10 Role of Entertainment Media in Promoting or Discouraging Tobacco Use

Popular entertainment media are a powerful force in the lives of Americans. In particular, young Americans have been shown to spend an average of more than five hours per day exposed to a variety of media channels. This chapter examines the role of entertainment media in encouraging or discouraging tobacco use, including aspects such as

n Channels of media exposure, particularly for children

n Studies performed on tobacco use in the movie industry, ranging from trends in tobacco prevalence by movie type to issues such as how tobacco use is depicted, not portraying the health consequences of smoking, and brand-specific exposure

n Studies examining the influence of smoking in the movies on the social attitudes and behaviors surrounding smoking

n A summary of research on the portrayal of tobacco use in other media channels, such as television, music, magazines, and the Internet

n Current and future strategies for reducing public exposure to tobacco use in entertainment media, including policy interventions, efforts at industry self-regulation, and advocacy efforts aimed at both the public and the entertainment industry

The total weight of evidence from cross-sectional, longitudinal, and experimental studies indicates a causal relationship between exposure to movie smoking depictions and youth smoking initiation. Further research to better understand this relationship and to evaluate strategies to reduce youth exposure to tobacco portrayals in entertainment media is warranted.

357 1 0 . R o l e o f E n t e r t a i n m e n t M e d i a

It’s the movies that have really been running impact of entertainment media on behavior things in America ever since they were focuses on these media. The next sections of invented. They show you what to do, how this chapter describe what is known about to do it, when to do it, how to feel about it, the smoking images contained in movies and how to look how you feel about it. and how viewing them affects attitudes and behavior. The text begins with the historical —Andy Warhol (1928–87) relationship between the tobacco and movie industries, both of which came of age during the early 1900s in the United States. Introduction The chapter also summarizes research on portrayal of tobacco in other forms of This chapter examines and summarizes entertainment media including television, what is known about the use of tobacco music, magazines, and the Internet. Finally, in entertainment media and its effect on efforts to reduce audience exposure to tobacco use in the population. A detailed tobacco-related media content are discussed, look at the influence of one of America’s and overall chapter conclusions are drawn. oldest entertainment media—the movies— is followed by a discussion of how today’s What Are Entertainment Media? overall media environment can influence tobacco use and steps that can be taken Entertainment media include print media to reduce public exposure to tobacco use (books and magazines), audio media in the media. Given the continued rapid (radio and music), and audiovisual media growth in media access, particularly among (television, movies, Web-based media, and young people, reducing tobacco use in the video/computer games). Just two decades media could serve as an important factor in ago, options for media delivery in the changing social attitudes toward smoking. home increased with the introduction of the videocassette. Today, the options also It has long been believed that the include digital media (digital versatile discs entertainment industry has a profound [DVDs], compact discs [CDs], video games) impact on behavior, especially when it and access to entertainment programming comes to what is perceived as fashionable. through cable/satellite and the World The entertainment industry produces Wide Web. The Web provides unique stars who introduce large segments of the entertainment options through Web sites population to new products and behaviors that deliver everything from traditional depicted in mass media. To the extent venues, such as news, to options for playing that viewers form personal connections interactive video games with multiple with these stars through their use of the players and downloading podcasts of movies media, the viewers’ own behavior may be and television shows. The increase in home influenced. The entertainment industry options for media and the multiplication also serves to maintain behaviors already of media viewing sites within the home established in the population. (60% of U.S. households contain three or more television sets) have transformed This chapter begins with a look at the home media viewing from a family event media environment and its evolution as a to a much more individualized and tailored backdrop for examining media channels that pattern of media viewing among family could potentially model smoking behavior. members. For example, parents who grew Perhaps because television and movies up before video games or Music Television are so prominent in people’s leisure time (MTV) may know little about the specific entertainment, most of the research on the content of the video games their children

358 M o n o g r a p h 1 9 . T h e R o l e o f t h e M e d i a

play or the music videos and other video have a television in their bedrooms (65% of podcasts their adolescents watch because the children and adolescents older than age 7). parents generally do not play or watch them. Most adolescents also have a radio and a CD player in their bedrooms.1(p.13) About one-half Surveys of media availability in U.S. of families report that the television is almost households reveal broad access to each of always on, and 58% watch television during the home media channels, with electronic mealtimes.1(p.15) Average media exposure media gaining market share over traditional among children is 5.3 person-hours per day media venues. Two studies that surveyed (3.3 hours for 2–7 year olds and 6.4 hours for representative samples of U.S. families 8–18 year olds). Average media exposure is with children found similar results. about one hour less for high-income families Roberts and colleagues1 surveyed more than for low-income families.1(p.19) than 3,000 families in 1999. Woodard and Gridina,2 surveyed some 1,200 families one One study noted that children and year later. The proportions of families with adolescents distribute their time in using two or more media delivery devices were entertainment media in the following 88% for televisions, 58% for videocassette proportions: television, 46%; CDs and tapes, recorders, 85% for radios, 71% for tape 12%; movies and videos, 11%; print media, players, 59% for CD players, 38% for video 11%; radio, 10%; video games, 5%; and game players, and 21% for computers. computer, 5%.1(p.20) In addition, most families reported having access to a wide variety of television channels, As children age, one-half of the additional with about three-quarters of American time spent with media is due to an increase families having cable/satellite television.1(p.9) in television viewing; the remainder is due The only media services strongly related to increases in time spent watching taped to socioeconomic status were computer television shows, taking trips to the movie ownership and Internet access. All other theater, listening to the radio and music, products were equally distributed across and, for boys, playing video games.1(p.20–21) socioeconomic groups. For example, the Note that television viewing comprises both median number of televisions in households the viewing of television programming was 2.8 for families with incomes under (traditional programming and movies from $25,000, 3.0 for those with incomes between movie channels) plus nontraditional venues $25,000 and $40,000, and 3.0 for families such as MTV. Thus, the viewing of television with incomes above $40,000. The percentages programming and movies takes up more with cable/satellite television access for these than one-half of the five to six hours that income groups were 71%, 73%, and 77%, children use media each day. respectively. However, the percentages with Internet access were 23%, 42%, and 58%, All of these media have the potential to respectively.1(p.11) influence the attitudes and behavior of young consumers toward tobacco products. Media Use A large body of research exists on the impact of tobacco use in movies on attitudes toward The national surveys cited above also assessed smoking. This medium therefore serves media use by children and adolescents. as a valuable exemplar for further study in These young Americans are considered most how various mass media might influence vulnerable to the effects of media messages, the potential for tobacco use. Thus, movies and much of the research discussed here are the primary focus of this chapter. Later addresses the effects of media on their use sections examine research findings regarding of tobacco. About one-half of U.S. children exposure to tobacco in other media. Together

359 1 0 . R o l e o f E n t e r t a i n m e n t M e d i a

with the existing body of knowledge surrounding the portrayal of tobacco use in movies, this chapter forms a base for future work on the impact of entertainment media on tobacco-related health issues. Historical Perspective: Movies

Examination of the role of entertainment media in tobacco marketing is increasingly becoming an area of active research. Most of this work has focused on portrayal of tobacco in movies. Quantitative studies suggest that youth exposed to on-screen smoking are Early Lucky Strike advertisement more likely themselves to initiate smoking.3–9 targeted at women These reports should prompt more careful examination of the historical role that the to a woman who asks him to “blow some entertainment industry may have played my way.” The company also recognized the in the marketing of tobacco. Pierce and role movie stars play in establishing social Gilpin10 have identified four key periods in a trends and recruited prominent actresses of historical analysis of tobacco marketing and the time to endorse the brand in their print smoking initiation among U.S. adolescents advertisements. Chesterfield advertisements and young adults. Tobacco companies regularly featured glamour photographs marketed cigarettes to men during the of a Chesterfield “girl of the month,” first period, from the inception of the primarily fashion models and Hollywood industry’s marketing practices in the 1880s starlets. Some endorsers were actresses, to about 1920. By 1920, the market for men including Joan Bennett, Claudette Colbert, was established and considered mature.11 Joan Crawford, Betty Grable, Rita Hayworth, The industry then turned its attention to Marion Hutton, and Rosalind Russell. increasing sales among women.12 For the During the late 1940s, the advertisements next two decades, the industry added to continued to feature glamorous women but its marketing portfolio messages aimed also included male stars. Star endorsements at women. Campaigns explicitly targeted during this period included Charles Boyer, women, as exemplified by the Lucky Strike Perry Como, Bing Crosby, Arthur Godfrey, “Reach for a Lucky Instead of a Sweet” print Bob Hope, Dorothy Lamour, Virginia Mayo, media campaign during that period.13 Ethel Merman, Gregory Peck, Basil Rathbone, Ann Sheridan, Jo Stafford, and James Stewart. This specific campaign focused on weight control. However, the cigarette also was From 1943 through 1946, advertisements positioned as a symbol of independence for the Regent brand of cigarettes featured and equality for women. At about the same drawings of celebrities, including Fred time, Chesterfield rolled out a campaign Astaire, Diana Barrymore, Joan Blondell, aimed at changing social norms regarding Bing Crosby, Robert Cummings, Jinx smoking, with an emphasis on the social Falkenberg, Arlene Francis, June Havoc, interaction between men and women. The Celeste Holm, Guy Lombardo, Merle Oberon, campaign was launched by a 1926 billboard and Jane Wyatt.14 These advertisements depicting a man who is smoking, seated next provide historical evidence of a strong,

360 M o n o g r a p h 1 9 . T h e R o l e o f t h e M e d i a

that era were beginning to perceive the potential power of celebrities and the media (including motion pictures) as ways to change social norms around smoking. The work by public relations pioneer Edward Bernays15 is particularly relevant; for example, he sponsored, on behalf of the American Tobacco Company’s Lucky Strike cigarettes, demonstrations in 1929 in which fashion models gathered on street corners to smoke their “torches of freedom.”

The tobacco industry advertising campaign aimed at women is credited with the steady increase in cigarette smoking initiation Chesterfield cigarette advertisement rates among women during this period featuring actress Joan Crawford (1925–39) (figure 10.1). After 1939, and Note: from Ladies Home Journal 1949 through the mid-1960s, tobacco marketing no longer focused on any particular mutually beneficial relationship between the subgroup.10 However, smoking initiation cigarette industry and the movie industry. rates among women continued to increase It would be reasonable to assume that the at the same pace as they did through the stars were paid for their appearances in 1920s and 1930s. Attending motion pictures the advertisements as well as receiving was a national pastime by 1940, with nonmonetary benefits, such as increased Americans spending almost one-quarter exposure. Public relations specialists of of their total recreation dollars on movies

Figure 10.1 Smoking Initiation Rates Among U.S. Males and Females Ages 14–17 Years, by Year

Period 2 3 4 5 0.12 0.11 0.10 0.09 0.08 0.07

0.06 0.05 Initiationrate 0.04 0.03 Females 0.02 Males 0.01 0 1910 1920 1930 1940 1950 1960 1970 Calendar year Note. From Pierce, J. P., and E. A. Gilpin. 1995. A historical analysis of tobacco marketing and the uptake of smoking by youth in the United States: 1890–1977. Health Psychology 14 (6): 500–08. Copyright © 1995 American Psychological Association. Reprinted with permission.

361 1 0 . R o l e o f E n t e r t a i n m e n t M e d i a

Smoking: A Requirement of the Role

One case report describes an actor being introduced to smoking on the set of his first movie. In a New York Times Op Ed column,a Kirk Douglas states he never smoked during his Broadway career in the early 1940s. Mr. Douglas goes on to describe his first movie role, in 1946. “My first picture was The Strange Loves of Martha Ivers, with Barbara Stanwyck and Van Heflin, in 1946. I was intimidated, but proud to be playing the role of Miss Stanwyck’s husband. I arrived at the set, very excited, to do my first scene with her. But I had spoken only a few lines when the director, Lewis Milestone, stopped the action and said, “Kirk, you should be smoking a cigarette in this scene.” “I don’t smoke,” I replied timidly. “It’s easy to learn,” he said, and had the prop man hand me a cigarette. I continued with the scene, lighting and smoking my first cigarette. Suddenly, I began to feel sick to my stomach and dizzy. “Cut,” yelled the director. “What’s the matter with you, Kirk? You’re swaying.” I rushed to my trailer to throw up. But Mr. Milestone was right. It’s easy to learn to smoke. Soon I was smoking two to three packs a day.”a aDouglas, K. 2003. My first cigarette, and my last. New York Times, May 16.

(compared with only 2% today). Weekly for such scenes, they mirror the romantic attendance at U.S. theaters was more than themes included in cigarette advertising at 90 million.16 By 1940, depictions of actors the time, as illustrated by the Lucky Strike and actresses smoking in movies were an advertisements from the mid-1930s. established routine. The use of stars to endorse cigarettes in An example of how smoking depictions in advertisements continued into the 1950s, movies might have affected the population’s with Chesterfield endorsements from women social perceptions of smoking is the 1942 movie celebrities, such as Dorothy Lamour, movie Now, Voyager, starring Bette Davis Virginia Mayo, Ethel Merman, Ann Sheridan, and Paul Henreid. Bette Davis plays a young and Jo Stafford. In addition to leading Boston socialite who has been repressed ladies, the advertising of the 1950s heralded and dominated by her mother. She smokes new young stars, such as James Dean who surreptitiously until she meets and falls in depicted rebellious adolescent characters love with an older man (Paul Henreid) on and consolidated the image of the “bad a cruise. boy” smoker. In Rebel Without a Cause, the image of Dean smoking a cigarette was The sequence is captured at the close of the so intertwined with his character image that voyage, when Henreid lights two cigarettes smoking was incorporated into publicity and hands one to his lover just before a posters for his movies. Thus, smoking parting embrace. Given the popularity was promoted in another way—through of this movie and these stars at the time, publicity photographs and posters distributed this sequence may have influenced the worldwide (as the German rendition of the socialization of women to take up smoking, poster illustrates). in part by teaching men a novel way to offer a cigarette to a woman. Although no direct As television began to become a mass evidence supports an advertising motive medium, the tobacco industry began

362 M o n o g r a p h 1 9 . T h e R o l e o f t h e M e d i a

Scenes from Now, Voyager (1942)

Magazine advertisements for Lucky Strike documenting thematic similarities between cigarette advertising and movie depictions of smoking sponsoring television shows, providing cash movie industries emerged upon the release to this fledgling entertainment industry of tobacco company documents.18 Other before it had a sizable audience to attract documents indicate that several movie stars, other types of mainstream advertising.17 including Pierce Brosnan, James Coburn, Tobacco companies remained prominent Roger Moore, and Charlie Sheen, were sponsors until television advertising of recruited to represent a James Bond type tobacco was banned in the United States of figure in an advertising campaign for in January 1991. Television advertisements Lark cigarettes during the 1980s in Japan.19 produced during the 1950s included Chapter 4 describes in detail paid product endorsements by prominent movie stars. For placement of tobacco images in movies. example, John Wayne appeared in a number Although these documents pertain to brand of Camel commercials during this period. placements in movies produced during the 1970s and 1980s only, the practice probably The extent to which the tobacco industry preceded those decades. Schudson20 argues played a role in tobacco product placement in that the practice of deliberately mentioning movies was speculative until specific evidence or picturing particular products in films of financial links between the tobacco and occurred earlier. “In the 1930s and 1940s,

363 1 0 . R o l e o f E n t e r t a i n m e n t M e d i a

tobacco industry dates back to the inception of the media industry. The first focus was on marketing cigarettes to the U.S. population by securing endorsements from prominent stars and through prominent depiction of smoking in motion pictures. There is no early evidence of paid placement of tobacco products in movies. However, it seems likely that the depiction of smoking in films contributed to the Promotional posters for Rebel establishment of social norms that Without a Cause (in English encouraged women to smoke as a and German) mark of independence and equality, as a way to establish a conversation De Beers increased the role of diamonds (break the ice) between men and women, in Hollywood films, just as cigarette and in ways that paralleled other cigarette manufacturers saw to it that leading actors advertising themes at that time. Early movie and actresses smoked cigarettes in movies images of male smokers as tough and in the 1920s.”20(p.101) It would be surprising independent also may have promoted to if A. D. Lasker, Edward Bernays, and other men the appeal of tobacco use. In addition, public relations specialists of that era failed the entertainment industry was key in to recognize the potential power of motion establishing the prototype of the rebellious pictures as a way to change social norms adolescent cigarette smoker. This prototype concerning smoking. As discussed below in continues to attract adolescents to smoking “Movie Content,” smoking continues to be in the present. depicted in movies. Cigarette brands also appear, although movie scenes showing actors actually using a specific brand have declined. Movie Content

In summary, the relationship between the Content analysis refers to a research method media entertainment industry and the in which coders systematically count and

Tobacco Portrayal Goes Beyond the Movie Itself

Tobacco product exposure in movies is not necessarily limited to the actual film content. The depiction of smoking and brands in promotional photographs still occurs. For example, the photograph shown here, released with a set of promotional photos by Screengems Productions for the movie Snatch, was widely published in newspapers across the United States. The photograph shows Brad Pitt sitting at a desk with a pack of Marlboro Golds. Interestingly, no cigarette brand appeared in the actual movie. The practice of showing smoking and cigarette brands in movie promotional products has not been studied systematically. Therefore, it is difficult to determine how Publicity photograph released with important these materials are from a communications standpoint. the movie Snatch, Screengems, 2000.

