In the paper “Movie Smoking and Youth Initiation: Parsing Smoking Imagery and Other Adult Content” [1], Farrelly and colleagues reported results from a longitudinal 4-wave survey of New York adolescents in which movie smoking exposure was estimated along with exposure to “adult” movie content from 71 movies. Consistent with the large literature that led the US National Cancer Institute [29] and US Surgeon General to conclude that exposure to smoking in movies caused youth to smoke [22], they found that exposure to smoking in movies predicted smoking initiation and more favorable attitudes towards trying smoking. Moreover, they found a dose-response relationship: The higher the exposure, the more likely youth were to smoke and the more positive their attitudes were to smoking. Finally, they estimated that youth smoking would have been twenty percent lower had the youths not been exposed to movie smoking, which is within the range of estimates from a number of previously reported longitudinal studies from around the world [2-5]. Surprisingly, Farrelly et al. concluded that the effect they observed could not be attributed to movie smoking, advancing three arguments to support their conclusion.

First, Farrelly et al. argued that the association between exposure to movie smoking and youth smoking could be spurious because high-risk-for-smoking adolescents are attracted to movies with adult themes. They suggest that this self-selection confounds the effect of movie smoking on behavior, which would exonerate any aspect of movie exposure—smoking or otherwise—as a cause of youth smoking. This is a reasonable hypothesis but one that has been posed and rejected in nearly 20 studies published over the past 10 years [2-20]. Taken together, these studies have controlled for a wide range of relevant covariates (demographics, parental smoking behavior, parenting and parental restrictions, adolescent personality characteristics, extracurricular activities, even weekly spending money) with many confounders—e.g., sensation seeking, self-esteem, impulsivity—entered explicitly to address self-selection. Regardless of the study sample and the mix of control variables, the published studies have consistently reported a statistically significant doubling of smoking risk given higher exposure to movie smoking, as shown by a meta-analysis in the 2012 Surgeon General’s report on adolescent smoking [22]. The consistency of the findings across diverse populations, covariate controls, and study designs provides strong evidence against the self-selection hypothesis.

In addition, after Farrelly et. al. submitted their manuscript, Sargent et. al. [21] published a study that directly tested whether smoking in adult-rated (R) movies was associated with more impact than smoking in youth-rated (PG-13) films. Sargent et. al. disaggregated total movie smoking exposure into smoking from G/PG movies, smoking from PG-13 movies, and smoking from R-rated movies to capture the additional impact of movies with more of the adult behaviors (profanity, violence, illicit drug use, and sex) considered by the Motion Picture Association when it rates them. In this approach, the R rating serves as a red flag to draw deviant adolescents to movies that address adult themes. If the concerns of self-selection are valid, R-rated smoking should be a much stronger predictor of smoking compared with PG-13 smoking. But instead of smoking in R rated movies being a stronger predictor of youth smoking (as Farrelly et. al hypothesize), the dose-response relationship for exposure to smoking in R-rated movies was indistinguishable from smoking in PG-13 movies. These findings also provide further evidence to reject a self-selection hypothesis.

Second, Farrelly et. al. speculated that other types of exposures such as “adult content” are the active ingredient responsible for the observed impact of movie smoking exposure on adolescents’ behavior. Critically, they do not explain how the presence of other adult behaviors (but not smoking) might affect adolescent smoking, and without a plausible explanation their argument falls short. This specificity issue was directly and empirically addressed in a recently published epidemiological study by Morgenstern et al. [4], which performed analyses with data from a large European sample to test for the specificity of the movie smoking-youth smoking association. These investigators found a significant unique association for movie smoking exposure (based on their exposure to smoking in the 50 movies each youth was asked about and had seen)—net of exposure to movies generally (based on the number of movies they had seen out of the 50 they were asked about). The movie exposure – adolescent smoking association was found despite a correlation of 0.82 between these two variables. This is an area for further research, but the finding of a specificity effect in the study by Morgenstern creates a substantial problem for the argument that the impact of movie exposure is not attributable to the smoking portrayed in movies (as opposed to other ingredients they might have in them).

