BACKGROUND. Several studies have linked seeing smoking in movies with adolescent smoking, but none have determined how much movie smoking adolescents see.
OBJECTIVE. Our aim was to determine exposure to movie smoking in a representative sample of young US adolescents.
METHODS. We surveyed 6522 nationally representative US adolescents aged 10–14 years. We content analyzed 534 contemporary box-office hits for movie smoking. Each movie was assigned to a random subsample of adolescents (mean: 613) who were asked whether they had seen the movie. Using survey weights, we estimated the total number of US adolescents who had seen each movie and then multiplied by the number of smoking depictions in each movie to obtain gross smoking impressions seen by adolescents.
RESULTS. The 534 movies were mainly rated PG-13 (41%) and R (40%), and 74% contained smoking (3830 total smoking occurrences). On average, each movie was seen by 25% of the adolescents surveyed. Viewership was higher with increased age and lower for R-rated movies. Overall, these movies delivered 13.9 billion gross smoking impressions, an average of 665 to each US adolescent aged 10–14 years. Although this sample's R-rated movies contained 60% of smoking occurrences, they delivered only 39% of smoking impressions because of lower adolescent viewership. Thirty popular movies each delivered ≥100 million gross smoking impressions. Thirty actors each delivered >50 million smoking impressions, such that just 1.5% of actors delivered one quarter of all character smoking to the adolescent sample.
CONCLUSIONS. Popular movies deliver billions of smoking images and character smoking depictions to young US adolescents. Removing smoking from youth-rated films would substantially reduce exposure from new box-office hits. Furthermore, the popular actors who frequently smoke in movies could have a major impact on adolescent movie smoking exposure by choosing not to portray characters who smoke.
Reports of an association between seeing smoking in movies and adolescent smoking1–7 have prompted greater scrutiny of smoking depicted in movies. Movie smoking has been quantified through content analyses, in which a specified sample frame of movies is assessed according to a set of predefined criteria. The published content analyses agree on multiple points: smoking is depicted in most movies8–14; movie smoking increases as the Motion Picture Association of America (MPAA) rating increases from G to R8,13,14; movie smoking rarely is associated with negative health outcomes8,9,11,14; and movie characters who smoke are more affluent than the typical US smoker.8,11 Studies that report the prevalence of character smoking concur that 20% to 25% of characters in popular contemporary movies smoke12,15,16 and that smoking rates are higher among male characters. One area that has been less thoroughly assessed is how much of the smoking in movies actually reaches adolescents. Movie smoking seen by a relatively small proportion of adolescents would be unlikely to have a large population effect on smoking. For movies popular among adolescents, however, a single smoking depiction may reach a large audience. In this context, as it is in the advertising arena, an assessment of the number of smoking depictions delivered to a defined audience is considered a key metric of the success of the campaign.
One 2004 non–peer-reviewed publication estimated the number of smoking impressions delivered to US adolescents.17 By using an Internet parental review/screening service (www.screenit.com) for some 800 box-office releases to estimate smoking content and Nielsen figures to calculate the youth audience, they determined that the selected movies delivered some 8.3 billion smoking scenes to children and teens, or ∼1350 per capita. Despite a higher concentration of smoking in R-rated movies, youth-rated movies delivered approximately half of the smoking scenes. The limitations of the study notwithstanding, this report underlines the potential impact of movie smoking because of the extensive reach into the adolescent population. A recent article looking at images of smoking included in 31 television trailers for movies found that even these brief advertisements resulted in substantial exposure, with 270 million smoking impressions delivered to youth in a single year.18 Our aim in this study is to assess exposure to movie smoking in US adolescents aged 10 to 14 years using more rigorous quantitative methods that allow for an evaluation of smoking within a popular contemporary sample of films and aggregated at the levels of movie and leading actors.