364 M o n o g r a p h 1 9 . T h e R o l e o f t h e M e d i a

Thank You for Smoking

Jason Reitman’s 2006 satirical film, Thank You for Smoking,a based on Christopher Buckley’s novel, highlights some of the realities of the relationship between the media and tobacco. The main character in the movie, Nick Naylor, is a spokesperson for the fictional Academy of Tobacco Studies run by cigarette manufacturers. Naylor suggests that declining rates of teen smoking can be turned around through the use of smoking in upcoming Hollywood films. He travels to Los Angeles to meet with an agent and negotiate the use of cigarettes in a futuristic film “where smokers and nonsmokers live in perfect harmony.” Both Naylor and the agent acknowledge that the use of cigarettes by Catherine Zeta Jones and Brad Pitt will “sell a lot of cigarettes.” Real-life tobacco companies have been banned from sponsoring Hollywood films since the 1998 Master Settlement Agreement. However, the use of cigarettes in movies is still prominent, and studies examined later in this chapter show a positive correlation between exposure to on-screen smoking and smoking initiation rates for adolescents. One studyb of 6,522 randomly selected participants suggests that exposure to on-screen smoking is the primary independent risk factor for teen initiation rates. So Naylor’s prescription to have actors smoke on screen in order to “sell a lot of cigarettes” is, at least among adolescents, supported by academic research. The correlation between on-screen smoking and smoking initiation rates has led to some tobacco control groups pushing for more restrictive ratings for movies portraying tobacco use. So far, these efforts have been unsuccessful. It is unlikely that these groups will switch to Thank You for Smoking’s final tobacco control idea: digital replacement of cigarettes in classic films with candy canes, steaming mugs of cocoa, and drum sticks. aReitman, J. 2006. Thank You for Smoking [Motion picture]. United States: Fox Searchlight Pictures. bSargent, J. D., M. L. Beach, A. M. Adachi-Mejia, J. J. Gibson, L. T. Titus-Ernstoff, C. P. Carusi, S. D. Swain, T. F. Heatherton, and M. A. Dalton. 2005. Exposure to movie smoking: Its relation to smoking initiation among US adolescents. Pediatrics 116 (5): 1183–91.

characterize media inputs. Published content Fields] OR (“motion pictures”[MeSH analyses examining depictions of tobacco use Terms] OR motion picture[Text Word]))— in entertainment media have focused almost 103 records obtained, May 9, 2006. exclusively on movies. Less information is available concerning tobacco-related content A search of PsycINFO using the key words in other entertainment media. ((“tobacco” OR “smoking”) AND (“movie” OR “motion picture”)) and restricted to Study Selection journal articles written in English identified no additional articles on movie content A number of content analyses have been analysis than those already captured by the conducted of portrayal of tobacco in popular MEDLINE search (23 articles retrieved, by movies. Fourteen peer-reviewed studies PsycINFO, May 9, 2006). were identified as published in the medical literature (in English) by using a PubMed Citations in some of the above papers21 search strategy on MEDLINE with the identified one more peer-reviewed paper following search terms and Medical Subject that examined tobacco as well as other Headings (MeSH): health-relevant behaviors in movies. Further citations to a study by Mekemson ((“tobacco”[MeSH Terms] OR tobacco[Text and colleagues,22 a Web-based report,23 Word]) OR (“smoking”[MeSH Terms] OR provide additional findings from the smoking[Text Word])) AND (movie[All American Lung Association’s “Thumbs Up!

365 1 0 . R o l e o f E n t e r t a i n m e n t M e d i a

Thumbs Down!” ongoing content analysis. location of setting; that study excluded Four additional published reports on movies in which cigarette smoking was a this subject were identified that were of central motif.38 methodological quality comparable with the peer-reviewed studies.24–27 These reports were Another study identified the “top 10” most commissioned by public agencies, including popular actresses per year for a given the White House Office of National Drug period, then randomly sampled movies Control Policy and the U.S. Department in which each played a leading role. of Health and Human Services’ Substance A number of studies have excluded from Abuse and Mental Health Services their samples movies that were not set in Administration;27 Center for Tobacco the present—that is, period dramas and Control Research and Education, University science fiction set in the future.21,38 Despite of California;26 the Health Education sampling differences among some studies, Authority in the United Kingdom;24 and most have used sampling criteria based on the Massachusetts Public Interest Research audience reach. Therefore, the media inputs Group (a nongovernmental, voluntary they documented are likely to provide a organization).25 Table 10.1 summarizes the valid indication of the amount and nature methods of movie selection and coding of of on-screen tobacco content presented to tobacco use for the respective studies. viewers. Polansky and Glantz26 extended their content analysis data to generating Methodological Issues quantitative estimates of audience reach (see “Audience Reach” below). Together, various studies have sampled and coded tobacco content in popular movies Studies also vary in how they capture released from 1937 through 2003. However, tobacco use, especially in terms of their the studies’ methodological differences unit of analysis. Many divided their movie make it difficult to compare the results. samples into five-minute intervals and The most common criterion for selecting then counted the number of tobacco movies was based on their revenue status occurrences per five-minute interval of as “top box-office” movies, mostly in the film.21,27–31,40 Others viewed and coded United States. Some studies28–30 selected a movies as a whole, counting tobacco random sample of top box-office movies for occurrences within movies.22,24–26,32–38 Some a given period. Others coded the top 10,24,31 included as one occurrence all smoking 25,32,33 50,22 100,34 125,35 or 200 movies per by one character during the course year,27 or those grossing at least $500,000 at of a movie scene.32 Others counted an the box office26 for a given period of years. occurrence every time a cigarette entered In general, the longer the period examined, the screen.22 These differences obscure the fewer movies per year were coded. Other comparisons in the absolute numbers of studies have selected the movie sample tobacco depictions reported among the based on genre or rating only (e.g., G-rated studies. Moreover, it is not clear how well animated movies)36,37 or a combination of the various measures correlate or whether rating and box-office revenue (e.g., top 10 measurement affects trend analyses. PG movies and video rentals).25 One study However, Polansky and Glantz26 found that examining the prevalence of smoking parents’ qualitative ratings of the amount among characters in contemporary of smoking in movies (using a six-point American movies about American life in the ordinal scale ranging from “none” = no 1990s relative to U.S. population smoking tobacco content through “extreme” = movie rates selected movies on the basis of box- is full of tobacco scenes) bore a statistically office revenue, rating, genre, and time and significant correspondence with coding

366 M o n o g r a p h 1 9 . T h e R o l e o f t h e M e d i a other variables Reported results of SC, HC, Chg, TT SC, HC, Chg Chg, Brand SC, HC, Chg, CX, TT SC, HC, Chg, CX SC, HC, CX, TT, MT (coded, but results not reported), Brand (coded, but results not reported) SC, HC, Chg, CX, TT SC, HC, Chg, MT Amount of smoking Presence of tobacco % of characters who No. of tobacco % tobacco use per Presence of tobacco No. of tobacco use; duration smoked at least once in target films incidents/hr of movie time 5-min interval of movie (includes use and implied) % movies depicting tobacco; duration of smoking scenes Tobacco incidents use (includes and implied) occurrences; time of tobacco use Unit of analysis Whole movie Whole movie 5-min intervals of movies Each movie divided Each movie divided into Whole movie into 5-min intervals ( n = 1,505 5-min intervals) Whole movie Each movie divided into 5-min intervals 5-min intervals 28 Interrater reliability Not reported 1 rater only Ranged from 0.74 to 0.94. At least 3 trained raters rated each film. One coder, cite validity of this approach as per Hazan et al. 1994 0.92 (sd = 0.07) 95% Not reported At least 70% agreement on all measures Movie criteria All G-rated animation movies released by major production cos. ( n = 50) All G-rated animated feature movies ( n = 81) Random sample of 5 films per decade sampled from top 20 films per year for period ( n = 100) Random sample of 20 top films released 1950–59, and 5 of top 20 films for 2001–02 ( n = 30) 2 movies randomly selected from top 20 per year, for each year ( n = 62) Top 20 movies per year ( n = 169) Top 10 movies per year ( n = 110) Top 25 movies per year ( n = 250)

(release) Movie years 1937–99 1937–2000 1940–89 1950–59, 2001–02 1960–99 1977–88 1985–95 1988–97 U.S. U.S. U.S. U.S. U.S. U.S. U.S. U.S. Country

1999 2001 1998 2004 1994 1991 1998 2002 year Publication 37 37 36 39

d Yokota d 31

32 29 28 21

n an n Authors able 10.1 in Movies Summary of Methods for Content Analysis Studies: Tobacco Thompso McIntosh et al. Glantz et al. Terre et al. Everett et al. Dalton et al. Goldstein et al. Hazan et al. Note. cos. = companies; SC smoker characteristics; HC health consequences; Chg change over time; TT type of tobacco; CX context; min minutes; no. number; hr hour; Brand brand appearances; sd = standard deviation; MT movie type; MSA Master Settlement Agreement; +ve positive verbal reference; –ve negative sci-fi science fiction. T

367 1 0 . R o l e o f E n t e r t a i n m e n t M e d i a other variables Reported results of SC, HC, Chg, CX, TT, Brand, MT Chg: pre–post MSA, SC, HC (antismoking SC, HC (verbal SC, HC, CX, TT, MT SC, HC, CX, TT, Brand, SC (minors smoking), Brand, MT signs or comments), smoke), TT, Brand references: +ve/–ve), Chg, CX, Brand MT HC (–ve statements about tobacco use), Chg, CX (of brand appearances), TT, Brand, MT Chg, CX (motivation to

5-min movie

/

Amount of smoking Tobacco incident counted each time tobacco appears on screen % of movies with Occurrence of smoking in each interval coded Tobacco appearances % tobacco use per No. of smoking Length of tobacco tobacco use % of 5-min intervals that depict smoking (use and implied) frequency of tobacco reported 5-min interval of movie incidents interval use and tobacco appearance

Unit of analysis Whole movie Whole movie Each movie divided into Each movie divided into 5-min intervals Whole movie Whole movie? (not Whole movie 5-min intervals ( n = 1,116 5-min intervals) clear)

, kappa kappa ,

d by by d s >0.85 >0.85 s

s code s

y score y

n reviewers n

l movie l

Interrater reliability 3 tee 3 reliabilit Al Double coding of brand appearances, not reported One coder, not reported 99% agreement on smoking variables Not reported Not reported Not reported Movie criteria ( n = 20) per year (1991–2000: n = 497) (1994–2003: n = 498) ( n = 42) Top 25 movies per year ( n = 250) Random sample of 5 from top 20 movies per year ( n = 35) Top 10 U.K. box-office movies for these years actress, 10 top each For selection random coded of 5 movies in which they had leading roles ( n = 50) Top 50 U.S. movies 200 most popular movie video rentals for 1996 and 1997 ( n = 200) Top 10 PG-13 movies and top 5 PG-13 video rentals for these years (release) Movie years (+1994–2003) MSA) 1988–97 1990–96 1990 and 1995 1993–97 1991–2000 1996–97 1996–97, 1999–2000 (i.e., pre–post U.S. U.S. U.K. U.S. U.S. U.S. U.S. Country 2001 1997 1998 2000 2004 1999 2002 year Publication 24 24 30 22

25 40 d Owen d 33 27 23

n an n Authors Association of Sacramento Stockwell and Glantz MacKinno Escamilla et al. Mekemson et al. American Lung Sargent et al. Roberts et al. Ng and Dakake Table 10.1 Table in Movies (continued) Summary of Methods for Content Analysis Studies: Tobacco

368 M o n o g r a p h 1 9 . T h e R o l e o f t h e M e d i a other variables Reported results of Chg, MT SC, MT SC, MT SC, HC, CX, MT Amount of smoking Qualitative descriptors % of top 5 characters % of major adolescent % of movies with of no. tobacco incidents per movie who smoke at least once in target films (excluding smoking by non-U.S. citizens or outside U.S.) characters who smoke at least once in target films tobacco use, % of characters who smoke at least once in target films Unit of analysis Whole movie Whole movie Whole movie Whole movie

32 Interrater reliability for equivalent films Not reported, although good correspondence with tobacco incident data in a study by Dalton and colleagues Not reported Ranged from 0.77 to 1.00 Ranged from 0.84 to 0.99. Movie criteria U.S. produced, English- speaking movies grossing at least $500,000 at box office ( n = 776) U.S. movies in top 10 weekly box-office list, excluding G-rated, animated, and sci-fi movies, movies not set in 1990s, and movies with cigarette smoking as a central motif ( n = 447) Of 125 top-grossing movies per year, coded those where at least 1 adolescent was a central character ( n = 43) 100 top-grossing films in 2002 ( n = 100) (release) Movie years 1998–2003 1990–98 1999–2001 2002 U.S. U.S. U.S. U.S. Country 2004 2006 2005 2005 year Publication 26 41 34 Authors 35 Polansky and Glantz Omidvari et al. Stern Dozier et al. Table 10.1 Table in Movies (continued) Summary of Methods for Content Analysis Studies: Tobacco Note. cos. = companies; SC smoker characteristics; HC health consequences; Chg change over time; TT type of tobacco; CX context; min minutes; no. number; hr hour; Brand brand appearances; sd = standard deviation; MT movie type; MSA Master Settlement Agreement; +ve positive verbal reference; –ve negative sci-fi science fiction.

369 1 0 . R o l e o f E n t e r t a i n m e n t M e d i a

conducted by Dalton and colleagues32 of Characteristics of smoking role models and the number of tobacco incidents for a depictions of contexts and consequences sample of 389 movies coded by both studies associated with smoking also have been (p < 0.001). This finding suggests a strong recorded. Some studies examined the correspondence between the two different types of tobacco presented (e.g., cigarettes, methods of coding the amount of on-screen cigars, chewing tobacco), the appearances smoking used in these studies. of specific tobacco brands, and whether tobacco portrayal varied with movie release The studies also vary in how rigorously they year, Motion Picture Association of America describe their coding procedure. Of the (MPAA) rating, or genre. Common themes studies reviewed here, only eight reported recurred in the findings of these studies, interrater reliability agreement, with values despite their methodological differences. ranging from 70% to 100% on key coding The results of these studies are summarized variables.21,22,28,32,34,35,39,40 Most studies used below. adults to code movie content, the exception being the “Thumbs Up! Thumbs Down!” Tobacco Use in Movies project.22,23 The latter study trained teams of young people aged 14–22 years to code films Prevalence by Movie Type according to a standard protocol. The adult coders in the study reported by Polansky Mekemson and colleagues22 found that and Glantz26 were parents working for a most top box-office movies from 1991 to parental review and screening service at 2000 had some tobacco use. Polansky and ScreenIt.com, a movie content database. Glantz 26 found that, of U.S. films released between 1999 and 2003, 80% included The criteria for coding tobacco events smoking. Similarly, content analyses of top also varied. Explicit depictions of tobacco box-office movies from 1988 to 1997 indicate use refer to instances in which the use of that most movies (87%) portrayed tobacco tobacco was directly portrayed (e.g., the use. However, tobacco use accounted for actor smokes on screen). Incidental only a small proportion of screen time.32 depictions of tobacco refer to those in which In 75% of movies, tobacco exposure the use of tobacco was implied, without accounted for less than 4% of total screen being explicitly portrayed (e.g., the actress time. Cigarettes were the predominant is shown placing a cigarette pack in her form of tobacco used, followed by cigars, handbag), or when smoking-related props with little use of smokeless tobacco.27,32 were shown (e.g., an ashtray on a table in a However, in children’s animated movies, movie set). Some content-analysis studies cigar use was most common.36 Tobacco use only coded explicit depictions of tobacco typically increased with the “adultness” of use.32,38 Others differentiated between types the MPAA rating. R-rated movies contained of tobacco depictions.27 Some counted more tobacco occurrences and were more explicit and incidental depictions of tobacco likely to feature major characters using together as tobacco events.29,31 Studies tobacco.22,26,27,32,34 For U.S. movies released with broader criteria for a tobacco incident from 1999 to 2003, a higher proportion of tended to report higher rates of depiction as R-rated movies included smoking (90%) a result of their more inclusive measure. compared with PG-13 (80%) and G/PG movies (50%). However, because of a decline There is, however, considerable overlap in in the total number of R-rated movies the content variables the studies attempted released between 1999 and 2003, a shift to assess (table 10.1). All quantified the occurred in the total distribution of movies amount of smoking in their movie samples. containing smoking. Most of the movies