Third, Farrelly et.al suggested that there is such a high correlation (0.99) between the presence of smoking and adult behaviors that it is impossible to disaggregate the effects of smoking from the effects of seeing other behaviors. This correlation is much higher than we have found in our studies. (We find the range of correlations between screen smoking, sex and alcohol among movies to be 0.10-0.43 in movies released since 2006 and the correlation between exposure to movie depictions of these same risk behaviors among adolescents to be 0.76-0.84). The almost perfect correlation Farrelly et. al. report may reflect problems with the crude content coding they used (a point they recognize) or poor reliability and correlated errors (content coding for their study was not reliability tested). In addition, the fact that Sargent et al. [21] found essentially the same dose-response relationships for the effects of smoking in R-rated and PG-13-rated films provides some additional evidence that the effects of smoking in movies on smoking behavior is independent of the effects of exposure to other adult behaviors.

As mentioned in the introduction of Farrelly et. al., several randomized experimental studies have manipulated the dose of movie smoking and shown that it enhances attitudes toward smoking [23-27], and other studies have shown that exposure to movie alcohol use prompts actual drinking behavior [28]. Experimental studies provide strong evidence for judging the causality of movie effects, but Farrelly et. al. dismiss such research as having less relevance to behavioral transitions. However, it has been known for years that the attitudes and cognitions targeted in the experiments predict transitions to smoking. Moreover, because these are controlled experiments with random assignment, dismissing the results as having no bearing on arguments about causality is not logically plausible, particularly since the results are largely consistent with the epidemiology.

In sum, the issues that Farrelly et. al. raise are not new and have been addressed herein and in prior published studies. Their longitudinal analysis of the effects of exposure to smoking in movies is consistent with the conclusions by the National Cancer Institute [29] and the U.S. Surgeon General [22]: Based on the combined weight of evidence from well-controlled observational studies and randomized experiments, movie smoking is one cause of adolescent smoking. Public health approaches to limiting movie smoking are warranted.

