Advertisement
Essay

Essay Essays are opinion pieces on a topic of broad interest to a general medical audience.

See all article types »

Four Arguments against the Adult-Rating of Movies with Smoking Scenes

  • Simon Chapman mail,

    simon.chapman@sydney.edu.au

    Affiliation: University of Sydney, Sydney, Australia

    X
  • Matthew C. Farrelly

    Affiliation: RTI International, North Carolina, United States of America

    X
  • Published: August 23, 2011
  • DOI: 10.1371/journal.pmed.1001078

Reader Comments (8)

Post a new comment on this article

Unaddressed Methodological Issues in Observational Studies on Smoking in the Movies

Posted by MatthewFarrelly on 08 Sep 2011 at 13:28 GMT

Our primary concern is that the influence of smoking in the movies on youth smoking is overstated in the existing observational studies due to unaddressed methodological issues. As a result, one is led to believe that if we eliminate youths’ exposure to smoking in the movies that youth smoking would be greatly reduced. If it is true that the influence of smoking in the movies is overstated, it puts the tobacco control community on the wrong trail and risks consuming scarce resources and political capital that should be focused elsewhere.

There are two related methodological issues that have yet to be addressed in the existing literature on smoking in the movies. The first has to do with what Manski (1993) refers to as context effects. Movies that contain a significant number of smoking scenes have other attributes that should be accounted to control for their potential influence on youth smoking initiation (i.e., construct confound). Examples of content that should be accounted for, as they may be particularly attractive to sensation seeking adolescents, include suspense and violence. Although it is not clear to what extent this other content may cause smoking, this empirical question can be easily addressed by coding for other features in movies seen by youth. If there is sufficient variation in the movies that youth choose—movies with a significant amount of smoking scenes and not a lot of other risk behaviors and adult themes; movies with the latter and not the former; and movies with neither—it is possible to better isolate the influence of seeing smoking in the movies on youth smoking. However, in a sample of movies seen by adolescents in New York State, the correlation between seeing smoking in the movies and significant amounts of other adult content was so high (0.99), it may be difficult to parse smoking from other content (Farrelly et al, 2011). Other observational studies with data on a larger database of movies may have a greater opportunity to do this. However, if this correlation remains high, this remains a potential threat to validity for both observational studies and experimental studies that remove smoking scenes from movies and leave other content unchanged.

A more vexing and fundamental issue with the existing measures of adolescents’ exposure to smoking in the movies is that this construct is not an exogenous variable. Adolescents’ underlying preferences (e.g., for risk taking) represents a third variable that jointly influences their selection of movies and their decision to smoke. Failing to disentangle how smoking in movies may influence youth smoking from how adolescents’ selection of movies indicates their interest in risk taking, including smoking, can lead to an overestimate of the influence of smoking in the movies on youth smoking initiation.

This concern about the existing estimates of the magnitude of the influence that exposure to smoking in the movies has on smoking prevalence/initiation may best be explained by an analogous problem of understanding peer influences on smoking. Cross-sectional and longitudinal studies find that having one or more close friends who smoke is highly correlated with an adolescent’s own choice to smoke (Kobus, 2003). However, it has long been recognized that this strong correlation does not merely reflect the influence that peers have on their friends smoking. It reflects a combination of the influence that peers have on an adolescent’s choice to smoke and an expression of that adolescent’s preference for peers who engage in risky behaviors, including smoking (Kandel, 1978, Manski, 1993). In fact, the tobacco industry likes to trot out these types of statistics to say that parents and peers are the main culprit for youth smoking, not them. It was also the justification for Philip Morris’ failed “youth prevention” campaign, “Think. Don’t Smoke.”

Adolescents’ underlying preference for risk taking jointly influences decisions to smoke and affiliation with friends who engage in risky behaviors. Unless we can purge the latter phenomenon from the relationship between choosing to smoke and having friends who smoke, the influence of peers on the choice to smoke is overstated. Some researchers who conduct social network analyses and/or employ instrumental variable models have taken up the difficult challenge of trying to disentangle these two effects with some success (Ennett and Bauman, 1994, Hoffman et al, 2007, Norton, Lindrooth and Ennett, 2007).