370 M o n o g r a p h 1 9 . T h e R o l e o f t h e M e d i a

released in 2002 and 2003 that contained to exceed historical trends for smoking smoking scenes had a youth rating (PG-13 prevalence. or G/PG).26 However, trends in the sheer frequency Tobacco use was more common in dramas with which tobacco appears in movies than in comedies, science fiction, or child across years do appear to be discordant and family genres.32 Similarly, Dozier with declining smoking rates in the actual and colleagues34 found that characters population. In a sample of top box-office in comedies smoked less frequently U.S. films from 1950 to 2002, the number of than in other genres among 2002’s top- smoking incidents per 5-minute interval of grossing movies. The amount of tobacco film declined from 10.7 incidents per hour use in movies did not have a significant in 1950 to a minimum of 4.9 in 1980 to 1982 association with the movies’ box-office but increased to 10.9 in 2002.28–30 Another success.32 This finding may suggest that study found that, after an initial drop in including tobacco in movies provides the frequency of depicting tobacco in the no direct economic benefit to the 1970s and mid-1980s, the rate subsequently entertainment industry. This notion is increased.21 Dalton and colleagues32 found bolstered by experimental evidence that that the number of tobacco occurrences among adolescent moviegoers, stripping in top box-office U.S. movies remained the smoking from a movie does not constant between 1988 and 1997, despite affect their satisfaction with the movie or declining trends for smoking prevalence in willingness to recommend it to a friend.42 the actual U.S. population. Mekemson and others22 found a weak decline in the amount Trends in the Amount of Tobacco of tobacco use per minute of film between Depicted in Movies Across Years 1991 and 2000. However, these rates appeared to increase again between 2001 Examination of changes over the years and 2003.23 MacKinnon and Owen24 found in the frequency of on-screen depiction that smoking was depicted more frequently of tobacco highlights some discrepancies in movies released in 1995 than in 1990. between movie portrayals of smoking and the social reality of smoking. In a content The depiction of smoking in children’s analysis by Dalton and colleagues32 of the animated films did not decrease between top 25 box-office hits from 1988 to 1997, 1937 and 1997.36 Later analyses of the the rate of tobacco use among 1,400 major “Thumbs Up! Thumbs Down!” content characters was 25%. This finding was analysis dataset23 found that in PG-13 films, not discordant with the prevalence of the total number of tobacco incidents smoking among U.S. adults during that depicted per year increased substantially period. McIntosh and colleagues39 found between 2000 and 2003. Thus, the argument that the proportion of leading characters that on-screen smoking reflects social who smoked increased from 20% in the realism does not hold up as a reason for 1940s to 31% in the 1950s. The proportion trends in the rate of smoking depiction then declined to 18% in the 1960s, 17% in in movies across the years. Movie content the 1970s, and finally 12% in the 1980s. appears to be out of step with declining Omidvari and others38 found that, among smoking rates in the U.S. population. contemporary U.S. movie characters during These results raise questions about the 1990s, smoking prevalence was similar the role of films in amplifying notions to that in the general U.S. population. of tobacco smoking being widespread. In these three studies, the proportion of A number of movie content analysis studies characters who smoked does not appear observed a pattern of increased depiction

371 1 0 . R o l e o f E n t e r t a i n m e n t M e d i a

of smoking in the late 1980s and early replicating sampling and coding methods 1990s. This time span follows the period over time will be necessary to confirm during which there is documented evidence whether a significantdecline has occurred in of paid tobacco product placement deals on-screen smoking among teen characters. occurring in relation to film. Examination Dalton and colleagues32 found that only 3% of trends in the rate of movie depictions of of tobacco occurrences were adolescents tobacco in relation to key tobacco-control smoking and that the typical smoker in events suggests these events have not movies was white, male, middle-aged, and of precipitated marked reductions in on-screen high socioeconomic status—traits possessed tobacco portrayals.33,36 by most leading characters. Omidvari and colleagues38 found that among leading Characteristics of On-Screen Smokers American movie characters portrayed in the United States in the 1990s, smoking As indicated earlier, smoking prevalence on-screen was associated with being male among characters in films was not markedly and of lower socioeconomic class. discordant with smoking prevalence in the actual population (i.e., 25%).32 However, The different findings of these studies in Dalton and colleagues32 found that the social relation to the apparent class of on-screen characteristics of leading characters were smokers may reflect the different sampling atypical (e.g., attractive, high socioeconomic methods used. Dalton and colleagues32 and status) so the characters represented as Dozier and colleagues34 selected movies smokers did not reflect the social reality solely on box-office rating. Omidvari and of smoking. Hazan and colleagues28 found others38 selected a subset of top box-office that between 1960 and 1990, the prevalence movies based on a range of exclusion of smoking among major characters with criteria (table 10.1). The findings of Dalton high socioeconomic status was nearly and colleagues provide an account of three times as high as among people of smoking prevalence among prominent similar socioeconomic status in the actual movie characters during the 1990s across U.S. population. In the 1980s, tobacco movies of all genres set in all eras. However, events involving young adults (aged 18–29 Omidvari and colleagues38 evaluated smoking years) more than doubled compared with prevalence among U.S. movie characters the previous two decades. However, tobacco in films of realistic genres set in the 1990s. events involving somewhat older adults These researchers focused on this subset (aged 30–45 years) fell by nearly one-half.28 of movies on the grounds that they were More recent movies tended to portray examining how movies portrayed smoking smoking by adults more often than smoking prevalence in contemporary life. Films set by adolescents. For popular movies from in the present may present smokers as more 1996 and 1997, smoking rates of 17%, 26%, socially disadvantaged than did films in and 25% were recorded for major characters previous eras. The study by Omidvari and aged younger than 18, 18–39, and older colleagues provides a useful snapshot of how than 39 years, respectively.27 contemporary on-screen smoking depictions compare with smoking prevalence in the Stern35 found an identical smoking general U.S. population. However, they do prevalence (17%) among major teen movie not represent a complete picture in terms characters for top-grossing films from of audience reach and impact of on-screen 1999 to 2001. Dozier34 found that only 2% smoking (this was not their aim). As Glantz of teenagers smoked in top-grossing films and Polansky43 argue, there is no evidence for 2002. The on-screen smokers tended to that viewers, particularly adolescents, be adult, white, and male. Future studies distinguish between portrayals of tobacco

372 M o n o g r a p h 1 9 . T h e R o l e o f t h e M e d i a

in historical, contemporary, and futuristic antagonists.38 Two cross-sectional surveys films or between portrayals of tobacco in of movie content report that in movies American and non-American films to which released during the 1990s, smoking was they are exposed. increasingly associated with stress reduction and hostility.24,28 It is unclear whether this The concern about the types of characters shift in imagery reflects changes in social who are predominantly depicted as norms concerning smoking, cinematic style, smokers in movies is that smoking is or commercial factors. modeled by characters bearing aspirational traits—such as good looks, maturity, Health Consequences affluence, and power—similar to the sorts of images traditionally promoted in tobacco A key concern about depictions of smoking advertisements. Theories of media influence on screen is that the health consequences of and persuasion predict that role models smoking are rarely shown. Content analyses bearing such traits are the most influential of children’s animated films released to audiences.44,45 As described later in this between 1937 and 1997 indicated that chapter, in “Effects on Attitudes, Beliefs, and more than two-thirds of the films included Behavior: Movies,” some audience studies tobacco use without clear verbal messages suggest that the sheer frequency of exposure of any negative long-term health effects of (across all movie genres and settings) smoking.36 Similarly, Hazan and colleagues28 is important to media impact. Audience found that most tobacco events in movies studies have not yet examined whether from 1960 to 1990 did not include health responses vary with the historical setting of messages. Roberts and others27 found that, smoking. Evidence is emerging, however, among the 200 most popular movie rentals that responses vary with character traits of for 1996 and 1997, negative long-term smoking models. health effects associated with substance use (smoking, drug use, or alcohol Other Social and Emotional Imagery consumption) were rarely depicted (in less than 7% of movies). Similarly, an analysis McIntosh and colleagues39 found that in by Everett and colleagues31 of top box-office popular films from 1940 to 1989, smokers U.S. films from 1985 to 1995 indicated that were depicted as more romantically and on average only 3.5% of tobacco events sexually active and marginally more were antitobacco, compared with 32.3% intelligent than nonsmokers. However, of tobacco events that were categorized as smokers and nonsmokers did not differ in protobacco. In top-grossing films for 2002, terms of their attractiveness, goodness, most (92%) incidents involving tobacco socioeconomic status, aggressiveness, were portrayed without consequences.34 friendliness, or outcome at film’s end. In movies released from 1988 to In another study, youth viewers found that 1997,32,34 smoking often is depicted 74% of the top 50 movies between 2000 (1) in association with intimacy and and 2003 that depicted tobacco contained social activity; (2) as motivated by certain protobacco messages.23 Dalton and mood states (e.g., agitation, sadness, colleagues32 found that negative reactions happiness, relaxation, pensiveness); or to tobacco use (e.g., comments about health (3) in conjunction with other risk-taking effects or gestures such as coughing) were behaviors (e.g., drug use or violence).32 depicted in only 6% of tobacco occurrences. Among American movie characters Escamilla and others40 found that movies portrayed as contemporary in the 1990s, rated as PG/PG-13 were less likely than smoking was more common among R-rated movies to contain negative messages

373 1 0 . R o l e o f E n t e r t a i n m e n t M e d i a

about smoking. In PG/PG-13 films, only one-half of box-office receipts for these films 9 of 22 tobacco messages were antitobacco, are from overseas.33 compared with 21 of 31 messages in R-rated/ unrated films. It is especially of concern Often, brand appearances involve only that health effects may be more frequently glimpses of cigarette packaging in the omitted from movies targeted toward ambient scene environment. A subset of younger audiences. As demonstrated by brand appearance of particular concern, social learning theory,45 showing hazardous termed actor endorsement, is display of behaviors in the absence of negative the tobacco brand while an actor handles consequences is likely to make viewers more or uses a product.33 It is reasonable to inclined to mimic them than if the negative single out actor endorsement, because the consequences were shown. film industry does so in its negotiations for placements for various products, often Brand Appearances asking for a higher payment when an actor uses a particular brand.33 Table 10.2 is Content analyses suggest that appearances derived from an ongoing content analysis of specific tobacco brands in movies occur of the top 100 box-office hits and covers frequently, despite a voluntary agreement the years 1996–2002. The table lists all on the part of the tobacco industry to actor endorsement tobacco events captured stop paying for their brands to appear during the seven-year period. The table (the Cigarette Advertising and Promotion documents 46 tobacco brand endorsement Code incorporated a voluntary ban on scenes from 43 of the 700 movies, thus paid product placement circa 1991). In a giving a measure of the scope of the 10-year sample of top box-office films from activity. Table 10.2 also illustrates that 1988 to 1997, the most highly advertised foreign cigarette brands are rarely depicted, U.S. cigarette brands also accounted for the the Marlboro brand captures most actor most brand appearances in the movies, and endorsements (25 of 46 endorsements), no decline occurred after 1991.33 Most (85%) actor endorsement is not limited to one or of the films contained some tobacco use, two actors, and actor endorsement usually with specific brand appearances in 28% of occurs only once or twice during the the total film sample. Brand appearances course of a movie. The one exception is the were as common in films suitable for movie 28 Days, which contains nine actor adolescent audiences as in films for adult endorsements of Marlboro. audiences. Although 27 tobacco brands were depicted in the movies sampled, 4 cigarette Audience Reach brands accounted for 80% of brand appearances. The brands were Marlboro One issue limiting the utility of content (40%), Winston (17%), Lucky Strike (12%), analysis studies is that most do not include and Camel (11%). Other content analyses an estimate of reach. Reach typically is of movies sampled from the late 1990s defined as the number of people who see a have found that brand appearances for particular form of advertising.46 Polansky and Marlboro occurred five to six times more Glantz26 estimated reach among adolescents frequently than those for other tobacco for smoking in movies released at the box brands.24,27 The U.S. film industry’s use of office between 1999 and 2003. They first the most heavily advertised tobacco brands estimated the number of smoking depictions (see chapter 4 for advertising expenditures contained in 776 movies released during by brand) in internationally distributed this period by using data from ScreenIt.com films suggests that film serves as a global (i.e., about 5,500 tobacco incidents in all advertising medium for tobacco, as about movies). They then used box-office data

374 M o n o g r a p h 1 9 . T h e R o l e o f t h e M e d i a

Table 10.2 Brand Cigarette Use Depicted in Contemporary Movies Number of endorsement Year of Actor name Brand endorsed scenes Movie name release Drescher, Fran Marlboro 1 Jack 1996 Eldard, Ron Marlboro 1 Sleepers 1996 Davis, Geena Parliament 1 Long Kiss Goodnight, The 1996 Addy, Mark Foreign Brand 1 Full Monty, The 1997 Carlyle, Robert Foreign Brand 1 Full Monty, The 1997 Roberts, Julia Marlboro 2 My Best Friend’s Wedding 1997 Sheen, Charlie Marlboro 1 Money Talks 1997 Franz, Dennis Camel 1 City of Angels 1998 Newman, Paul Camel 1 Twilight 1998 Sarandon, Susan Camel 1 Twilight 1998 Hawke, Ethan Kool 1 Great Expectations 1998 Cage, Nicolas Marlboro 1 Snake Eyes 1998 Janssen, Famke Marlboro 1 Rounders 1998 Keaton, Michael Marlboro 1 Desperate Measures 1998 Reno, Jean Marlboro 1 Godzilla 1998 Eastwood, Clint Camel 2 True Crime 1999 Bujold, Genevieve Foreign Brand 1 Eye of the Beholder 1999 Leguizamo, John Marlboro 1 Summer of Sam 1999 Quaid, Dennis Camel 1 Frequency 2000 Bullock, Sandra Marlboro 4 28 Days 2000 Buscemi, Steve Marlboro 1 28 Days 2000 Dooly, Mike Marlboro 1 28 Days 2000 Pratt, Wendee Marlboro 1 28 Days 2000 Santoni, Reni Marlboro 1 28 Days 2000 Skye, Azura Marlboro 1 28 Days 2000 Vaughn, Vince Marlboro 1 Cell, The 2000 Carrey, Jim Marlboro 1 Me, Myself & Irene 2000 Wilhoite, Kathleen Marlboro 1 Pay It Forward 2000 Schwimmer, Rusty Marlboro 1 Perfect Storm, The 2000 Fisher, Carrie Marlboro 1 Scream 3 2000 Scott, Dougray VF 1 Mission: Impossible II 2000 West, Dominic Winston 1 28 Days 2000 Washington, Denzel Kool 1 Training Day 2001 Barrymore, Drew Marlboro 1 Riding in Cars with Boys 2001 Rockwell, Sam Marlboro 1 Heist 2001 Zahn, Steve Marlboro 1 Riding in Cars with Boys 2001 Germann, Greg Parliament 1 Joe Somebody 2001 Crowe, Russell Winston 1 Beautiful Mind, A 2001 de Matteo, Drea Winston 1 Swordfish 2001 Hoechlin, Tyler Bugler 1 Road to Perdition 2002 Johnson, Carl J. Marlboro 1 Men in Black II 2002 Note. From a content analysis of the top 100 movies each year from 1996 through 2002.

375 1 0 . R o l e o f E n t e r t a i n m e n t M e d i a

from the National Association of Theatre studies provide informative descriptive Owners and Nielsen data on average data but do not provide conclusive audience share by age as well as the MPAA information as to impact of the media. ratings to determine the number of children A search of PubMed identified seven such 6–17 years of age who purchased tickets to studies by using the following strategy: see these movies. The MPAA is the lobbying arm of the film industry. The researchers ((“focus groups”[MeSH Terms] OR focus estimated that the thousands of smoking group[Text Word]) OR qualitative[All incidents in hundreds of movies multiplied Fields]) AND ((“tobacco”[MeSH Terms] OR by the number of tickets purchased to see tobacco[Text Word]) OR (“smoking”[MeSH these movies resulted in about 8.2 billion Terms] OR smoking[Text Word])) smoking depiction impressions for children AND (movies[All Fields] OR (“motion and adolescents during the five-year period. pictures”[MeSH Terms] OR motion Although these estimates are subject to picture[Text Word]) OR media[Text Word])) error and may be overestimated, they are 41 records obtained, May 9, 2006. a general measure for the very large scale of exposure from a population standpoint. Five of the studies reported on focus groups They also do not include viewings of movies conducted with adolescents;47–51 one was on as DVD releases or on television in the years focus groups and interviews with college following the theatre release dates. students;52 and one was on interviews conducted with a convenience sample of writers, actors, directors, producers, studio Effects on Attitudes, executives, and others involved in the film Beliefs, and Behavior: industry.53 Two additional relevant focus group studies were identified via citations in Movies other papers by MacFadyen and colleagues54 and the World Health Organization (WHO).55 Content analysis studies are useful for All of these studies used an acceptable documenting media inputs, but they qualitative research methodology. do not provide evidence concerning audience responses to such content. This Similar results concerning young people’s section reviews the results of research on interpretations of smoking imagery in audience responses to tobacco content in film have been found for focus group entertainment media. Most of the media- studies conducted with college students in effects research on tobacco in entertainment India (8 groups, N = approximately 50)52 media has focused on movies rather than and adolescents in Australia (16 groups, on other forms of entertainment media. N = 117),47 New Zealand (approximately This section focuses, therefore, on the 10 groups, N = 76;48 and approximately findings of that movie research. 10 groups, N = 88),49 India (8 groups, number not reported),55 and the United States Qualitative Studies (178 groups, N = 1,175;51 and 31 groups, N = 205).50 Young people reported that Researchers taking a cultural studies movies are an important source of approach to media research place a information about smoking and that these heavy emphasis on the subjectivity of images convey the notion that smoking interpretation of media messages. They tend is a normative, acceptable behavior; offers to use qualitative methods to investigate a means of stress relief; conveys a certain interpretations of media among small social image; and may serve as a marker numbers of audience members. These of adult independence. Together, these

