen drinking to some extent, it would be unlikely to pass constitutional muster. Furthermore, the research evidence is so mixed that it is hard to know whether a total ban would have a significant impact. Consider, for example, the persistence of illegal drug use among teens, in the complete absence of advertising. Relying on a reduction in alcohol advertising to reduce drinking will have little or no impact.

The Alcohol Policy Index (API) data has been used in a number of recent studies to examine relationships between alcohol control policies and adolescent alcohol use. API is a composite indicator used to compare and rate the alcohol policies of different countries and generate scores (based on policies from five regulatory domains). Examining the relationship between policy score and per capita alcohol consumption, analysis revealed a strong negative correlation between score and consumption. (14) The authors believe the study revealed a clear inverse relationship between policy strength and alcohol consumption.

A 2007 study examined this relationship and concluded that “more comprehensive and stringent alcohol control policies, particularly policies affecting alcohol availability and marketing, are associated with lower prevalence and frequency of adolescent alcohol consumption and age of first alcohol use;” but it also found that most of the relationships between API, alcohol availability and advertising control and drinking prevalence rates were attenuated and no longer statistically significant when controlling for per capita consumption in regression analyses. This suggested, the study said  “that alcohol use in the general population may confound or mediate observed relationships between alcohol control policies and youth alcohol consumption.”(15) It is difficult to change youth drinking without affecting the consumption of legal aged adults.

Conclusion
In a forthcoming paper for the Journal of Economic Surveys, Jon Nelson, Professor Emiritus of Economics at Penn State, warns that “a great deal of work remains to be done if this literature is to serve as a basis for sound public policy.” A lack of methodological uniformity and statistical robustness has been compounded by a "dissemination bias," whereby fundamentally weak associations are repeatedly cited until they appear to gain the force of consensus. This is particularly true of research produced by CAMY, says Nelson. (16)

Those who sell ads for alcohol and those who oppose alcohol advertising each have their own agendas and believe advertising to be powerful force, albeit for very different reasons and to different ends. Research shows the effects of advertising to be far more complex than either side admits and ignoring this critical fact doesn't help set effective alcohol policy aimed at reducing underage drinking.

Measures aimed at increasing alcohol prices could significantly affect youth drinking, but they would also be unpopular with many adults and could even have a negative impact on cardiovascular disease if they cut moderate drinking rates among adults. Moderate drinking by adults has been found to reduce heart and blood vessel diseases, which are major killers in the older age groups. A more effective and less onerous way of cutting college drinking could be to limit the number of alcohol sellers near campuses, ban price promotions like "all you can drink," and reduce volume discounts allowed on items like kegs.

Counter-advertising could also be effective, particularly ads that illustrate the potential negative consequences of drinking that are of particular concern to young people rather than ads that rely on fear-mongering or warn of long-term dangers.

Again, counter-intuitively, adults seem to be more affected by anti-alcohol and other anti-drug public service announcements than adolescents. This may result from the fact that many such ads use fear as a tactic, which is far more effective with adults than youth.

Essentially, this research shows that young people "don't think it will happen to me." Like adults, they are genuinely shocked and dismayed by horrifying images of auto accidents, overdose victims and smokers with artificial voice boxes. Like adults, they are able to recall ads that feature such images. Yet none of this seems to influence their behavior. While most adults have learned from painful experience that they are not immune to negative consequences, youth tend to feel invulnerable (and may need to maintain these feelings in order to accomplish their developmental goals of leaving their family nest).

As a result, every "Scared Straight" type program ever evaluated has been a failure. These programs are huge hits with parents - and teens tend to say that they were profoundly affected by them - but the data consistently shows either no effect or the unintended consequence of a slight increase in the behavior the intervention was attempting to prevent.

This doesn't mean that teens are impervious to anti-drinking messages: it's just that the ads that affect them aren't the dramatic ones that gain public attention and win advertising awards. In the tobacco area, for example, studies have found that ads which focus on the immediate negative effects of smoking, like bad breath or impaired athletic performance, are more effective with youth than those which warn of the likelihood of lung cancer decades hence. Similar attention to how alcohol could make one lose control in the social situations most important to teens might be a promising approach.