As adolescents age they have increased freedom to express their own individuality, including the friends they choose, the clothes they wear, the music they listen to, the movies they watch, and the risk behaviors in which they engage. An adolescent who is contemplating smoking may seek movies that portray risk-taking, including smoking. Adolescents are voting with their feet (or eyeballs) in their choice of movies and these choices both reveal their underlying appetite for risky behaviors and other adult themes and expose them to influences that can shape their decisions to engage in risk behaviors, including smoking. As a result, we are concerned that the existing studies overstate the influence of smoking in the movies on adolescent smoking because they do not adequately parse out these two issues reflected in the single measure of exposure to smoking in the movies. Glantz et al and Sargent et al argue vigorously that the existing studies account for this by controlling for potential confounding with covariates such as sensation seeking and parental controls over what movies adolescents are allowed to see. These controls are certainly an important start, but they do not directly disentangle the process of movie selection and the influence that smoking in the movies has on youth smoking.

To solve this issue with observational studies is as challenging as it has been to parse the influence of peer smoking from selection of peers who smoke. In this case, it requires instrumental variable models where the identifying variables explain the choice of movies with smoking but are not highly correlated with the choice to smoke (Greene, 1999). Other solutions include experimental studies, of which there are just a couple to date (Shmueli, Prochaska and Glantz, 2010; Wagner et al, 2011) and/or interventions that may alter what movies youth are permitted to view. Although experiments help demonstrate the effect of smoking scenes on youth smoking, it may be challenging to translate how identified effects translate to population effects.

Another way to examine the reasonableness of the existing estimates of attributable risk of smoking in the movies on smoking is to examine recent trends in youth smoking and the quantity of smoking scenes. Trends in youth smoking and the quantity of movie scenes with smoking support our suggestion that the reported magnitude of the attributable risk of smoking in the movies is overstated. A recent study from CDC shows that the amount of smoking in the movies has declined markedly in recent years (CDC, 2011). At the same time, youth smoking prevalence has reached a plateau during a period when many other environmental influences have been trending in a positive direction—higher state and federal taxes and more and more smoke free air laws. However, given the 72% decline in the amount of smoking in movies rated G, PG, or PG-13 from 2005 to 2010 and the potential attributable fraction for smoking initiation for being exposed to smoking in the movies is estimated at 44% (Millet and Glantz, 2010), one would expect a substantial decline in youth smoking, all else constant. During this same period, the prevalence of smoking among 8th and 10th graders from Monitoring the Future declined by 14%. By applying the attributable fraction of 44% to the 72% decline in exposure to smoking in the movies, one would expect to see a 32% decline in initiation over this time period. Although changes in initiation do not directly translate to changes in smoking prevalence, it does give a sense of the magnitude of change we might expect. Of course, other factors changed during this time period that help explain the 14% decline in smoking. For example, the price of cigarettes increased by 57% over the same time period. Based on the effect of cigarette prices on youth smoking prevalence, such an increase would lead to a decline in smoking of 13% to 32% (Carpenter and Cook, 2008; Nonnemaker and Farrelly, 2011)—possibly explaining the entire decline over this time period.

Does this analysis indicate that smoking in the movies does not influence youth smoking? No, it does not, but it does suggest that the magnitude of its influence may be overstated. Given the reduction of funding for tobacco control in the U.S., the flattening of youth and adult smoking prevalence, and the need to counter tobacco industry marketing activities through the Family Smoking Prevention and Tobacco Control Act, the tobacco control community must use the strongest possible science when prioritizing public health interventions. Our concerns regarding the science of smoking in the movies are valid and should not be considered demeaning to the researchers who have published in this area or the public health organizations that have called for reductions in smoking scenes in youth-rated movies.

Matthew C. Farrelly, RTI International
James Nonnemaker, RTI International
Simon Chapman, University of Sydney

References

Carpenter, Christopher and Philip J Cook (2008), “Cigarette Taxes and Youth Smoking: New Evidence from National, State, and Local Youth Risk Behaviour Surveys”, Journal of Health Economics, 27:287-299.