376 M o n o g r a p h 1 9 . T h e R o l e o f t h e M e d i a

findings indicate that young people perceive that the findings were generalizable images of smoking in movies as leading to (see table 10.3 for summary scores of the positive social or personal consequences studies). On the basis of these criteria, rather than as presenting information two cross-sectional studies were excluded about the negative health consequences from the review5,6 because they included of smoking. Qualitative research further no controls for covariate influences. indicates that other mass media with a The remaining studies—seven published visual component (e.g., television, magazines) and one unpublished—involved four cross- convey mainly protobacco information about sectional analyses of three U.S. samples7,9,56,57 smoking to youth audiences (12 groups, and one unpublished Australian sample N = 70 approximately;54 and 178 groups, of adolescents.58 N = 1,175).51 As shown in table 10.3, researchers have Cross-Sectional Studies tended to use two general measures of movie influence. One assesses the smoking 4,9,56,58 Cross-sectional studies attempt to quantify status of favorite movie stars, and the 3,7,57 the relationship between exposure to other relies on movie title recognition. media and attitudes, beliefs, or behavior in The first measure, smoking status of favorite population-based samples. One unpublished movie stars, is an exposure measure that and eight published cross-sectional studies taps the self-concept and the prototypical of the relationship between exposure to smoker. People choose behaviors that are 59 smoking in movies and adolescent smoking consistent with their self-concepts. Self- were identified. Articles from the medical concept ratings of adolescent smokers, literature were identified through the as well as susceptible nonsmokers, are more following PubMed search strategies: similar to their ratings of the prototypical smoker than are the self-concept ratings 60–62 1. (“Smoking”[MeSH] OR “Tobacco”[MeSH]) of nonsmokers. In theory, adolescents AND “Motion Pictures”[MeSH], also may initiate behaviors as they modify 79 records obtained, May 10, 2006 their self-images. Behavioral depictions by favorite stars shape that process by 2. (“Smoking”[MeSH] OR “Tobacco”[MeSH]) determining what is “cool,” attractive, AND (“movie star” OR “movie stars”), and grown up. To the extent that smoking 5 records obtained, May 10, 2006 portrayals are consistent with adolescents’ actual or ideal self-images or a prototype of Articles from the literature on psychology, the ideal group member (that is, appearing marketing, and communications were grown up), adolescents will be motivated to identified by searching PsycINFO, using the smoke to align their self-perceptions with following search strategy and limiting to personal ideals.63,64 articles in English: In determining the smoking status of KW=(smoking or tobacco) and favorite stars, Distefan and colleagues4,56 KW=(movies or (motion picture), and Dixon58 asked adolescents to list their 26 records obtained, May 10, 2006 favorite male and female movie stars. The researchers developed lists of the top 10 The studies were reviewed for inclusion of male and female actors and subsequently design characteristics that increased the used content analysis to determine the on- reviewer’s confidence that the relationship screen smoking status for these individuals. demonstrated in the studies was a true The Distefan study also determined these media effect for the study sample and stars’ real-life smoking status. Other

377 1 0 . R o l e o f E n t e r t a i n m e n t M e d i a

Table 10.3 Summary of Results of Cross-Sectional and Longitudinal Studies: Smoking and Movies

Media influence Study Study design Recruitment Subjects Country measure Distefan et al. Cross-sectional Random digit dial 3,053 adolescents U.S. Chooses favorite 199956 aged 12–17 years movie star of ever (vs. never) smokers Dixon 200358 Cross-sectional School based 2,610 adolescents Australia Movie smoking aged 12–18 years; status of favorite attitudes assessed star among subgroup of 1,858 never/ experimental smokers Tickle and Cross-sectional School based 632 adolescents U.S. Movie smoking Sargent 20019 aged 10–19 years; status of favorite attitudes assessed star among subgroup of 281 never smokers Sargent and Cross-sectional School based 4,919 adolescents U.S. Two-stage direct Beach 20017 aged 10–15 years; measure (movie Sargent et al. attitudes assessed title recog × amt 200257 among subgroup of of smoking) 3,766 never smokers

Sargent et al. Cross-sectional Random digit dial 6,522 adolescents U.S. (national Two-stage direct 200565 aged 10–14 years sample) measure (movie title recog × amt of smoking)

McCool et al. Cross-sectional School based 3,041 adolescents New Zealand Perceived 200566 aged 12–16 years frequency of viewing films (cinema and video)

Dalton et al. Longitudinal School-based 2,603 adolescents U.S. Two-stage direct 20033 recruitment with aged 10–15 years measure (movie teleph F/U at inception title recog × amt of smoking)

Distefan and Longitudinal Random digit dial 2,084 adolescents U.S. Movie smoking Pierce 200467 aged 12–17 years status of favorite at inception star Note. Teleph F/U = telephone follow-up; recog = recognition; amt = amount; S = sociodemographics; P = personality characteristics; Sch = school attachment and function; SI = other social influences (friend and family smoking); PS = parenting style; M = other media/advertising influences. aStatistically significant relation (p < .05) between movie smoking exposure and this outcome after covariate adjustment. bSignificant correlation (no covariate adjustment).

378 M o n o g r a p h 1 9 . T h e R o l e o f t h e M e d i a

Additional Smoking Covariate outcome outcome Measure of Association adjustment Validity, reliability measures measure association size categories Not reported Susceptibilitya 0 0 0 S, P, Sch, SI, M

Not reported Intentions Index Adjusted 1.16a S, Sch, SI proportional odds

Not reported Susceptibilitya Initiation Adjusted odds 1.5a S, Sch, SI, M

3-week test–retest (average Susceptibilitya Initiation Adjusted odds 1.7–2.7a S, P, Sch, PS, percent agreement) 92%. Norms—adulta SI, M Correct recall of titles seen up Norms—peer a to 1 year prior = 90%. Positive expect Recalls having seen a sham title 3%. 3-week test–retest (average 0 Initiation Adjusted odds 1.7–2.6 S, P, Sch, PS, SI percent agreement) 92%. Correct recall of titles seen up to 1 year prior = 90%. Recalls having seen a sham title <2%. Cronbach’s alpha = 0.65 Norms—moviesa 0 0 0 S Nonchalance— moviesa Norms—peerb Judgment—peera Intentions 3-week test–retest (average 0 Initiation Adjusted relative 2.0–2.7 S, P, Sch, PS, percent agreement) 92%. risk SI, M Correct recall of titles seen up to 1 year prior = 90%. Recalls having seen a sham title 3%. Not reported 0 Initiation Adjusted odds 1.3a S, Sch, PS, SI, M

379 1 0 . R o l e o f E n t e r t a i n m e n t M e d i a

researchers9 asked adolescents to name and intentions to smoke in a sample of their favorite stars and determined smoking Australian adolescent never smokers and status in recently released movies for any experimental smokers. star chosen by five or more adolescents. One problem with favorite star measures It is unclear whether the lack of association was the loss of sample size due to the for intentions observed in Dixon’s study great diversity of stars adolescents chose in contrast to the U.S. studies is due to as “favorite.” Adolescents were excluded if a cultural difference in responsiveness their chosen star did not make the top 10 to on-screen smoking by stars or due list—51% were excluded in the Distefan to methodological differences between study,56 and 37% were excluded by Dixon58— the studies. For example, the Australian or because fewer than five adolescents chose adolescents in Dixon’s study may have been the star (50% excluded in a study by Tickle less susceptible to the influence of smoking and colleagues).9 in movies because it did not resonate with their other media exposure in relation to All studies have examined associations tobacco. Unlike in the United States, most between stars’ on-screen smoking status direct forms of tobacco advertising are and adolescents’ attitudes toward smoking. illegal in Australia. Cross-cultural surveys Two used an adolescent smoking measure using identical methods would be necessary termed susceptibility to smoking, which to test these hypotheses. captures an individual’s inability to rule out smoking in the future or to rule out Two studies9,58 also examined whether smoking if a peer offers cigarettes; this the smoking status of favorite stars was measure has been found to be a strong linked with adolescent smoking. Overall, predictor of future smoking.68 Distefan and the relationship between favorite stars’ colleagues56 determined the favorite movie smoking and adolescent smoking was stars for a random sample of California statistically significant in both cases. Dixon58 adolescent smokers. They found that estimated the effect on a smoking uptake adolescent never smokers who preferred the index with a proportional odds model favorite star of smokers were more likely (adjusted proportional OR = 1.16). Tickle to be susceptible to smoking. The favorite and colleagues9 estimated the effect on stars of smokers also were more likely to trying smoking with a logistic regression have smoked on screen and in real life. (adjusted OR = 1.5 [95% confidence Tickle and colleagues9 determined favorite interval (CI), 1.01–2.32] for adolescents movie stars for a school-based sample of whose favorite stars smoked in two recent northern New England adolescents. Among movies and 3.1 [95% CI, 1.34–7.12] for never smokers, those choosing stars who adolescents whose favorite stars smoked smoked were significantly more likely to be in three or more movies). Dixon separated susceptible to smoking. For each of these the effect by whether the favorite actor studies, the adjusted odds ratio (OR) was the was male or female and the gender of the measure of association with smoking and subject. She found that the association was susceptibility to smoking. For the study by significant for male actors’ smoking, and Distefan and colleagues, the adjusted OR was only in girls. Tickle and colleagues found 1.3 for adolescents who chose a favorite star no such gender-based interactions. among smokers. For the study by Tickle and others, the adjusted OR was 4.8 if the star The second approach to measuring exposure had smoked in two or more recent movies. to smoking in movies is a two-stage method Dixon58 found no relationship between the that directly estimates exposure to smoking on-screen smoking status of favorite stars in movies.3,7 The first stage involves

380 M o n o g r a p h 1 9 . T h e R o l e o f t h e M e d i a

content analysis to determine the amount who participated in their longitudinal pilot of smoking contained in the movie sample study. As part of the pilot, students were of interest. Because adolescents cannot be called once a month for 12 months; they surveyed on all movies, the second stage were asked at each interview what movies of this method requires special survey they had seen in the past week. One year techniques that present the adolescent with after the final interview, adolescents were a movie title list (Sargent and colleagues7 asked whether or not they had seen items on chose to include 50 titles) that was randomly a list of 50 movies. Each list contained up selected from the larger content-analyzed to 30 movie titles they reported having seen sample (table 10.4). This method has the the previous year (average = 19), 10 false advantage that exposure to smoking in movie titles with real stars, 10 false movie movies can be estimated directly and in an titles with false stars, and other real movie unbiased fashion for all adolescents in the titles to complete a list of 50. As shown survey sample. in table 10.4, adolescents had excellent recognition of the movies they had seen The method relies on adolescents’ ability to and were very unlikely to report seeing recall accurately whether or not they had false movies, even when associated with seen a movie, when prompted by the movie real actors. title, and has been extensively validated by Sargent and colleagues.65 As a test of Sargent and colleagues57 used the direct face validity, these researchers evaluated method described above to estimate whether box-office success was related to the exposure to smoking in movies from a probability adolescents would say they had sample of 601 popular contemporary movies seen a movie. In their cross-sectional study, among 4,919 adolescents in northern there was a high correlation (r = –0.73) New England. The movie exposure measure between the box-office success of the provided an estimate of lifetime exposure top 100 movies released the year before the to smoking scenes from the 601 movies. survey and the percentage of adolescents The subjects had seen an average of 30% who had seen these films. Two of the movies of the movie sample; in these, they were included were foreign films not released in exposed to an average of 1,160 depictions the United States and served as a validation of smoking in movies (interquartile range against false reports. Of the students queried 640–1,970).69 A smoothed curve for the dose regarding the two foreign films, only about response shows a direct linear relationship 1% or less reported that they had seen the between higher exposure to smoking in unreleased movies. These were the two movies and higher rate of smoking through lowest viewing rates reported for the survey. most of the exposure range, with the dose To further evaluate validity, Sargent and response flattening out past the 95th colleagues7 recontacted the 49 students percentile of exposure (figure 10.2).

Table 10.4 Validity of Adolescents’ Recognition of Movie Titles Have you seen this movie? (ascertained in 2001) Movie category Yes No Don’t know Adolescent reported seeing it in 1999 87.2% 12.6% 0.6% False movie title, real actors 2.7% 96.7% 0.5% False movie title, false actors 3.0% 96.4% 0.6% Other movies 41.1% 54.2% 4.6% Note. Data derived from research by Sargent, J. D., M. O. Beach, M. A. Dalton, and T. F. Heatherton.

381 1 0 . R o l e o f E n t e r t a i n m e n t M e d i a

Figure 10.2 Lowess Smoothed Curve Showing Cross-Sectional Relationship between Exposure to Movie Smoking Depictions and Adolescent Smoking Initiation in a Study of Northern New England Adolescents

1.0

17.3% of the sample had tried smoking 0.8

0.6 95th percentile for exposure

0.4

Smoothed point estimate: 0.2 Prevalence of ever tried smoking

0 0 2000 4000 6000 Estimated exposure to movie smoking depictions* Note. Based on sample described in Sargent, J. D., M. L. Beach, M. A. Dalton, L. A. Mott, J. J. Tickle, M. B. Ahrens, and T. F. Heatherton. 2001. Effect of seeing tobacco use in films on trying smoking among adolescents: Cross sectional study. British Medical Journal 323 (7326): 1394–97. *From 601 popular contemporary motion pictures.

There was almost no smoking among sample.57 Exposure to smoking in movies adolescents with little exposure to movies, was associated with susceptibility to and smoking peaked at almost 40% above smoking, an indexed measure of positive the 95th percentile. The relationship expectations for smoking, and normative between viewing smoking in movies and beliefs about adult smoking. The measure adolescent smoking remained after a broad of association was the adjusted OR. Ranges range of confounders was controlled.57 (for the three higher quartiles) for the The measure of association was the adjusted effect size for the association with exposure OR, with the adjusted odds of trying smoking to smoking in movies were 1.2–1.7 for being 1.9 (95% CI, 1.3–2.7), 2.6 (1.8–3.7), susceptibility to smoking, 1.2–1.4 for the and 2.5 (1.7–3.5) for quartiles 2, 3, and endorsement of adult smoking as normative, 4, respectively, compared with quartile 1. and 1.2–1.4 for the endorsement of positive The effect of moving to a higher category of smoking expectations. Exposure to smoking exposure to smoking in movies was similar in movies was not associated with normative to the adjusted OR for having siblings who beliefs about peer smoking, a finding that smoke (1.7 [95% CI, 1.3–2.1]); the effect was is consistent with the predominantly adult higher than the effect of having parents who nature of depictions of smoking in movies. smoke (1.3 [95% CI, 1.1–1.6]) or owning This finding is consistent with content tobacco-branded merchandise (1.2 [95% CI, analyses showing that movies rarely depict 0.97–1.5]) and lower than the effect of having adolescent characters as smokers.32 peers who smoked (5.1 [95% CI, 4.0–6.4]). Sargent and colleagues65 used the direct The relationship between exposure to method described above to estimate smoking in movies and attitudes toward exposure to smoking in movies from a smoking also was assessed among never sample of 532 popular contemporary movies smokers in the northern New England among a nationally representative sample of

382 M o n o g r a p h 1 9 . T h e R o l e o f t h e M e d i a

6,522 U.S. adolescents. Adolescents’ level of exposure to movies had an indirect effect exposure to smoking in movies was divided on intentions, through its influence on into quartiles. Compared with adolescents mediating cognitions. Thus, this study, like in quartile 1, the adjusted ORs for having that of Dixon,58 failed to find a statistically tried smoking were 1.7 (95% CI, 1.1–2.7) significant association between the movie for quartile 2, 1.8 (95% CI, 1.2–2.9) for exposure measure and smoking intentions. quartile 3, and 2.6 (95% CI, 1.7–4.1) for Owing to differing methods in the studies, quartile 4 after controlling for potential it is not clear whether the lack of association confounders. This association between observed with intentions is because on- exposure to smoking in movies and screen smoking does not directly affect smoking initiation was similar in size to the smoking intentions, whether the two studies association with parent and sibling smoking that examined intentions used measures of (adjusted odds of smoking 1.8 [95% CI, exposure to media that lacked specificity in 1.5–2.3] and 2.3 [95% CI, 1.8–2.9], quantifying actual exposure to on-screen respectively) and held true within broad smoking, or whether the tobacco control racial and ethnic categories, and regardless environments in those countries (Australia of residential location. The association and New Zealand) “dampen down” the was lower than the association with peer protobacco effects of on-screen smoking. smoking (OR 3.3 [95% CI, 2.6–4.2]). Intercountry surveys that use identical An adjusted attributable risk fraction methods (including more direct measures indicated that among 38% of adolescents of on-screen smoking) would be necessary who had tried smoking, exposure to to test these hypotheses. smoking was an independent, primary risk factor for smoking initiation. The cross-sectional surveys not included (because of the lack of controls for In addition to the measures of smoking confounding) are still interesting, status and movie title recognition, a because they suggest that an association third measure of movie influence—used between exposure to smoking in movies in a single study—asked adolescents and youths’ smoking also occurs in non- their perceived frequency of viewing Western countries. However, because of movies. Using this crude estimate of the limitations of these studies, further exposure to on-screen smoking, McCool research is needed to establish more and colleagues66 examined a sample of clearly the effect of smoking depicted in 3,041 New Zealand adolescents. The self- movies on adolescents in non-Western reported frequency of movie exposure countries. A survey of 1,338 Thai adolescents was positively associated with perceived (aged 14–17 years) found that exposure to smoking prevalence among adolescents American movies was related to heightened and among people in movies, and with levels of smoking-related behavior but not nonchalance/apathy concerning smoking to smoking intentions.6 In addition, a survey in films, when controlling for demographic of more than 1,700 Hong Kong adolescents variables. These researchers did not find a indicated that viewing a greater number statistically significant association between of movies was significantly associated with exposure to film and smoking intentions being more likely to have ever smoked and (“smoking expectations”). However, path with intentions to smoke.5 analytic techniques revealed that certain smoking belief variables that bore a direct Longitudinal Studies association with movie exposure also were significantly associated with smoking Longitudinal studies attempt to quantify intentions, leading the authors to argue that the relationship between exposure to

383 1 0 . R o l e o f E n t e r t a i n m e n t M e d i a

media and behavior in population-based directly.3 Figure 10.3 shows a smoothed samples by using multiple-wave survey curve for the dose response. As shown in the designs. These studies have the advantage cross-sectional sample, there was a direct of determining more clearly whether the linear relation between higher exposure exposure precedes the adoption of the to smoking in movies and a higher rate behavior. Never smokers in two U.S. samples of smoking through most of the exposure were followed longitudinally to determine range. The dose response flattened past the which persons initiated smoking in the 95th percentile of exposure. Smoking during future as a function of baseline movie follow-up was almost zero for adolescents exposure.3,4 A longitudinal study published with minimal exposure to smoking in in 2004 examined the status of smoking in movies at baseline and approached 20% for movies by favorite stars (assessed at baseline) adolescents in the highest exposure range. as a predictor of trying smoking in the future.4 This study identified “favorite stars” The effect persisted when controlling who smoked in at least two movies during for a large set of covariates, including the three-year period prior to the survey. other social influences, advertising Consistent with Dixon’s cross-sectional influences, personality characteristics study,58 female, but not male, adolescents (e.g., rebelliousness), and parenting style. who chose stars who were smokers were The effect size, measured as adjusted relative significantly more likely to initiate smoking risk of smoking initiation, with baseline during the follow-up period. movie exposure categorized into quartiles, was 2.0 (95% CI, 1.3–3.2), 2.2 (95% CI, Initiation of smoking also was determined 1.4–3.4), and 2.7 (95% CI, 1.7–4.3) for for never smokers in the study of northern quartiles 2, 3, and 4, respectively, compared New England adolescents in which exposure with quartile 1. This range of relative risks to smoking in movies was estimated was similar in magnitude to the relative

Figure 10.3 Lowess Smoothed Curve Showing the Longitudinal Relationship between Exposure to Movie Smoking Depictions and Adolescent Smoking Initiation in a Study of Northern New England Adolescents

10% of the sample tried smoking during follow-up period 0.8

0.6

0.4 95th percentile for exposure Smoothed point estimate: point Smoothed

0.2

0 incidence of smoking initiation during follow-up follow-up during initiation smoking of incidence 0 1000 2000 3000 4000 5000 Estimated baseline exposure to movie depictions* Note. Based on sample described in Dalton, M. A., J. D. Sargent, M. L. Beach, L. Titus-Ernstoff, J. J. Gibson, M. B. Ahrens, J. J. Tickle, and T. F. Heatherton. 2003. Effect of viewing smoking in movies on adolescent smoking initiation: A cohort study. Lancet 362 (9380): 281–85. *From 601 popular contemporary motion pictures.