Between June and October, 2003, we conducted a random-digit dial telephone survey of 6522 US adolescents aged 10 to 14 years. We have previously reported details of the survey, including the survey completion rate (66%).1 Briefly, the telephone surveys were conducted by trained interviewers using a computer-assisted telephone interview system, and interviewers were trained to administer the survey in English or Spanish. We obtained parental consent and adolescent assent before interviewing each respondent. To protect confidentiality, adolescents indicated their answers to sensitive questions by pressing numbers on the telephone, rather than speaking aloud. All aspects of the survey were approved by the institutional review boards at Dartmouth Medical School and Westat. The sample was broadly representative of the US adolescent population (as assessed by the US Census) with respect to age, gender, household income, and census region (Appendix 1). Compared with the 2000 US Census, our unweighted sample had higher percentages of Hispanics and slightly lower percentages of blacks; we adjusted for these differences using a poststratification weighting procedure.
One aim of the survey was to assess the proportion of adolescents who had seen each movie of a sample of popular contemporary movies. We selected the top 100 US box-office hits per year for each of the 5 years preceding the survey (1998–2002; N = 500) and 34 movies that earned at least $15 million in gross US box-office revenues during the first 4 months of 2003. The computer-assisted telephone interview survey was programmed to randomly select 50 movie titles from the larger pool for each adolescent interview, such that adolescents were randomly assigned to movie titles, and a representative subsample of adolescents (mean number surveyed: 613; SD: 27) was queried on each title. Movie selection was stratified by the MPAA rating so that the distribution of movies in each list reflected the distribution of the full sample of movies (19% G/PG, 41% PG-13, and 40% R). Respondents were asked whether they had ever seen each movie title on their unique list. To assess the possibility of false-positive responses, we asked all of the adolescents whether they had seen a sham movie title, Handsome Jack, and <2% reported having seen it. We have shown previously that adolescents reliably remember movies they have seen 1 to 2 years before a survey.2
Trained coders counted the number of smoking occurrences in each of the 534 movies using previously validated methods.8 A smoking occurrence was counted whenever a major or minor character handled or used tobacco in a scene (defined as a “smoking episode”) or when background smoking was present (defined as a smoking “incident”). The smoking occurrence was used as the unit of movie exposure in characterizing adolescent movie smoking exposure by movie. To determine exposure to movie smoking according to actor, we restricted the analysis to the smoking episodes by major and minor characters. A major character was defined by the movie coders as someone “central to the development of the plot”; coders identified a median of 6 (interquartile range: 5–8) major characters in each movie. Smoking episodes were counted irrespective of the scene's duration or how many times the tobacco product appeared during the scene. We used 2 movie coders and double-coded 10% of the movies. As a measure of interrater agreement, the Pearson correlation coefficient for the number of smoking episodes (character smoking) contained in each double-coded movie was 0.99 and for smoking incidents (background smoking or signage) was 0.86.
In assessing advertising campaigns, it is important to determine how many individuals or households are exposed to a particular advertising media or message; this is what marketing researchers refer to as a determination of reach. Because adolescents were randomly assigned to a movie, we assumed that the proportion of adolescents within each subsample who had seen each movie reflected the proportion of adolescents in the US population who had seen the movie, hereafter referred to as the “movie viewership.” For each movie, we multiplied the movie viewership proportion by the number of US adolescents aged 10 to 14 years in 2002 (20.88 million) to obtain an estimate of the number of US adolescents who had seen the movie, our measure of reach. “Gross impressions” is the total number of exposures delivered by a media schedule. For movies, gross smoking impressions were determined by multiplying the estimate of the number of US adolescents who had seen the movie by the number of smoking occurrences in the movie. Per capita gross impressions of movie smoking were obtained by dividing the total number of gross smoking impressions across all of the movies in the sample by the total US population of adolescents. This measure is similar to that used to determine the success of a marketing campaign, the gross rating point, except that our measure is conservative in that it does not account for multiple viewing of 1 movie by an adolescent. In our determination of gross smoking impressions delivered by leading actors, we aggregated by actor, the movie smoking occurrences delivered by all of the major characters in ≥1 movie.