CDC. (2011). Smoking in Top-Grossing Movies—United States, 2010. Morbidity and Mortality Weekly Report, 60(27);909-913.

Ennett, S.T., Bauman, K. E. (1994). The contribution of influence and selection to adolescent peer group homogeneity: The case of adolescent cigarette smoking.
Journal of Personality and Social Psychology, Vol 67(4), 653-663.

Farrelly, M. C., Kamyab, K., Nonnemaker, J. M. and Crankshaw, E., (2011). Movie Smoking and Youth Initiation: Parsing Smoking Imagery and Other Adult Content. Available at SSRN: http://ssrn.com/abstract=...

Greene, William H. (1999) Econometric Analysis. Prentice Hall.

Hoffman, B.R., Monge, P.R., Chou C., Valente, T.W. (2007). Perceived peer influence and peer selection on adolescent smoking, Addictive Behaviors, Vol 32(8), 1546-1554.

Kandel, D. B. (1978) Homophily, selection, and socialization in adolescent friendships. American Journal of Sociology, 84, 427–436.

Kobus, K. (2003) Peers and adolescent smoking. Addiction, 98 (Suppl 1), 37–55.

Manski, C. F. (1993) Identification of endogenous social effects: the reflection problem, Review of Economic Studies 60 (3) (1993), pp. 531–542.

Millett C, Glantz SA. Assigning an '18' rating to movies with tobacco imagery is essential to reduce youth smoking. Thorax. 2010 May; 65(5):377-8.

Nonnemaker, J.M., & Farrelly, M.C. (2011). Smoking initiation among youth: The role of cigarette excise taxes and prices by race/ethnicity and. Journal of Health Economics, 30 (3):560-567.

Norton, E. C., Lindrooth, R. C., and Ennett, S. T. (1998). Controlling for the
Endogeneity of Peer Substance Use on Adolescent Alcohol and Tobacco Use. Health
Economics, 7: 439-453.

Shmueli D, Prochaska JJ, Glantz SA (2010) Effect of smoking scenes in films on immediate smoking: a randomized controlled study. Am J Prev Med 38: 351-358.

Wagner DD, Dal Cin S, Sargent JD, Kelley WM, Heatherton TF (2011) Spontaneous action representation in smokers when watching movie characters smoke. J Neurosci 31: 894-898.

No competing interests declared.

RE: Unaddressed Methodological Issues in Observational Studies on Smoking in the Movies

MatthewFarrelly replied to MatthewFarrelly on 08 Sep 2011 at 18:42 GMT

One point of clarification. We had intended to say "However, if this correlation remains high, this remains a potential threat to validity for both observational studies and experimental studies that do not separate smoking scenes from other adult content within the same movie" at the end of the second paragraph.

Matthew C. Farrelly, RTI International
James Nonnemaker, RTI International
Simon Chapman, University of Sydney

No competing interests declared.

RE: Unaddressed Methodological Issues in Observational Studies on Smoking in the Movies

jpk1 replied to MatthewFarrelly on 09 Sep 2011 at 02:06 GMT

You write "Given the reduction of funding for tobacco control in the U.S. . . . the tobacco control community must use the strongest possible science when prioritizing public health interventions".

I absolutely agree. However I must report that the vast majority of tobacco prevention funded in the U.S. has far less scientific support than smokefree movies. For example, school based "resist pressure, don't smoke" programs; youth access programs; media campaigns using themes not measurably effective; programs limited to youth; the list goes on and on. For that matter, the narrow product ingredients focus of the "Family Smoking Prevention and Tobacco Control Act" has little scientific backing.

So in questioning how much smoking results from on-screen smoking, a little context is in order. We already have more evidence that smokefree movies will be effective and cost effective than we have for the vast majority of funded programs. Do we have the level of evidence we have for smokefree policies? No; and until we try it, we can't get to that level of evidence.

No competing interests declared.