384 M o n o g r a p h 1 9 . T h e R o l e o f t h e M e d i a

risk of smoking associated with having versus nonsmoking footage) to be parents who smoke (1.6 [95% CI, 1.2–2.0]), manipulated and allows controlled and higher than the relative risk associated assessment of audience reactions to such with friends’ smoking (1.1 [95% CI, content. This method overcomes a key 0.87–1.5]) or ownership of tobacco-branded limitation of cross-sectional studies— merchandise (1.1 [95% CI, 0.85–1.5]). It is the inability to control for unknown or also notable that the estimates of the effect unmeasured confounders. In experimental of viewing smoking in movies on smoking studies, randomization of subjects to initiation in both longitudinal studies were exposure categories is used to control almost identical to estimates obtained for for known and unknown confounders. the cross-sectional samples. This finding The limitations of experimental studies suggests that exposure to smoking in are that the viewing conditions tend to be movies and its effect on adolescent smoking nonnaturalistic and it generally is feasible persist over time. to assess only short-term responses to relatively brief media exposure. Nonetheless, Taken together, these cross-sectional these studies complement the cross- and longitudinal studies provide strong sectional studies and provide further support for a direct association between insights into the impact on audiences of exposure to smoking in movies and attitudes movie depictions of tobacco and tobacco use. toward smoking and smoking initiation. The cross-sectional study of attitudes among The PubMed and PsycINFO searches never smokers57 suggests that exposure to reported under cross-sectional studies smoking in movies enhances perceptions yielded two experimental studies42,70 about the utility of smoking and increases and two quasi-experimental studies71,72 adolescents’ intentions to try smoking. assessing reactions to depictions of tobacco The longitudinal studies provide evidence in movies. The latter two studies are best of a temporal association—that is, exposure classified as quasi-experimental, as they to on-screen smoking precedes smoking assessed naturalistic exposure to whole behavior among adolescents. The strongest movies among actual cinema audiences.71,72 associations have been demonstrated in The strength of these studies was their studies using a direct measure of exposure. larger audience sample size relative to Cigarette smoking by a favorite movie star the other studies. Their limitation was has a weaker association, probably because that viewers were not randomly allocated tobacco use by favorite stars is not a true to conditions. The authors identified two measure of exposure to all smoking depicted further peer-reviewed experimental studies: in movies but instead taps the much one published73 and another conducted as narrower effect mediated by the adolescent’s part of a doctoral dissertation.58 identification with his or her favorite star. If this is the case, the gender findings in Table 10.5 summarizes the methods and the studies by Dixon58 and Distefan and findings of the respective experimental colleagues4 indicate that, in relation to studies assessing reactions to on-screen movies, identification processes are more portrayals of tobacco. Most designs of important in determining smoking onset the studies included an experimental for girls than they are for boys. manipulation that compared audience responses to movie footage depicting Experimental Studies smoking (intervention) with responses to movie footage that did not depict smoking Experimental research enables media (control). Some studies included further content variables of interest (e.g., smoking experimental manipulations, such as varying

385 1 0 . R o l e o f E n t e r t a i n m e n t M e d i a

Table 10.5 Summary of the Methods and Results of Experimental Studies Assessing Responses to On-Screen Tobacco Use Methods

Subgroups Stimulus Experimental Study Subjects examined Country movie manipulation Jones and Carroll 51 college students n = 40 females, Australia Video clips (role Smoking compared with 199873 n = 11 males plays, not actual nonsmoking footage movie footage) Pechmann and Shih 607 ninth graders, — US Scenes from Smoking compared with 1999 (study 1)42 nonsmokers Reality Bites and nonsmoking footage x high Wild at Heart compared with low positive arousal elicited by scenes Pechmann and Shih 232 ninth graders, — US Whole movie Smoking compared with 1999 (study 2)42 nonsmokers Reality Bites nonsmoking footage x prefilm antismoking advertisement compared with no advertisement Gibson and Maurer 120 college n = 36 smokers, US 20-minute clip of Smoking, nonsmoking 200070 students n = 84 nonsmokers Die Hard footage Hines et al. 200074 151 college — US 6 scenes from Smoking compared with students 6 popular films nonsmoking footage Dixon et al. 200171 383 adult cinema n = 192 who Australia Whole movie Antitobacco message patronsa completed follow-up (The Insider) compared with interview within control film (Erin Brokovich) 2 weeks of seeing movie Edwards et al. 2,038 female n = 186 smokers, Australia Whole movies Prefilm antismoking 200472 adolescent cinema n = 1,852 (depicting smoking) advertisement compared with patronsa nonsmokers no advertisement Dixon 200358 374 seventh and — Australia 2 x 5 minute clips Smoking compared with eighth graders from popular movies nonsmoking footage of different character types Note. – = variable not assessed; ns = variable not significantly affected by experimental manipulation. aQuasi-experimental study, using subject’s self-selected cinema exposure. *p < .05. **p < .01. ***p < .001. Variable significantly affected by experimental manipulation (lowest p value achieved for variables in this response category).

the level of emotional arousal for the sample (intervention) produced a different audience movie footage (study 1)42 or varying the response than the response to viewing a social characteristics of the characters in movie that did not contain such content the movie footage.58 Two studies assessed (control). Most of the studies used actual whether exposure to an antismoking movie footage or whole movies for their advertisement (intervention) before stimulus material, often with some editing viewing a movie that featured smoking performed to achieve the experimental promoted different audience responses manipulation. The exception, the study compared with responses to viewing a movie by Jones and Carroll,73 used video clips without such an advertisement (control). of role plays produced specifically for the One study assessed whether including study. For studies using actual movie antitobacco content within the movie71 footage as stimuli, the strength is that

386 M o n o g r a p h 1 9 . T h e R o l e o f t h e M e d i a

Response variables Beliefs Beliefs Beliefs about Personal about the Ratings of Ratings of Ratings of about personally intentions tobacco the movie characters actors smokers smoking to smoke Arousal industry — ** (females) — — — — — — ns (males)

— — — ** * — ** —

* ** — ** * * ** —

— * (smokers) * (smokers) * (nonsmokers) — ns (nonsmokers) — — ns (nonsmokers) ns (nonsmokers) — *** — — — * — —

— — — — — * (completed — *** follow-up within 2 weeks of movie) ns (smokers) — — — — * (smokers) — — *** (nonsmokers) ns (nonsmokers)

— ** — ns * ns — —

the stimuli represent those the viewers the footage is of nonprofessional quality, might be exposed to in the “real world.” limiting generalization of the results to the The disadvantage of this method is that likely effects on audiences of “real world” to achieve the intended experimental movie viewing. manipulation (e.g., smoking versus nonsmoking footage), it is not always Most of the studies consisted of a posttest­ possible to obtain directly comparable only design in their assessment of the control footage.58 Conversely, studies audience’s tobacco-related attitudes, beliefs, using nonprofessionally produced footage and intentions. Only one71 used pretest and can more readily produce stimuli that posttest assessments of smoking-related are identical, with the exception of the beliefs, which would have increased the experimental manipulation.73 However, power to detect the effects of the media

387 1 0 . R o l e o f E n t e r t a i n m e n t M e d i a

manipulation within subject analyses. reactions to smoking in films, power However, several of the studies did include a calculations were performed, using the pretest assessment of participants’ smoking results observed in studies in which status and demographic characteristics. significant effects of the experimental This information enabled examination of manipulations were found on smoking- responses as a function of key audience related beliefs and intentions, with the use subgroups or inclusion of these variables as of Power and Precision software. To perform covariates in data analyses.70,74 such calculations comparing mean response scores postintervention, it was necessary The main methodological difference to specify means, standard deviations, and between the studies related to their cell sizes for each experimental condition. respective sample sizes. The smallest This process was possible for all of the audience sample consisted of approximately experimental studies, except for two that did 40 subjects, with about 20 viewers per not publish standard deviations with their condition.73 The largest audience sample results.42,70 The effect sizes achieved were consisted of 2,038 subjects, with about within a range similar to those observed 1,000 viewers per condition.72 Despite in the above meta-analyses of media these marked differences in sample size, experiments on other health topics (absolute even the smaller studies found some values 0.1 through 0.8). The strongest effect statistically significant effects of the size, 0.8 (95% CI, 0.41–1.19), was observed experimental manipulation on viewers’ in the study by Jones and Carroll73 for the responses. effects of a video character’s on-screen smoking status on perceptions of that To help inform the assessment of the effect character’s social characteristics. According sizes of these experimental studies, the to Cohen’s77 effect size conventions, this authors examined meta-analyses of effect observation would be viewed as a “large” sizes observed in experimental research effect for social science research. The effect assessing the effects of violent media sizes observed for more self-referent beliefs depictions on viewer aggression75 and of about smoking (e.g., intentions) tended thin media models on body dissatisfaction.76 to be “small” (range: 0.1–0.3), as might be The meta-analysis of media violence studies expected for studies assessing reactions found a mean effect size for laboratory to a brief media exposure. However, it experiments of approximately 0.25 (95% CI, is theoretically plausible that recurrent, 0.23–0.28) and for field experiments naturalistic exposure to movie images of approximately 0.2 (95% CI, 0.15–0.25). smoking have a larger cumulative effect The absolute values for effect sizes in on viewers’ propensity to smoke, and the the body image studies were of a similar findings of cohort studies3,4 are consistent magnitude. The mean effect size across with this hypothesis. studies was –0.31 (95% CI, −0.40 to –0.23). (The positive direction of the effect in the Effects of On-Screen Smoking on violence studies reflects increased aggression following exposure to violent movie content. Viewers’ Smoking-Related Beliefs The negative direction of the effect in the body image studies reflects more negative Theories of media influence predict that body image perceptions following exposure role models bearing favored social attributes to thin models in the media.) are likely to be especially persuasive.44,45 Several experimental studies have assessed To determine the effect sizes observed in whether stars who smoke on screen promote experimental research assessing audience prosmoking beliefs among audiences.42,70,74

388 M o n o g r a p h 1 9 . T h e R o l e o f t h e M e d i a

Results of experimental studies suggest that Exposure to on-screen smoking also has viewing movie characters who are smoking been found to influence viewers’ beliefs enhances viewers’ perceptions of how about the social consequences of personal socially acceptable smoking is. Pechmann smoking. Pechmann and Shih42 digitally and Shih42 found that exposure to movie changed the image frame to edit smoking scenes of popular, young stars smoking out of the 1990s film Reality Bites. (versus nonsmoking) prompted adolescent Comparing adolescents’ responses to the viewers to report that adolescent smokers original versus the nonsmoking version of had higher social stature. This finding the movie, they found that adolescent never was replicated in a second experiment smokers exposed to the original version that assessed reactions to a whole movie showed enhanced perceptions of how their (Reality Bites) depicting smoking compared social stature would be viewed by others if with an edited version of the movie that they were to personally smoke. The video excluded smoking depictions. Similarly, manipulation had no significant effects on Gibson and Maurer70 found that, among participating adolescents’ perceptions of nonsmoking college students, viewing how popular, vital, or poised they would a movie clip of a leading male character look if they were to smoke. Dixon58 found smoking (versus a comparable clip in which that beliefs about the social consequences this character does not smoke) resulted of personal smoking were affected in a greater willingness to become friends differentially, depending on the social with a smoker. However, further analyses characteristics of the on-screen smoker. revealed that this effect was most marked Among adolescent viewers, attractive, high- for viewers low on “need for cognition” status characters who smoked on screen (a trait predicted to render someone more promoted positive beliefs about the benefits susceptible to persuasion via the peripheral of smoking. However, unattractive, low- route).78 This finding suggests that some status characters who smoked on screen people may be more susceptible than others detracted from such beliefs. to the persuasive impact of movie depictions of smoking. Pechmann and Shih42 also found that exposure to the original version of Dixon58 found evidence suggesting that Reality Bites promoted increased personal adolescents who watched footage of movie intentions to smoke among adolescent adult characters smoking on screen never smokers. For older viewers, two perceived adult smoking prevalence studies (with sample sizes of 150 or in the “real world” to be higher than more) found a significant effect of on- did adolescents who watched footage screen tobacco depictions on personal of nonsmoking movie characters. This intentions to smoke.71,74 However, another effect occurred irrespective of the social study (examining a smaller subgroup of characteristics of the on-screen smokers 84 nonsmokers) did not find such an effect.70 that students viewed. Together, these Hines and colleagues74 found that college findings suggest that movie depictions of students who viewed movie scenes in which smoking may promote perceptions that the main characters smoke were more likely smoking is a normative behavior in the than those who viewed nonsmoking scenes real world. These findings are of concern, to indicate a likelihood to smoke in various since social learning variables, “especially situations in which smoking is likely to peer smoking and approval, prevalence occur. This effect persisted with controls estimates, and offers/availability”79(p.1171) for the smoking status of the participant. have been found to be strongly predictive Furthermore, among male viewers who of smoking onset. were regular or occasional smokers, the

389 1 0 . R o l e o f E n t e r t a i n m e n t M e d i a

smoking film footage also promoted a Pechmann and Shih42 found that showing higher current desire to smoke. In contrast, youth an antismoking advertisement the study by Gibson and Maurer,70 with less immediately before viewing a movie statistical power, found that nonsmoking depicting popular young stars smoking college students were no more likely to inoculated them against the prosmoking report intentions to smoke in the future influence of the movie footage. The after exposure to movie footage of a leading advertisement also generated more character smoking (versus nonsmoking). negative thoughts toward the leading However, the direction of the trend in the movie characters, but it did not detract overall cell means was toward smoking from the ratings of the movie’s overall scenes promoting slightly higher scores on action or storyline, or from the likelihood intentions. Because the sample size for this of recommending it to a friend. In fact, analysis was small (N = 84), it is likely that those who saw a movie preceded by an this study had insufficient power to detect a antismoking advertisement rated the small or moderate effect size, if it existed. movie storyline more favorably than those who saw a movie without such an Dixon and colleagues71 found that viewing a advertisement. These findings are of great movie that portrayed the tobacco industry in practical importance in providing evidence a negative light and included information on concerning the efficacy of one possible the negative health consequences of smoking strategy for reducing the negative impact within the story (The Insider) promoted on-screen smoking has on youth audiences. a short-term reduction in intentions to That is, screening an antismoking smoke among adult smokers and former advertisement before the movie immunized smokers. Content analyses suggest that young viewers against the prosmoking portrayal of information about the negative effects of the movie, without detracting from health consequences of smoking is a rare their overall enjoyment of the movie. phenomenon. Experimental research indicates, however, that inclusion of such This approach was subsequently evaluated information in a movie can promote an using a quasi-experimental study of antitobacco message. Dixon and colleagues71 2,037 female adolescent moviegoers in also found that viewing The Insider Australia who had self-selected to see movies promoted more negative views among depicting smoking.72 The intervention group audience members of the tobacco industry’s who viewed an antismoking advertisement business conduct. These results have some before the movie was compared with parallels with findings of evaluations of a control group who did not view an public responses to antitobacco media antismoking advertisement screened campaigns exposing industry manipulation. before the movie. Among nonsmoking Surveys indicate that cigarette consumption viewers, those who saw an antismoking declined in association with California’s advertisement before the movie showed Proposition 99 media campaign.80 Moreover, stronger disapproval of smoking by evaluation results for Florida’s “truth” characters in the movie. Among viewers campaign advertisements show evidence of a who were current smokers, those who saw decline in youth smoking and a relationship the antismoking advertisement showed between youth smoking behavior and significantly reduced intentions for future changes in youth attitudes toward the smoking. The antismoking advertisement tobacco industry’s manipulation.81 did not affect nonsmokers’ intentions to Chapter 12 on the effectiveness of mass smoke. Most nonsmoking subjects (95%) media in discouraging smoking includes in both conditions reported they were details of these antismoking campaigns. unlikely to be smoking at this time next