As reported previously,1 the movie sample was composed mainly of PG-13 (41%) and R-rated (40%) movies (Table 1) and contained a total of 3830 smoking occurrences. Sixty percent of all smoking occurrences appeared in R-rated movies, 36% in PG-13 movies, 4% in PG movies, and <1% in G-rated movies. Mean movie viewership was 0.25 (SD: 0.21). Figure 1 shows the distribution for movie viewership by MPAA rating for the 534 movies. The mean of movie viewership was similar for G-rated movies (0.51) and PG-rated movies (0.47) but significantly lower for PG-13 movies (0.27) and for R-rated movies (0.14; P < .001). Figure 2 illustrates the distribution of movie viewership by age of the respondent. Movie viewership increased in a linear fashion with age (P < .001 for trend), from a mean of 0.19 for 10-year-old adolescents to 0.34 for 14-year-olds.
Gross Smoking Impressions Delivered Overall
In aggregate, the 534 movies delivered some 13.9 billion impressions of smoking to this age group (Table 1). Examined on a per capita basis, an average of 665 impressions of smoking were delivered to each US adolescent aged 10 to 14 years as of the date of the survey, September 2003. Despite containing a higher number of smoking occurrences, R-rated movies accounted for a lower proportion of gross smoking impressions delivered to US adolescents (39%) than PG-13 movies (50%) because of lower viewership of R-rated movies. Figure 3 illustrates this: youth-rated movies contained ∼40% of smoking occurrences, yet they delivered ∼60% of all gross smoking impressions to young adolescents.
Gross Smoking Impressions Delivered According to Movie
Table 2 lists all of the movies that delivered ≥100 million gross smoking impressions (N = 30). The movies on the list shared 2 common characteristics: they included many smoking impressions and were popular, seen by 2 to 15 million of the estimated 21 million US 10- to 14-year-olds. Not all of the movies were intended for youths in this age range. For example, the R-rated movie Hannibal contained graphic violence scenes, yet was seen by ∼25% of adolescents in this age range. Most of the movies on this list depicted cigarette smoking. However, Lord of the Rings: The Fellowship of the Ring depicted pipe smoking, Wild Wild West depicted cigar smoking, and Atlantis: The Lost Empire depicted animated cigarette smoking.
Gross Smoking Impressions According to Major and Minor Characters
In this movie sample, we identified 7720 major or minor movie characters (Table 3). The smoking rate among minor characters was much lower than for major characters. Among the 3813 major characters, 71% were male (we did not code gender for minor characters), and the smoking rate was slightly higher among males. Movie characters accounted for 2760 (72.1%) of the 3830 smoking occurrences in the movies, delivering 9.88 billion gross impressions of character smoking to US adolescents. The majority of the character gross smoking impressions (65.2%) were delivered by male major characters.
Of the 1961 actors who played character roles in these films, 499 (25.4%) smoked in ≥1 role. Table 4 shows the number of smoking characters, total smoking episodes, and gross smoking impressions for 30 actors who delivered ≥50 million gross impressions of character smoking to US adolescents (see Appendix 2) for details on movies and names of smoking characters). These 30 actors accounted for 2.4 billion gross smoking impressions, or 24.6% of all of the gross smoking impressions delivered by characters in these movies. Of the 30 actors shown, only 3 were women, in keeping with the observation that most actor smoking is done by males. Generally, high impact actors played smoking characters in several movies released over the study period; but some actors, such as Keanu Reeves, delivered all of their smoking in a single popular movie (Hardball). Some of the characters delivering the high numbers of smoking impressions were not human: Florence Stanley delivered the highest number of smoking impressions of any actor by playing the voice for Mrs Packard, the chain-smoking communications officer on the submarine in the animated movie Atlantis: The Lost Empire, and Ian McKellan played Gandolf, a pipe-smoking wizard in The Lord of the Rings.