The belief that advertising induces teenagers to drink underscores attempts by lawmakers and anti-alcohol groups to impose stricter regulation and restrictions on how alcohol is marketed. Though all these efforts have the laudable goal of trying to reduce alcohol abuse among young people, the science upon which they are based has yet to provide clear evidence that such restrictions – or for that matter, advertising campaigns to reduce drinking - will have much, if any, effect.

1. Johnston LD, O’Malley PM, Bachman JG, Schulenberg JE. Monitoring the Future national survey results on drug use, 1975-2007: Volume I, Secondary school students (NIH publication No. 08-6418A). See Table 2-2..

2. 2007 National Survey on Drug Use and Health: national findings (NSDUH Series H-34, DHHS publication No. SMA 08-4343), 2008: Figures 3.1, 3.4, and 3.6.

3. Lipsitz, A.; Brake, G.; Vincent, E.J.; and Winters, M. “Another round for the brewers: Television ads and children's alcohol expectancies.” J Appl Soc Psychol 23(6):439-450, 1993..

4. Hastings G, Anderson S, Cooke E, Gordon R.“Alcohol marketing and young people's drinking: a review of the research,” Journal of Public Health Policy, 2005 Sep;26(3):292-5.

5. GfK Roper Youth Report, 2009.

6. McClure AC, Dal Cin S, Gibson J, Sargent JD. “Ownership of alcohol-branded merchandise and initiation of teen drinking,” Archives of Pediatrice Adolescent Medicine, 2006 Jan;160(1):18-24.

7. Hurtz S.Q, Hnriksen L, Wang Y, Feighery, E.C., Fortmann, S.P “The Relationship between Exposure to Alcohol Advertising in stores, owning alcohol promotional items, and adolescent alcohol use,” Journal of Alcohol and Alcoholism, 2007, Mar-Apr; 42(2): 143-9.

8. Collins RL, Ellickson PL, McCaffrey D, Hambarsoomians K.“Early Adolescent Exposure to Alcohol Advertising and its Relationship to Underage Drinking,” Journal of Adolescent Health, [Epub April 12 2007]

9. Nicholson M, Hoye R. “Reducing adolescents' exposure to alcohol advertising and promotion during televised sports.” JAMA. 2009;301(14):1479-1482

10. Hastings G, Anderson S, Cooke E, Gordon R. “Alcohol marketing and young people's drinking: a review of the research,” Journal of Public Health Policy, 2005 Sep;26(3):292-5.

11. Nelson, J.P. “What is Learned from Longitudinal Studies of Advertising and Youth Drinking a Smoking? A Critical Assessment,” Int J. Environ. Res. Public Health 2010, 7(3), 870-926; doi:10.3390/ijerph7030870

12. H. Saffer and D. Dave. "Alcohol Advertising and Alcohol Consumption by Adolescents," NBER Working Paper No. 9676, May 2003.

13. Cited in Nelson: Babor et al., Alcohol: No Ordinary Commodity—Research and Public Policy; Oxford University Press: Oxford, UK, 2003.2003, p. 272.

14. Brand DA, Saisana M, Rynn LA, Pennoni F, Lowenfels AB. “Comparative analysis of alcohol control policies in 30 countries,” Department of Medicine, New York Medical College, Valhalla, New York, United States of America.

15. Paschall MJ, Grube JW, Kypri K, Prevention Research Center, Berkeley, CA. “Alcohol control policies and alcohol consumption by youth: a multi-national study.” PLoS Med. 2007 Apr;4(4):e151. PLoS Medical, 2007 Apr;4(4):e151.

16. Nelson, J.P. “Alcohol Marketing, Adolescent Drinking, and Publication Bias in Longitudinal Studies: A Critical Appraisal Using Meta-Analysis,” Journal of Economic Surveys (forthcoming).