390 M o n o g r a p h 1 9 . T h e R o l e o f t h e M e d i a

year. The results of these two studies young female smoking rated her as more suggest that screening antismoking outgoing, more sophisticated, not as easy advertisements before movies depicting to manipulate, and less emotional about smoking is an effective strategy for reducing breaking up with her boyfriend than those the prosmoking persuasive effect of on- women who viewed a control video in screen tobacco use by movie stars. which the young female did not smoke. In a study examining reactions to different Effects of Smoking Depictions on movie character depictions of smokers, Dixon58 found that adolescents associated General Reactions to Movies smoking by female antagonists with low social status. Ratings of the male characters In discussing audience reactions to smoking did not differ in this way. Together, these in movies, it also is relevant to examine results suggest that audience members responses from the perspective of audiences’ may identify more with movie characters entertainment experience. Evidence is mixed of similar smoking status. Moreover, as to whether audience perceptions of movie on-screen smoking by female characters characters are affected by their on-screen appears to carry some negative social smoking. Pechmann and Shih42 found connotations. that, among adolescent never smokers, there were no significant differences in the Pechmann and Shih42 found that, in more number of negative, neutral, or positive general reactions to on-screen smoking, thoughts about the leading characters in viewing movie scenes depicting smoking a movie as a function of whether scenes evoked higher levels of positive arousal of their smoking were viewed. Similarly, than did viewing similar scenes without Gibson and Maurer70 found that, among smoking. Despite the effects of smoking college students who were nonsmokers, on viewers’ emotional arousal, Pechmann viewing movie scenes of a leading male and Shih42 found that adolescents’ ratings character smoking (versus nonsmoking) of a movie’s action or storyline or their did not markedly affect their ratings of that willingness to recommend the movie to character. However, among college students friends was no different for a version of who were smokers, viewing such movie the movie that edited the smoking out scenes led them to rate the male actor and of the scene, compared with the original the character he played as more likeable version of the movie. This finding has when he smoked, compared with when he relevance to filmmakers in suggesting was not depicted as a smoker. Reactions that excluding smoking from films does appear to vary, however, depending on the not detract from their overall appeal. movie character’s gender—smoking by This argument is further corroborated by females may be associated with negative Dalton and colleagues.32 They found that character traits. Hines and colleagues74 the amount of tobacco use depicted in found that female characters depicted as movies is not significantly associated with smokers were rated less favorably on a range box-office success. Pechmann and Shih42 of social characteristics (e.g., attractive, also found that, for adolescent viewers who sexy, popular), but they found no such were shown an antismoking advertisement effects for male characters. Smoking before viewing a movie depicting smoking, by female characters also led audience the effect of smoking depictions in the members who were occasional smokers or movie on arousal, perceptions of a smoker’s nonsmokers to perceive themselves as less social stature, and personal intent to smoke similar to the character. Jones and Carroll73 were eliminated. This finding and those found that young women who viewed a of Edwards and colleagues72 imply that

391 1 0 . R o l e o f E n t e r t a i n m e n t M e d i a

showing antismoking advertisements before movies with smoking could modify the Tobacco Content in effect of prosmoking movie depictions on Other Media the audience’s smoking behavior. Television Conclusions Concerning Media Effects Research Television began a close relationship to the tobacco industry in the 1950s. As it The findings from experimental studies became clear that smoking was a cause of contribute to the understanding of how cancer, and with the elimination of cigarette vicarious learning effects may occur in advertising in the broadcast media in 1971, response to smoking behavior symbolically tobacco use also dropped out of network modeled in movies. Along with the television in the United States. This resulted, results of cross-sectional and longitudinal in part, from the Public Airways Act.82 population-based studies, experimental research indicates that images of smoking Several authors have analyzed content in film can influence people’s beliefs samples of prime time television about social norms for smoking, beliefs programming for smoking depictions. Breed about the function and consequences of and De Foe’s83 content analysis of prime smoking, and ultimately their personal time U.S. television dramas and situation propensity to smoke. Certain movie comedies produced between 1950 and 1982 depictions may be more likely than others found a steady drop in the use of cigarettes to promote prosmoking beliefs. Audience over the three decades. In the period before members’ responsiveness to such imagery the release of the first Surgeon General’s may vary as a function of their personal report (1950–63), nine times more characteristics (especially smoking status cigarettes were used per hour than for the and gender). Experimental studies found season 18 years later. Several authors have many statistically significant effects—of a found that television smoking is more similar magnitude to the effects observed common in dramas than in other genres.83,84 in experimental media research on other Table 10.6 lists the number of smoking acts health topics—for only brief exposure to per hour observed in samples of television movie images of smoking. dramas selected for content analyses of television programming. The studies used Across the different study designs used similar coding methods but differed slightly to assess audience responses to on- in their methods of sampling television screen tobacco use, there is considerable content. Taken together, the results suggest convergence in findings. Protobacco that the rate of smoking in prime time film content has been found to promote television dramas declined dramatically prosmoking beliefs and intentions in both from 1950–63 (4.52 smoking acts per hour) experimental and cross-sectional studies. to 1981–83 (0.35 smoking acts per hour). Exposure to on-screen smoking has been However, studies conducted in 1984 and associated with smoking behavior in cross- 1993 found slightly higher smoking rates sectional studies and predictive of smoking (1.01 and 1.20 smoking acts per hour, behavior in longitudinal studies. A similar respectively). A content analysis of television convergence of findings across different drama aired on Japanese television between study types was observed in a meta-analysis 1995 and 1996, however, found a rate of examining the effects of media violence smoking depiction (4.22 per hour) similar to on aggression.75 that found on U.S. television in the 1950s.85

392 M o n o g r a p h 1 9 . T h e R o l e o f t h e M e d i a

Table 10.6 Number of Smoking Acts per Hour of Television Drama for Different Content Analysis Studies Conducted in the United States Year of programming Smoking acts per hour Study 1950–63 4.52 Breed and De Foe 198483 1964–70 2.43 1971–77 0.70 1981–82 0.35 1976–77 0.71 Fernandez-Collado et al. 197886 1976–77 2.19 Greenberg et al. 198487,a 1977–78 2.66 1984 1.01 Cruz and Wallack 198688 1993 1.20 Hazan and Glantz 199589 1998–99 Not reported Christenson et al. 200090 aUnlike the other studies, this one did not restrict its sample to prime time television.

Christenson and colleagues90 analyzed among major television characters was content of 168 episodes of top-rated quite low: 11% of males and 2% of females television dramas and situation comedies smoked. Similarly, Cruz and Wallack88 found broadcast from 1998 to 1999. Tobacco was that smoking was more prevalent among used in 19% of episodes. Comparing these male than female television characters. results with those obtained in their content Fernandez-Collado and colleagues86 found analysis of movies,27 they concluded that that in a sample of prime time dramatic young viewers were considerably less likely to television from 1976 to 1977, fewer smoking view smoking on television than in movies. incidents occurred per hour during television programming with the largest Gerbner and colleagues84 found that, in child audiences. Similarly, Christenson and a 10-year sample of prime time dramatic others90 found that in television programs television and a 3-year sample of television from 1998 to 1999, tobacco was used less commercials, the prevalence of smoking frequently in TVG-rated episodes (6%)

Smoking Shifts to the Bad Guys

Social trends can influence not only the quantity of tobacco portrayal on television but also the context in which it is portrayed. For example, Breed and De Foe observed a shift over time in the manner of portraying smoking on television. Between 1950 and 1963, “all kinds of adults— heroes and heroines as well as villains—were seen smoking.”a(p.263) Between 1971 and 1982, however, the typical smokers on television were villains or insecure characters; by the 1980s, scenes parodying cigarette smoking began to emerge. Cruz and Wallack, however, found that in prime time television in 1984, the majority of male smokers (70%) were in strong and enduring roles, with a minority viewed as antagonists.b aBreed, W., and J. R. De Foe. 1984. Drinking and smoking on television, 1950–1982. Journal of Public Health Policy 5 (2): 257–70. bCruz, J., and L. Wallack. 1986. Trends in tobacco use on television. American Journal of Public Health 76 (6): 698–99.

393 1 0 . R o l e o f E n t e r t a i n m e n t M e d i a

compared with TVPG-rated (20%) and TV14­ observation period than did those who rated (24%) episodes. In their total sample watched less than two hours per day. of television episodes, 8% of adult major A cross-sectional survey of adolescent characters used tobacco and no characters smokers in Belgium found a positive, younger than 18 years of age were portrayed curvilinear association between television smoking. Byrd-Bredbenner and colleagues91 viewing volume and smoking volume; the found that during 1998 prime time children’s relationship was stronger for higher levels television programming, depiction of tobacco of viewing.93 This association occurred was rare (shown in 2% of scenes), typically in a multivariate regression analysis that portrayed as a background activity performed controlled for other predictors of adolescent by adults, mostly men. smoking. Adolescent smokers who watched five or more hours of television per day Tobacco portrayal in prime time television smoked 60–147 more cigarettes per week is less common than in movies. Only a than those who watched one hour or less. minority of portrayals (23%) express Another longitudinal study of a New Zealand negative statements about smoking, almost birth cohort94 found an association between none (less than 1%) mention or portray higher exposure to television during negative consequences of smoking, and none childhood and smoking in young adulthood. of the major characters depicted as smokers This study controlled for childhood made on-screen attempts to quit smoking.90 socioeconomic status and parental smoking. These content analyses relate primarily to television programming in the United States. These studies suggest the possibility that The studies document some smoking television viewing could be linked with content but not to the extent seen in movies. smoking initiation and maintenance. If a social influence effect is assumed, it is not Three studies have examined the association clear how much of the effect is mediated by between television viewing and smoking. smoking seen in television programming One examined the association between versus smoking depicted in televised movies, viewing and smoking initiation for a sample because movies comprise a substantial share of U.S. adolescents.92 The authors examined of television programming. Additionally, smoking initiation among 592 adolescent in the longitudinal study by Dalton and never smokers enrolled in the National colleagues3 on the relationship between Longitudinal Study of Youth and for whom exposure to smoking in movies and data on television viewing were available adolescent smoking initiation, self-report at baseline (1990, when subjects were measures of exposure to daily television 10–15 years of age). Initiation of smoking were not associated with smoking initiation during the following two years was examined after controlling for other social influences as a function of baseline television viewing, (exposure to smoking in movies, friend controlling for several socioeconomic and smoking, family smoking). Therefore, demographic factors (ethnicity, household the argument for a social-influences link poverty, marital status, number of children in between exposure to smoking in television the household), maternal factors (education, programming and adolescent smoking is measured intelligence, employment), and not as well established as is the link for child factors (gender and baseline child exposure to smoking in movies. aptitude test scores). Children who watched more than five hours of television per day Popular Music (above mean exposure) had significantly higher adjusted odds of smoking initiation Roberts and colleagues27 analyzed the (adjusted OR of 5.99) during the follow-up content of lyrics for the 1,000 most

394 M o n o g r a p h 1 9 . T h e R o l e o f t h e M e d i a

popular songs from 1996 and 1997. They can present both positive and negative found tobacco references were relatively images and messages about smoking, these uncommon in song lyrics (3% of songs). studies have focused on two key questions. Tobacco references occurred more frequently First, what coverage do magazines give to in rap song lyrics than in other musical smoking and health, and is this coverage genres (7% of rap songs compared to 4% related to whether they accept tobacco of alternative rock songs and 2% or less of advertisements? Second, what is the nature other music genres). Similarly, a content and extent of positive images of smoking in analysis by DuRant and others95 of a sample editorial material, such as fashion pictures? of music videos (N = 518) televised during Both questions are addressed below, and 1994 found that rap music videos (30%) further discussion of the first question were most likely to depict smoking, followed appears in chapter 9 in the section “Tobacco in order by adult contemporary (23%), rock Industry Influence on News Reporting.” (22%), country (12%), and rhythm and blues (11%). A small number of videos (N = 11) Between 1967 and 1979, coverage of the contained 10 or more instances of smoking health hazards of tobacco smoking in major behavior. The results in these two studies women’s magazines in the United States suggest that visual references to tobacco was generally uncommon. Whalen and in popular music videos are more common colleagues96 found that editors of such than verbal references to tobacco in popular magazines frequently encouraged health song lyrics. However, because these studies writers to avoid the subject of tobacco. Those used different sampling methods, the results magazines that did run frequent articles on are not directly comparable. smoking and health did not accept tobacco advertising. Warner and others97 found, in DuRant and colleagues95 found that portrayal a sample of 99 U.S. magazines published of tobacco use was more common in music between 1959 and 1996, strong statistical videos televised on MTV (26%) than on other evidence that cigarette advertising in networks (Video Hits 1 [VH1], 23%; Black magazines was associated with diminished Entertainment Television, 17%; and Country coverage of the hazards of smoking— Music Television, 12%). Few videos contained especially in magazines directed toward branded tobacco advertising, and most of women. These studies’ findings suggest that those were on MTV (N = 4) and VH1 (N = 3). financial dependence on tobacco industry In music videos that portrayed smoking, the advertising may have influenced editorial lead singer was twice as likely to smoke as policy. In the United States, between 1996 a background singer or musician. Smokers and 1999, popular general interest and in music videos were mostly young adults health magazines covered tobacco less than (76%) and were more commonly Caucasian other health topics, and this discrepancy and male. Smoking scenes tended to have was more marked in the latter group.98 a positive emotional tone, but they were no The authors argue that the relatively low more likely to contain sexual content than coverage of tobacco and its hazards presents were videos that did not depict smoking. readers with a skewed account of the importance of smoking as a threat to their Magazines health relative to other health issues.

Numerous studies have examined the A survey of the tobacco policies of the amount and nature of tobacco-related most widely read European women’s content in high-circulation magazines, magazines published in 1996 found that particularly magazines for women and most of the magazines accepted cigarette young people. Recognizing that magazines advertisements, but a minority reported

395 1 0 . R o l e o f E n t e r t a i n m e n t M e d i a

having published a major article on smoking published in 1999 found major differences and health.99 Magazines that accepted between young women’s and young men’s tobacco advertising were slightly less magazines.102 Young men’s magazines likely to have covered smoking and health carried considerably more tobacco compared with magazines that did not advertising and positive images and accept tobacco advertising. Other apparent coverage of smoking in editorial pages than obstacles to coverage of the health effects did young women’s magazines. In addition, of smoking mentioned by editors were their very few young men’s magazines carried opinions about smoking, their perceptions any smoking-or-health coverage. Editorial of their readers, a perception that the images of smoking were most frequent in smoking story had been “done,” or, in some features about personalities, such as an countries, a general ignorance of the subject. interview accompanied by a picture of the In contrast, nearly half of the magazines celebrity smoking. Second most common allowed editorial images of smoking, such were smoking images in fashion pictures as models smoking on fashion pages and that included both posed as well as pseudo celebrities smoking in feature articles. “real-life” fashion shots. Similar, though less prevalent, were smoking images in “slice of In a study of popular Australian magazines, life” items about “real” people out having Chapman and colleagues100 found, after fun, for example, at nightclubs and music the introduction of a ban on tobacco events. The amount of prosmoking coverage advertising in print media in 1991, in the three most widely read young men’s an initial increase in incidental depictions magazines in 1999 averaged more than eight of smoking (6 months after the ban), pages per issue, an increase of more than followed by a reduction in such depictions 400% since 1991.102,103 in the subsequent 18 months. The authors found that photographs of smoking were Content analyses found that print media infrequent in Australian magazines, with coverage of cigars also increased during a mean of one incidental depiction of the 1990s. In a sample of high-circulation smoking per 147 pages. These findings U.S. newspapers and magazines, articles indicate that, in Australian magazines focused on cigars increased substantially produced in the context of bans on paid between 1987 and 1997.104 The articles tobacco advertising, incidental magazine tended to portray cigars and the tobacco content presents nonsmoking as normative. industry favorably but rarely mentioned In contrast, a study of cigarette advertising the health risks of cigar smoking. Between and health aspects of smoking in British 1992 and 1998, a significant upward trend magazines, before and after the introduction occurred in cigar images and images of of a voluntary restriction on cigarette women smoking cigars in U.S. women’s advertising in 1986, found that while the magazines with the highest readership of proportion of magazines accepting cigarette adolescent girls.105 Wenger104 found that advertising decreased, the new restrictions cigar “lifestyle” magazines recurrently did not cover the most popular magazines; presented content that associated cigars thus, protobacco content remained prevalent with business stories, social events in the highest circulation magazines.101 (including fundraisers for charities), and Furthermore, editorial coverage of the celebrities. Of the celebrities and public health aspects of smoking was low and did figures quoted or described in the articles, not increase following the voluntary ban. most (87%) were portrayed as having favorable attitudes toward cigars. Only 1% A content analysis of the most popular of cigar-focused articles focused primarily British young people’s style magazines on the health effects of cigars. Cigar use was