As mentioned above, the majority of actors did not smoke in films. Moreover, a number of actors starred in multiple films in this sample and did not smoke in any of them. Table 5 lists all of the actors who starred in ≥5 films during the study period and portrayed no smoking characters. Table 5 includes the names of many notable actors, some of whom (eg, Tom Cruise) have rarely, if ever, smoked in movies, and others (eg, Bruce Willis), who were known previously for playing smoking characters in highly popular movies (eg, Die Hard).
This is the first study to assess the delivery of movie smoking impressions to US adolescents using survey techniques that allow for a direct estimate of the exposure. The results indicate that this sample of popular contemporary movies delivered billions of smoking impressions to American youth and provide a basis, from a communications standpoint, for the large population effect (an adjusted attributable risk of 0.50 in 1 longitudinal study4) of movie smoking on adolescent smoking. A single popular movie with smoking can deliver tens of millions of smoking depictions to adolescents on the first run at the box office. Once that film also appears on DVD, video, and movie channels, the movie extends its reach and may ultimately deliver hundreds of millions of gross smoking impressions to youth. These findings represent a conservative estimate of the impact of movies, because we did not assess exposure to all of the films, did not determine how many times adolescents had seen each film, and restricted our survey to a both a narrow age range and date range for the movie sample. Indeed, children begin watching animated movies with smoking as preschoolers, and, with the penetration of VCRs and DVD players, are able to view their favorite films over and over throughout childhood and adolescence. Older movies are also readily available, resulting in the delivery of many more smoking depictions from these films than we were able to document in this study.
When examined according to movie rating, we found that, although R-rated movies accounted for the majority of occurrences of movie smoking, they delivered only ∼40% of smoking impressions to adolescents. This assessment corresponds closely with the findings by Polansky and Glantz.17 Youth-rated movies delivered proportionally more smoking because of higher viewership rates for these movies. One implication of our findings is an assessment of the likely impact of an R-rating for smoking on reach among young adolescents. An R rating for smoking is one of the policy planks of a public health advocacy movement aimed at decreasing the impact of movie smoking.19 Because movies are financed and produced with a target audience and rating in mind, an R rating for smoking would assure that movies intended for youth audiences would be smoke free. Assuming that new movies had a smoking mix similar to this sample and assuming that adolescent viewership rates for movies according to rating did not change, an R rating for smoking would decrease adolescent exposure to movie smoking in new releases by ∼60%, resulting in a substantial reduction in exposure over time. Thus, from a health perspective, if one accepts movie smoking as a causal influence in adolescent smoking, the goal to incorporate smoking into the ratings system has some appeal.
The decision to depict smoking in movies may be subject to forces other than industry rating policy. Our examination of character smoking according to actor shows that just 30 actors delivered one quarter of movie character smoking to young adolescents, representing the work of only 1.5% of the actors playing character roles in these movies and 6% of actors who smoked in movies. These 30 actors were primarily lead males who starred as smoking characters in multiple movies. A decision not to smoke in movies by any one of these characters would have had a substantial impact on current adolescent exposure to smoking. The behavior of these actors contrasts with other notable stars who played lead roles in multiple movies but did not smoke in any of them. For example, Ben Affleck, who has been shown smoking in “real-life” images captured by news magazines and tabloids, appears as 10 different characters within this movie sample, yet none of his characters smoked. This contrast suggests, for lead actors at least, a fair amount of autonomy with regard to the decision to depict smoking on screen, with many having chosen not to smoke. If this is so, another approach to decreasing smoking exposure to youth could be to encourage/pressure a relatively small number of high-profile actors not to smoke in films.