James Sargent
Thomas Wills
Mike Stoolmiller

REFERENCES

1. Farrelly, M.C., et al., Movie smoking and youth initiation: parsing smoking imagery and other adult content. PLoS One, 2012. 7(12): p. e51935.
2. Dalton, M.A., et al., Effect of viewing smoking in movies on adolescent smoking initiation: a cohort study. Lancet, 2003. 362(9380): p. 281-5.
3. Hanewinkel, R. and J.D. Sargent, Exposure to Smoking in Internationally Distributed American Movies and Youth Smoking in Germany: A Cross-cultural Cohort Study. Pediatrics, 2008. 121(1): p. e108-117.
4. Morgenstern, M., et al., Smoking in movies and adolescent smoking: cross-cultural study in six European countries. Thorax, 2011. 66(10): p. 875-83.
5. Titus-Ernstoff, L., et al., Longitudinal Study of Viewing Smoking in Movies and Initiation of Smoking by Children. Pediatrics, 2008. 121(1): p. 15-21.
6. Dal Cin, S., M. Stoolmiller, and J.D. Sargent, When movies matter: exposure to smoking in movies and changes in smoking behavior. J Health Commun, 2012. 17(1): p. 76-89.
7. Dalton, M., et al., Early exposure to movie smoking predicts established smoking by older teens and young adult. Pediatrics, 2009. 123: p. e551-e558.
8. Hanewinkel, R. and J.D. Sargent, Exposure to Smoking in Popular Contemporary Movies and Youth Smoking in Germany. American Journal of Preventive Medicine, 2007. 32(6): p. 466-473.
9. Hunt, K., et al., Exposure to smoking in films and own smoking among Scottish adolescents: a cross-sectional study. Thorax, 2011. 66(10): p. 866-74.
10. Jackson, C., J.D. Brown, and K.L. L'Engle, R-rated movies, bedroom televisions, and initiation of smoking by white and black adolescents. Arch Pediatr Adolesc Med, 2007. 161(3): p. 260-8.
11. Sargent, J.D., et al., Exposure to Movie Smoking: Its Relation to Smoking Initiation Among US Adolescents. Pediatrics, 2005. 116(5): p. 1183-1191.
12. Sargent, J.D., et al., Movie smoking and urge to smoke among adult smokers. Nicotine Tob Res, 2009. 11(9): p. 1042-1046.
13. Sargent, J.D., et al., Exposure to smoking depictions in movies: its association with established adolescent smoking. Arch Pediatr Adolesc Med, 2007. 161(9): p. 849-56.
14. Song, A.V., et al., Smoking in movies and increased smoking among young adults. Am J Prev Med, 2007. 33(5): p. 396-403.
15. Tanski, S.E., et al., Movie character smoking and adolescent smoking: who matters more, good guys or bad guys? Pediatrics, 2009. 124(1): p. 135-43.
16. Tanski, S.E., et al., Moderation of the Association between Media Exposure and Youth Smoking Onset: Race/Ethnicity, and Parent Smoking. Prev Sci, 2012. 13(1): p. 55-63.
17. Thrasher, J.F., et al., Does film smoking promote youth smoking in middle-income countries?: A longitudinal study among Mexican adolescents. Cancer Epidemiol Biomarkers Prev, 2009. 18(12): p. 3444-50.
18. Waylen, A.E., et al., Cross-sectional association between smoking depictions in films and adolescent tobacco use nested in a British cohort study. Thorax, 2011. 66(10): p. 856-61.
19. Wills, T., et al., Movie smoking exposure and smoking onset: A longitudinal study of mediation processes in a representative sample of U.S. adolescents. Psychol Addict Behav, 2008. 22(2): p. 269-277.
20. Wills, T.A., et al., Good self-control moderates the effect of mass media on adolescent tobacco and alcohol use: tests with studies of children and adolescents. Health Psychol, 2010. 29(5): p. 539-49.
21. Sargent, J.D., S. Tanski, and M. Stoolmiller, Influence of motion picture rating on adolescent response to movie smoking. Pediatrics, 2012. 130(2): p. 228-36.
22. U.S. Department of Health and Human Services. Preventing Tobacco Use Among Youth and Young Adults: A Report of the Surgeon General. Atlanta, GA: U.S. Department of Health and Human Services, Centers for Disease Control and Prevention, National Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health, 2012.
23. Dal Cin, S., et al., Smoking in movies, implicit associations of smoking with the self, and intentions to smoke. Psychol Sci, 2007. 18(7): p. 559-63.
24. Lochbuehler, K., R.C. Engels, and R.H. Scholte, Influence of smoking cues in movies on craving among smokers. Addiction, 2010.
25. Pechmann, C. and C.F. Shih, Smoking scenes in movies and antismoking advertisements before movies: effects on youth. J Marketing, 1999. 63(3): p. 1-13.
26. Shmueli, D., J.J. Prochaska, and S.A. Glantz, Effect of smoking scenes in films on immediate smoking: a randomized controlled study. Am J Prev Med, 2010. 38(4): p. 351-8.
27. Shadel, W.G., et al., Motives for smoking in movies affect future smoking risk in middle school students: An experimental investigation. Drug Alcohol Depend, 2011.
28. Engels, R.C.M.E., et al., Alcohol Portrayal on Television Affects Actual Drinking Behaviour. Alcohol and Alcoholism, 2009. 44(3): p. 244-249.
29. National Cancer Institute. The Role of the Media in Promoting and Reducing Tobacco Use. Tobacco Control Monograph No. 19. Bethesda, MD: U.S. Department of Health and Human Services, National Institutes of Health, National Cancer Institute. NIH Pub. No. 07-6242, June 2008.