396 M o n o g r a p h 1 9 . T h e R o l e o f t h e M e d i a

presented as normative behavior and as a differently.107 Young people rated perfectly key element of a successful lifestyle. matched (other than the presence or absence of a cigarette) smoking and nonsmoking The second question addressed in research fashion pictures taken from youth and style on tobacco-related content in magazines magazines on a range of attributes. The study is the nature and extent of positive images found that the presence of a cigarette affected of smoking conveyed in fashion pictures. how the pictures were rated and that the Magazines have a potentially important nature of this effect differed between pictures. influence on the social image of smoking, In general, the smoking images were rated as they often have high readerships; are as being more “druggy,” wild, and depressed. targeted toward and therefore tailored to Identical nonsmoking images were rated as appeal to different audiences on the basis of being more healthy, rich, nice, fashionable, age, gender, ethnicity, and socioeconomic slim, and attractive. On the surface, the status; and are printed so they remain smokers’ attributes were negative, but some available for longer periods than other of the attributes represented images that media (as reflected in the often high ratio young smokers aspired to and admired. of readership to circulation). Of particular Smokers, especially males, identified more concern are magazines aimed at young strongly with the smoking images and people. As discussed earlier, adolescence attributes than did nonsmokers.107 is a period of considerable change and transition during which young people The second study, by MacFadyen and engage in the active construction of their colleagues, used focus groups of first-year adult identities, not only about who a young college students, all smokers, to explore person wants to become, but also how an perceptions of smoking images in youth style image can be projected in particular social magazines and the relationship between contexts.106 Young people’s magazines, these perceptions and their own smoking by promoting certain styles, brands, and images and identities.54 The research images, not only help create the latest found the students perceived this imagery fashions but define what and who is “in.” to be, on the whole, attractive, sociable, To appeal to young readers, these magazines and reassuring. There was considerable attempt to embody attitudes and values by synergy among the image of smoking, incorporating them into fashion spreads the personality of the magazines, and and articles that tap into and articulate respondents’ self-images. The most popular what it means to be a young person today. magazines had personalities that were Thus, it is theorized, both the extent to similar to the students’ image of smoking— which magazines show smoking images and carefree hedonism, risky behavior, and the types of such images may be important antipolitical correctness. This finding in influencing young people’s perceptions suggests that the display of smoking in these of the desirability of adopting a smoking magazines was likely to reinforce positive identity and consequently affecting their perceptions of smoking and contribute to behavior. So far, however, very few studies the belief that smoking is a normative and have explored how young people engage important part of student culture. with magazine images of smoking or the effect of such images. The findings by MacFadyen and colleagues are similar to those from an Australian study Two British studies used different methods to that used focus groups to explore secondary explore this question. A study by Amos and school (both smoker and nonsmoker) colleagues examined whether young people students’ perceptions of smoking images perceived smoking and nonsmoking images in magazines and films.47 Smoking in

397 1 0 . R o l e o f E n t e r t a i n m e n t M e d i a

magazines and films was perceived as despite sexually explicit content and/or the normal and acceptable. Additionally, the capacity to order tobacco-related products young people felt that most of the images online on a number of these sites, most do used in the study portrayed smoking not require age verification procedures. positively in terms of mood attributes, Ribisl and others also found that one-third such as being in control or confident. of such Web sites featured smoking stories Such positive images of smoking portray that “instructed would-be smokers on the smoking in a way that young people merits of smoking and provided reasons interpret as being a normal part of life. for resuming smoking for those who have already quit.”109(p.74) Internet Further information on the use of the Hong and Cody108 conducted a content Internet in tobacco marketing appears in analysis of protobacco Web sites (N = 318). chapter 4. These sites were predominantly e-commerce sites (50%), followed by hobby/recreation Other Entertainment Media sites (19%), erotic/fetish sites (15%), other tobacco-related sites (8.8%), corporate Smoking content in newer forms of sites (5.7%), and smoker’s rights/lobbyist entertainment media, such as increasingly sites (2.5%). Ribisl and colleagues109 also realistic video games (e.g., cigar smoking conducted a content analysis of protobacco in the video game Halo 2), has been largely Web sites (N = 30 sites). However, their ignored despite the widespread use of these sample excluded sites for individuals or games (see chapter 4). T-rated (teen-rated) organizations that manufacture or sell video games comprised 28% of video and tobacco products. Despite the different computer sales in 2002.110 In a content sampling criteria used in these studies, analysis of T-rated video games, Haninger they yielded similar findings. On e-commerce and Thompson111 found that 5 (6%) of sites and sites featuring hobbies, recreation, 81 games showed tobacco use (either and “fetishes,” imagery depicting smoking in a character used tobacco or a tobacco association with glamour, relaxation, leisure, product otherwise appeared in the game). sex, or alternative lifestyles was prevalent; It is unclear what social normative effects negative health effects of smoking were (e.g., smoking norms) are associated with rarely depicted or mentioned.108,109 playing these games. However, in domains other than smoking, the games have The models portrayed on such sites were influenced behavior in children and young predominantly young (18–34 years old) adults. For example, playing violent video and Caucasian in appearance. Females games has been shown to increase aggression tended to be portrayed as attractive and in children and young adults.75 More research slim while males appeared more average is needed on these influences. Assessing in appearance.108 Hong and Cody argue that, whether tobacco is portrayed in a negative in addition to portraying predominantly or positive light also is important. Haninger young role models, many protobacco Web and Thompson111 state that a character in sites contained features characteristic the video game Shadow of Destiny decides of the Web sites young people frequent. to quit smoking cigarettes because, he says, For example, they contain content related to “I don’t want to die,” reinforcing negative “shopping, hobbies and recreation (including health consequences of cigarette smoking. entertainment), sites featuring celebrities and sites featuring sex or sexually arousing The effects of smoking by people performing visuals.”108(p.291) Both studies found that, in live concert and theater venues also

398 M o n o g r a p h 1 9 . T h e R o l e o f t h e M e d i a

might be studied. Some research on a tobacco use in its films. However, paid live theater production to encourage product placement deals between some nonsmoking has been reported.112 However, movie production companies and tobacco the effects of characters smoking on stage companies, and contracts precluding during live theater performances have unattractive movie depictions of smoking,18 not been examined. Some of the other reveal that some in the entertainment entertainment venues in which smoking industry have been compensated by the influences have been understudied include tobacco industry to add branded smoking smoking by musicians in live concerts, and other signage to their artistic output. depictions of smokers in comic books,113,114 Given the history of product placement and (noted earlier) smoking images in in movies and the similarities between movie promotional material. the social imagery of smoking in movies and in tobacco advertising, it is likely that the social iconography of smoking Efforts to Reduce in films derives in large part from images Exposure of smoking that the tobacco industry cultivated strategically.

Legal/Policy Issues: Artistic or In the past, the American movie industry Commercial Speech? was not afforded the First Amendment protections it now enjoys in the United One of the foundations of democratic States115 and was subject to censorship society involves freedom to express at both state and local levels. The movie a diversity of views (see chapter 8). industry fought censorship, arguing that it Expression of diverse viewpoints is interfered with First Amendment speech. valuable for enabling communicators to But in 1915, in Mutual Film Corporation espouse a cause or position and defend v. Industrial Commission of Ohio, the it. The expression of diverse viewpoints U.S. Supreme Court determined that provides audiences with material on which motion pictures did not constitute part of to base informed judgments about the the “press” and therefore were not entitled world around them. This freedom applies to First Amendment protection from not only to political commentary but also censorship. This case arose in response to to commentary on behaviors within the the passing of a statute creating a Board culture. Thus, most free societies give of Censors that had to approve all motion artists and other communicators the pictures prior to their exhibition. Localities ability to reflect on, depict, and comment continued to censor movies until 1952, on their perception of the world around when the Supreme Court granted full them. In the United States, this freedom is First Amendment protection to movies in incorporated into the constitution as the Joseph Burstyn, Inc. v. Wilson. At that time, First Amendment of the Bill of Rights. there was little or no product placement in movies, but this is no longer the case. Interviews conducted by Shields Paid product placement is an integral and colleagues53 with film industry commercial element in almost every movie. representatives illustrate the value Given the increasing number of product producers and actors place on freedom of placements in movies, the question is now speech and their fears about censorship. whether or not depictions of brands in The movie industry does not welcome movies should be reclassified as commercial public health strategies that advocate speech, which would be subject to a lower for restricting the freedom to depict level of First Amendment protection.

399 1 0 . R o l e o f E n t e r t a i n m e n t M e d i a

Self-regulation by eliminating cigarette and television programs unless smoking is brands already is happening in some movie an essential part of the action.118 The Indian production companies. For example, Government had planned to impose a ban on Robert Reiner requires justification for smoking scenes in new films and television smoking scenes in movies he produces for serials in July 2006.119 Thailand’s Film Castle Rock Entertainment.116 As a WHO Censorship Board has censored depictions document on this issue states, “The film of smoking in movies. For example, the industry cannot be accused of causing release of the movie Som + Bank (Bangkok cancer, but they do not have to promote a for Sale) was delayed, as the board required product that does.”117 In contrast to violence, that the images of smoking be blurred out.120 which may be linked with box-office success, In other countries, efforts are under way the evidence indicates that the inclusion of to incorporate smoking into government smoking is not necessary for the commercial censorship and movie rating systems. For success of movies.32,42 example, the Lung Association in Ontario, Canada, has called upon the government Product placement deals are not the sole to censor smoking.121 Some countries also reason for on-screen smoking. The decision censor aspects of films considered offensive to portray a character as a smoker may to most adults in their societies. For example, arise from a range of motives, such as a many Arab countries do not allow movies desire to make the character seem realistic, that depict use of tobacco and alcohol to reliance on cigarettes as a prop, and personal be shown in public places, because doing smoking behavior of an actor.53 Nevertheless, so violates mainstream religious beliefs movie characters for the most part represent (personal communication from R. Kelishadi, the affluent and most powerful segment of M.D., Isfahan University of Medical Sciences, society.3,28 When these actors smoke, whether Isfahan, Iran, to J. Sargent, 2004). they play the bad or good guy, the risk is that adolescents will emulate the behavior.3,4 Because of unique protections on First Amendment speech in the United States, this Movie Rating Systems country does not have censor boards. Instead, the United States is the only country that In most countries, movie rating systems allows its film industry to rate its own motion exist to protect children from exposure to pictures. Rating is done through the MPAA. forms of media society deems harmful or This rating system, established in November objectionable. The rationale for most rating 1968, has undergone only minor changes. systems is that society wishes to protect In the voluntary MPAA rating system, most children from seeing media that may have producers allow their films to be subjected undue influence on their behavior. Most to review by a rating board. Movies are countries have government-sponsored rated primarily according to what the board censor boards charged with evaluating the determines parents would find objectionable appropriateness of entertainment media for (or what Congress might regulate). In its children. The procedures of government- explanation of the ratings system, the MPAA sponsored censor boards are subject to lists violence, nudity, sensuality, language, regulation by government and to revision if and drug use as factors the board considers new data arise regarding a media threat to when rating movies. Board members must children. Governments in some countries have parental experience, and the board have attempted to regulate smoking content president is chosen by the MPAA’s president. in entertainment media. In 2001, Russia’s The MPAA and the National Association lower house of parliament passed a bill to of Theatre Owners presidents jointly set ban images of people smoking in movies decisions regarding rating criteria.122

400 M o n o g r a p h 1 9 . T h e R o l e o f t h e M e d i a

The MPAA promotes the ratings system as from directly or indirectly targeting youth a guide to parents. Some might argue that in marketing. No studies have yet been the real purpose of the voluntary movie published on cigarette brand placements ratings system is to protect the studios in movies since the signing of the MSA. from more intrusive government regulation. However, a number of movies released after In that regard, the film industry has operated this agreement have included cigarette brand in much the same way as the tobacco and placements. Because the U.S. attorneys alcoholic beverage industries, with the former general are charged with enforcing the MSA, changing its voluntary rating standard, the continued appearance of cigarette brands the Cigarette Advertising and Promotion in movies has become a topic of interest. Code, only when Congress was considering So far, the tobacco industry has denied stricter regulations (see the section “Failure violating the MSA by obtaining cigarette of Self-Regulation” in chapter 3). brand placements; the denials are in response to several inquiries by the state attorneys Voluntary Efforts general (for more information, see the statement by J. Joseph Curran, Jr., Attorney Tobacco Industry General of Maryland in appendix 10A).

Voluntary Advertising Standards Movie Industry U.S. tobacco companies’ voluntary Cigarette Advertising and Promotion Code was Before describing efforts by some in the modified in 1990 to prohibit paid product movie industry to limit the depiction of placement. The tobacco industry initiated the smoking, it is necessary to describe the voluntary ban on paid product placements industry. Although the industry changes in the same year that the U.S. Federal Trade from year to year with buyouts and mergers, Commission conducted an inquiry into the U.S. film industry in 2004 was organized product placement activities of various around seven major production companies tobacco firms. Little change occurred in the that finance and distribute motion prevalence of cigarette brand appearances pictures: Buena Vista Pictures (Disney), after the initiation of the voluntary ban.33 Sony Pictures, Metro-Goldwyn-Mayer, Moreover, the frequency of on-screen Paramount Pictures, Twentieth Century Fox, smoking increased in the 1990s, compared Universal City Studios, and Warner Brothers with the 1970s and 1980s, suggesting that Entertainment. Many of the names seen in the ban had little impact on either on-screen movies are subsidiaries of these companies. product placement or smoking practices.28–30 For example, Miramax is a subsidiary of Buena Vista Pictures. These large studios Master Settlement Agreement hire production executives responsible In 1998, the U.S. Master Settlement for financing their major in-house movie Agreement (MSA) prohibited participating efforts. Many independent film producers cigarette manufacturers (e.g., Brown & also make movies. For independent movies Williamson, Lorillard, Philip Morris, to be successful, the producer must R.J. Reynolds) from product placement partner with one of the major studios for activities. The settlement bans payments to the widespread distribution of the film. promote tobacco products “in any motion Other players in the industry (the artists) are picture, television show, theatrical production organized through guilds, bodies that serve or other live performance, live or recorded as financial advocates for their constituents performance of music, commercial film or (directors, actors, screenwriters, etc.) in video, or video game.”123(p.18) The MSA also much the same way that labor unions act on prohibits participating tobacco companies behalf of their members.

401 1 0 . R o l e o f E n t e r t a i n m e n t M e d i a

The MPAA represents the domestic interests health behaviors in entertainment media, of the major studios, and the Motion Picture particularly movies. Their strategies can Association represents the international be broadly categorized as collaborative or interests. The president of the MPAA is confrontational. also the chief lobbyist for the industry in Washington, D.C. When approached by the Collaborative Approaches state attorneys general in August 2003, Jack Valenti, the MPAA president at the The Council for Excellence in Government time, sponsored a series of meetings that and the University of Southern California, included himself, the NATO president, and Annenberg School for Communications, various guilds. However, Valenti declined to Norman Lear Center, published a review incorporate smoking into the MPAA rating of all efforts to engage the entertainment system. (For more information on the industry in developing prosocial messages dialogue between the state attorneys general into entertainment.127 The report, How and the motion picture industry, see the Pro-Social Messages Make Their Way into statement by Maryland Attorney General Entertainment Programming, summarizes Curran in appendix 10A.) Four years later, in these programs and provides a guide to February 2007, the Harvard School of Public some of the following discussion. Health recommended that the MPAA take action to “eliminate the depiction of tobacco Office on Smoking and Health smoking from films accessible to children and The Office on Smoking and Health (OSH) is youth.”124 In May 2007, 31 attorneys general a division of the National Center for Chronic wrote a letter to major movie studio heads Disease Prevention and Health Promotion supporting this recommendation and stating of the Centers for Disease Control and the dangers of exposing children to smoking Prevention (CDC). The OSH maintains a Web depictions in movies. In a response released page that encourages members of the public that same month, former congressman to work with the entertainment industry Dan Glickman, Valenti’s successor as to promote accurate depiction of tobacco president of MPAA, stated that the MPAA use and health information in movies, would begin to consider smoking depictions television, and other media.128 By “accurate,” when rating movies. However, a letter to the the group means that movies should show MPAA in June of 2007 from U.S. Senators the health consequences of smoking. Since Durbin, Kennedy, and Lautenberg described 1997, the OSH has developed a collaborative MPAA’s new policy as “not enough to curb the relationship with the entertainment industry influence of smoking in the movies on the to achieve three strategic aims: (1) educate health of children.”125 Six months after the and provide accurate science and resources new policy began, Polansky, Glantz, and Titus to the creative community for television reported that there was no substantial change programming and films containing tobacco- in the percentage of G, PG, or R-rated movies related themes; (2) develop public relations that included smoking depictions compared campaigns and provide media training for with the same time period in each of the four volunteer celebrity advocates who want to previous years.126 use their public profile to advance tobacco- free lifestyles; and (3) develop educational Efforts to Induce/Promote Change materials, with the cooperation of the entertainment industry, that can be used A number of interested government and in schools and by health partners to teach citizen groups have attempted to exert and reinforce messages about the dangers of influence on media policy and production tobacco use. The approach is to encourage in relation to tobacco use and other the entertainment industry to deglamorize