This study is subject to limitations, some of which are described above and relate to the underestimation of movie smoking reach among children and adolescents. Adolescent recollection of whether they had seen a movie title is subject to recall bias. However, adolescents correctly report having seen a movie 90% of the time more than a year after having seen it,2 so we think that their responses are accurate. Another limitation is that this study treats all movie smoking equally; it does not address any potential relationship between viewing different types of smoking and adolescent smoking. We would expect movie smoking to vary in its influence on adolescent smoking, depending on the individual characteristics of the adolescent and the context of the movie smoking depiction. At the movie level, our counts of movie smoking include character and background depictions of smoking and do not differentiate on the duration of the occurrence. In addition, our character smoking measure did not distinguish between different types of smoking (cigar, pipe, or cigarette) or other contextual elements, such as the type or attractiveness of character who smokes, whether the smoking occurred in an emotional or arousing scene, or the overall level of violence in the movie. One would expect, for example, that cigarette smoking depicted by Rusty Ryan (Brad Pitt) in Oceans Eleven might have more impact on adolescent cigarette smoking than pipe smoking by Gandolf (Ian McKellan) in Lord of the Rings. In addition, we do not consider differential impact by adolescent gender or race/ethnicity. More research needs to be done on whether these factors impact adolescent smoking behavior differentially.
Despite its limitations, this study offers the first direct assessment of the reach of movie smoking, providing a measure of the magnitude by which this medium delivers potentially powerful social influence impressions to a vulnerable population. The finding that movies deliver smoking depictions by the billions to US adolescents during a period when they are susceptible to social influences to smoke warrants further research to better understand the process by which these depictions might affect behavior. Of additional importance, with more than half of box-office receipts for US movies coming from overseas, US movies deliver smoking impressions to adolescents all over the world. We hope that this research will prompt those in the movie industry who make decisions about movie smoking to think carefully about the role they may be playing in the smoking epidemic both domestically and worldwide. Finally, we encourage pediatric practitioners to support the efforts of the American Academy of Pediatrics and other groups19 to pressure the movie industry to adopt voluntary incentives and other policies to limit smoking in movies. In addition, pediatricians should counsel parents regarding the media, monitoring steps that they can take to reduce movie smoking exposure to their children. As stated in the conclusion of a recently published article on the topic, “parental rules and monitoring of children's movie viewing may have a protective influence on children's risk for smoking and drinking, over and above parental monitoring of nonmedia-related behaviors.”20
We acknowledge funding from the National Cancer Institute.
We thank our collaborators, Madeline Dalton, Michael Beach, Jennifer Tickle, and Todd Heatherton, and also the content coders, Daniel Nassau, Balvinder Rakhra, and Elaini Bergamini, for their efforts.
- Accepted November 27, 2006.
- Address correspondence to James D. Sargent, MD, Dartmouth-Hitchcock Medical Center, One Medical Center Drive, Lebanon, NH 03756. E-mail:
The authors have indicated they have no financial relationships relevant to this article to disclose.
- ↵Sargent JD, Beach ML, Adachi-Mejia AM, et al. Exposure to movie smoking: its relation to smoking initiation among US adolescents. Pediatrics.2005;116 :1183– 1191
- ↵Sargent JD, Beach ML, Dalton MA, et al. Effect of seeing tobacco use in films on trying smoking among adolescents: cross sectional study. BMJ.2001;323 :1394– 1397
- ↵Tickle JJ, Sargent JD, Dalton MA, et al. Favourite movie stars, their tobacco use in contemporary movies, and its association with adolescent smoking. Tob Control.2001;10 :16– 22
- ↵Mekemson C, Glik D, Titus K, et al. Tobacco use in popular movies during the past decade. Tob Control.2004;13 :400– 402
- ↵Roberts D, Henriksen L, Christenson P. Substance Use in Popular Movies and Music. Rockville, MD: Office of National Drug Control Policy and Department of Health and Human Services Substance Abuse and Mental Health Services Administration; 1999.
- ↵Polansky J, Glantz S. First-Run Smoking Presentations in U.S. Movies, 1999–2003. San Francisco, CA: Center for Tobacco Control Research and Education; 2004
- ↵Glantz SA. SmokeFreeMovies homepage. Available at: http://smokefreemovies.ucsf.edu. Accessed October 27, 2005
- ↵Dalton MA, Adachi-Mejia AM, Longacre MR, et al. Parental rules and monitoring of children's movie viewing associated with children's risk for smoking and drinking. Pediatrics.2006;118 :1932– 1942
- Copyright © 2007 by the American Academy of Pediatrics