402 M o n o g r a p h 1 9 . T h e R o l e o f t h e M e d i a

and denormalize tobacco and its use. It is not “First Draft,” a technical resource service clear how successful the group has been in that provides information on request; persuading individuals in the entertainment “Spotlight on Depiction,” resources for industry to reduce or eliminate smoking. writers; and “Generation Next,” educational resources for film students. In addition, the Seeking Tobacco Alternatives with Realistic EIC annually presents the PRISM awards, a Solutions Project nationally televised awards show recognizing The American Lung Association of the accurate depiction of drug, alcohol, and Sacramento-Emigrant Trails initiated the tobacco use and addiction in film, television, Seeking Tobacco Alternatives with Realistic interactive, music, video, and comic book Solutions (STARS) project in 1998. The aims entertainment.130 Established in 1997, the of the project were to work PRISM awards honor productions that are powerfully entertaining and realistically with the entertainment industry to show substance abuse and addiction. The reduce the unintentional glamorization Robert Wood Johnson Foundation, the of smoking in film and television, provide OSH, the National Institute on Drug Abuse, media education to the general community and the CDC are among the groups that regarding pro-tobacco messages, and jointly sponsor these awards. The intent of conduct research regarding the impact of the PRISM awards is to encourage artists the tobacco industry on the entertainment to “make the most of their rights to free community and acts to reduce this creative expression, while at the same time impact.129(pp.10–11) showing the reality of substance abuse and addiction on screen, in song and on the With support from the California Tobacco page.” The awards serve to communicate and Control Program, STARS produced an award- reward realistic depictions of substance use. winning documentary, Cigarettes, Cinema, However, it is not clear to what extent the and the Myth of Cool.129 This film features awards foster change or even to what extent writers, directors, and actors speaking directors and screenwriters are aware of about social responsibility and smoking in them or use the resources the EIC provides. movies. During the course of the project, a Blue Ribbon Advisory Committee regularly Attorneys General/Master Settlement convened; the committee included Hollywood Agreement directors and producers. It is not clear that The state attorneys general have an interest progress was made during the project in in reducing youth smoking as part of their eliminating smoking from movies, and the involvement in the MSA (see appendix 10A, project ceased in 2003 because of lack of a statement from Maryland Attorney General funds. However, STARS did result in a well- J. Joseph Curran Jr., for details on this regarded documentary that showed both initiative). To this end, they have begun sides of the debate over smoking in film. to collaborate with the movie industry with the aim of decreasing the prevalence Entertainment Industries Council of depictions of smoking in movies. The Entertainment Industries Council The underlying concern raised by the (EIC) is a nonprofit organization that aims attorneys general is the role movies play in to provide information, awareness, and smoking by youth. In August 2003, 28 state understanding of major health and social attorneys general, led by Mr. Curran, issues among the entertainment industries approached Mr. Valenti, the MPAA president, and to audiences at large. The EIC was asking the organization to reduce smoking founded in 1983 by entertainment industry in movies. A letter from Mr. Valenti then leaders. The EIC has three areas of focus: invited the attorneys general to a series

403 1 0 . R o l e o f E n t e r t a i n m e n t M e d i a

of discussions on the issue (see letters 2001.132 The campaign aims to reduce the in appendix 10B). This letter may have impact of smoking in movies on adolescents been the first public statement made by a through four specific, voluntary changes in movie industry spokesperson on smoking movie industry policy: in movies, despite many press inquiries as a result of scientific publications that Rate new smoking movies R. Any film linked smoking in movies with teens’ that shows or implies tobacco use should smoking. The initial dialogue resulted be rated R. The only exceptions should be in a series of meetings among scientists, when the presentation of tobacco clearly several attorneys general, and movie and unambiguously reflects the dangers industry leaders. In May 2007, 31 attorneys and consequences of tobacco use or is general once again approached the MPAA, necessary to represent smoking by a real NATO, and major studio heads to decrease historical figure. depictions of smoking in movies directed at youth.131 It also led to a hearing convened Certify no payoffs. The producers should in April 2004 by the Senate Committee on post a certificate in the credits at the end Commerce, Science, and Transportation, of the movie declaring that nobody on to consider the impact of smoking in the production received anything of value movies on children. The Senate hearing is (cash money, free cigarettes or other evidence of an expanding demonstration gifts, free publicity, interest-free loans, of substantial interest in major political or anything else) from anyone in exchange institutions in the United States regarding for using or displaying tobacco or its use. tobacco use in movies and its potential impact on children. In addition to meeting Require strong antismoking advertisements. with industry representatives, the attorneys Studios and theaters should require a general have addressed the tobacco industry genuinely strong antismoking advertisement with respect to movie brand appearances. (not one produced by a tobacco company) Assistant Attorney General Dennis Eckhart to run before any film with any tobacco of California sent letters to the legal presence, regardless of its MPAA rating. counsels of tobacco companies whose brands appeared in movies after the MSA. In each Stop identifying tobacco brands. There case, the letters prompted communication should be no tobacco brand identification between counsel for the tobacco industry and no presence of tobacco brand imagery and counsel for the movie industry to verify (such as billboards) in the background of that there was no violation of the MSA in any movie scene. the form of a payment to place the brand (see example in appendix 10C). This legal The aim of the Smoke Free Movies activity was a sign to tobacco companies that campaign is to create a groundswell of they are being monitored. It is also possible support for these policy aims within the that, as a result, the motion picture industry public health community and, eventually, will act upon requests by tobacco companies among public policymakers to bring not to have their brands used in movies. pressure to bear on the industry. By 2004, the campaign gained the endorsement of Confrontational Approaches many mainstream health organizations, including WHO, the American Medical Smoke Free Movies and the Rate Smoking Association, the American Academy “R” Public Health Campaign of Pediatrics, and the American Heart Smoke Free Movies is a public health Association. The Smoke Free Movies campaign started by Stanton A. Glantz in media campaign began by rolling out a

404 M o n o g r a p h 1 9 . T h e R o l e o f t h e M e d i a

Example of a Smoke Free Movies advertisement

405 1 0 . R o l e o f E n t e r t a i n m e n t M e d i a

controversial print advertising campaign Another issue that has been raised is in March 2001 that was aimed at members whether the balance between adolescents’ of the movie industry. The campaign was desire to see R-rated movies and parental designed to raise awareness about the attempts to limit viewing of these movies effect of smoking in movies on adolescent weighs in favor of higher or lower exposure smoking; to place responsibility for change rates for R-rated movies among young on studio executives, theater owners, adolescents. If adolescents successfully and actors; and to suggest government circumvent attempts by parents and oversight.133 Along with the advertising theaters to restrict their exposure to campaign, Smoke Free Movies has organized these movies, their viewing rates would and maintains a network of public health be expected to be similar to other rating activists at state and local levels. These categories. The R rating for the smoking groups have developed awareness campaigns campaign, in this case, would be futile aimed at youth (in New York, Texas, and and possibly even counterproductive. Vermont, among others), have engaged in a If view rates for R-rated movies are in fact national letter-writing campaign to movie lower among young adolescents, then the stars, and have encouraged other forms argument could be made that rating movies of activism, such as e-mail messages to with smoking R could limit adolescent movie executives. exposure despite making them “forbidden fruit.” To shed light on these possibilities, The most controversial policy aim of researchers7,134 examined the reach of Smoke Free Movies is the R rating for movies, as determined by MPAA ratings, smoking. This policy aim has been for a sample of young adolescents. under the control of the movie studios and theater owners, the two entities The adolescents were part of an already that run the MPAA rating system. From published cross-sectional survey of the original perspective of the movie 4,946 students, 10–14 years of age, attending industry, the movie rating system was 15 junior high schools in New Hampshire designed for concerned parents and was and Vermont.7,134 Each student was surveyed not designed in relation to public health on whether he or she had seen a randomly considerations. However, the ratings do selected subsample of 50 movies, drawn include violence. After the Columbine from 601 popular contemporary movies High School shootings in 1999, public (based on year of release and box-office health considerations were added when success). Almost 50% of the movies were efforts by President Clinton, the Senate, rated R. Because movies were randomly and public health experts led to changes selected, each title appeared on an average in the movie industry’s depiction of of 470 surveys (standard deviation of seven). violence in R-rated films. The movie Therefore, it was possible to determine industry deleted the most violent scenes accurately the percentage of adolescents from soon-to-be released films and who had seen each title (termed reach in increased restrictions on how R-rated the marketing literature). G-rated movies movies are marketed. From a public health were seen by most of the adolescents, with perspective, limiting the portrayal of a median reach of 67% of adolescents. tobacco in movies is important because of As the rating becomes more restrictive its link to adolescent smoking (see earlier toward adolescents, reach drops. This is discussion) and the severity of the health especially true for the transition from consequences of smoking compared PG-13 rating to R rating, for which the with some other depictions of behavior median and interquartile ranges for (e.g., using foul language). reach drop substantially. Whereas the

406 M o n o g r a p h 1 9 . T h e R o l e o f t h e M e d i a

75th percentile for reach in PG-13 movies depictions of smoking, and some evidence was more than 60%, the 75th percentile supports this idea. Most research involves for R-rated movies barely exceeded 30%. restriction of access to movies in the These data provide convincing evidence R-rated category. that movies in the R-rating category are seen by many fewer young adolescents R-Rated Movie Restriction compared with movies that are not rated R. The prevalence of smoking depicted in This result is probably because parents movies increases with high levels of movie restrict access (see below) and because rating. In a sample of 250 contemporary theaters generally enforce the R-rating movies, Dalton and colleagues32 showed that as part of their participation in the MPAA the median number of smoking depictions ratings system. was 8.5 for R-rated movies, 4 for PG-13–rated movies, 3.5 for PG-rated movies, and 1 for Would the R rating for smoking have G-rated movies. About one-half of the movies a substantial immediate impact on produced in 1990 were R rated, and that adolescents’ exposure to smoking in percentage dropped to one-third after 2000. movies? Smoke Free Movies is calling Thus, by restricting access to R-rated movies, for the R rating to be applied only to new parents reduce movie exposure overall movies. Most adolescents’ exposure to by a factor of one-third to one-half and R movies is through seeing older movies eliminate movies that contain the highest on video and DVD. The prospective R rating concentration of smoking. for smoking would therefore substantially cut exposure to depictions of smoking at Two studies examining the effect of parental theaters that air new releases and would R-rated movie restriction on adolescent have a more pronounced impact over time smoking were identified. The studies of a because of the cumulative effects of the sample of Vermont and New Hampshire rating change. On the other hand, if the children aged 10–14 years at baseline R rating for smoking caused parents to assessed parental restriction of R-rated pay less attention to the ratings system, movies through the question, “How often it could result in the reach of R-rated movies do your parents allow you to watch movies increasing among younger adolescents. or videos that are rated ‘R’?” (never, once Because of these concerns, it may be wise to in a while, sometimes, all the time). also consider, along with implementation of In the cross-sectional study,135 90% of the this policy change, surveillance of R-rated 4,544 students were younger than 14 years movie viewership among adolescents of age. However, only 16% reported they and inclusion of a motivational effort to were never allowed to watch R-rated convince parents to take the ratings system movies. One-third (31%) indicated that literally and seriously. their parents never restricted them from viewing R-rated movies. Thus, restriction Other Potential Strategies of R-rated movies was not a major focus for most of the parents of the children in this sample. Among adolescents who reported Parental Supervision of Entertainment R-movie restriction, exposure to R-rated Media movies was about one-eighth as high as that for adolescents who reported no restriction. Most media exposure occurs in the Exposure to PG-13 movies was also reduced household. Therefore, parental supervision by about 50%. Thus, reports of R-rated of their children’s access to media could parent restriction seemed to be associated affect the children’s exposure to media with lower exposure to such movies.

407 1 0 . R o l e o f E n t e r t a i n m e n t M e d i a

Importantly, initiation of alcohol Devices That Restrict Access consumption and tobacco use was much This is a rapidly changing area as technology lower in adolescents reporting movie offers parents more control of the home restriction, even after controlling for a media environment. The shift toward number of other covariates. These variables automated control of home media was included sociodemographics, social spearheaded by the television V-Chip, influences (smoking by friends and a device that enables parents to block family), personality (sensation seeking, television channels and also to block based rebelliousness), and parenting style on television and movie ratings. In the (authoritative parenting). Compared Telecommunications Act of 1996,136 Congress with adolescents with no R-rated movie required manufacturers of televisions to restriction, the adjusted relative risk include a control device that could be used (95% CI) for smoking initiation was by parents to block unwanted programming. 0.74 (0.65–0.85) for adolescents with In the words of the legislation, the device partial restriction and 0.29 (0.19–0.45) for those who were completely restricted enables parents to block programming from viewing R-rated movies. based on identifying programs without ratings, The never smokers in the cross-sectional study were followed up one to two years is available to consumers at a cost which later. Smoking incidence (10% tried is comparable to the cost of technology smoking during the observation period) that allows parents to block programming was examined as a function of parental based on common ratings, and R-movie restriction at baseline.8 Adolescents allowed to see R-rated movies at baseline will allow parents to block a broad range were three times more likely to try smoking of programs on a multi-channel system (relative risks adjusted for a full set of as effectively and as easily as technology covariates) compared with those who were that allows parents to block programming never allowed to watch R-rated movies. based on common ratings … The effect was stronger for adolescents from nonsmoking families, among whom Since 2000, the V-Chip is included on all only 3 of 399 with complete R-rated movie televisions distributed in the United States restriction tried smoking. In this group, with screens larger than 13 inches. In the adjusted relative risk of smoking given addition to the V-Chip, many modern video no R-movie restriction was 10. Students and DVD players contain software that were asked again about movie restriction gives parents the ability to block television at follow-up. Most reported no change in programs by rating, so that their children restriction status, indicating that many cannot play material above a certain parents are able to continue enforcing threshold rating. Given the prevalence of restriction as adolescents age during junior this kind of technology and the interest in high school. Moreover, compared with protecting children from the ill effects of adolescents reporting no change, relaxation media, one would have expected a number of restriction was associated with higher risk of interventions involving the V-Chip. Yet a of smoking in each of the baseline restriction MEDLINE search on “V-Chip” conducted categories. This longitudinal study provides in September 2004 yields only four articles, strong evidence that supports interventions and a search on PsycINFO yields only six— to motivate and assist parents in enforcing none of which involves cross-sectional or media restrictions as a smoking prevention interventional data. Although this technology measure aimed at young adolescents. is in its infancy, the potential benefits of

408 M o n o g r a p h 1 9 . T h e R o l e o f t h e M e d i a

widespread application are clear. One study examining the effect of a blocking device Efforts to Modify that restricted television time showed that Response to Exposure mean daily television time for children in the intervention dropped, as did their increase in body mass index.137 This randomized clinical Antitobacco Advertising in trial provides strong evidence for a powerful Theaters intervention effect. As described in the experimental studies Internet section, there is some evidence that showing an antitobacco advertisement before a movie It may be too early to consider interventions with smoking blunts the movie’s effect on aimed at the Internet as relatively little is attitudes. On the basis of this evidence, one known about how people use it. In a study aim of Smoke Free Movies is to require published in 2004, a sample of underage the distributing production studio to pay adolescents were asked to purchase for antitobacco advertising in theaters. cigarettes over the Internet.138 The authors Another possibility raised in discussions reported that 29 of 30 subjects were able between the representatives of the National to make a purchase by using a parent’s Association of Attorneys General and the credit card, and 75% received the product movie industry is attaching an antismoking in the mail. This study shows that access to message ahead of any videotape or DVD that cigarettes by minors is possible. However, contained smoking. This action would cost as yet the prevalence of such purchasing the industry little or nothing. In 2007, at behavior among the adolescent population least one major studio executive announced is unknown. that the studio planned to add anti-smoking PSAs on DVDs of future films that feature Hong and Cody108 recommend the following cigarette smoking.139 actions to counteract the presence and influence of tobacco on the Web: (1) online As noted earlier, through the impetus of tobacco retailers should be required to use state attorneys general, the possibility of age verification and should not sell tobacco communications about smoking depictions products without a bona fide age check; in movies has been raised with the president (2) consumer awareness information on the of the National Association of Theatre hazards associated with smoking should Owners as well as owner-members. Because be displayed for visitors to protobacco movies appeal strongly to adolescents, movie Web sites; (3) popular portal sites for the theaters may be ideal places for antitobacco general public and adolescents should be advertising campaign messages. However, encouraged to provide links or banner the source of funding for such a campaign advertisements to sites on tobacco cessation is unclear. or to provide educational material on the health effects of smoking; and (4) tobacco Media Literacy control advocates should use the Web more proactively to advocate smoke-free, healthy Media literacy refers to educational environments (e.g., work to have a more approaches to help viewers better understand noticeable Web presence and use some media inputs. Some counteradvertising of the engaging, interactive features that campaigns and contests, discussed in appeal to audiences). By 2004, however, the section “Media Activism” in chapter Congress had not passed any restrictions on 11, can be considered a form of media Internet purchases. literacy. Critical viewing skills are a

409 1 0 . R o l e o f E n t e r t a i n m e n t M e d i a

major component of most media literacy after the intervention) knowledge about educational programs.140 From the health consequences, protobacco attitudes, standpoint of persuasion theory,44,78,141 and use of tobacco. The intervention these programs aim to affect the way the was associated with significantly higher recipient processes media information. knowledge scores, a decrease in protobacco Many of the media images viewers see are attitudes, and a decrease in current tobacco processed implicitly, without much thought. use. Limitations of the study include In theory, adolescents are affected in a measurement of short-term outcomes cumulative fashion by the images of smoking only and inability to attribute attitudinal in the media. As they see literally thousands and behavior change to the media literacy of depictions of smoking—by affluent component of the intervention. characters and without portrayal of negative health effects—in movies, television, and Another study examined the effect of a tobacco marketing materials, over time, they media literacy curriculum on attitudes associate smoking with positive expectations. toward alcohol use in a sample of third- By teaching about the mechanisms by grade students. Austin and colleagues143 which media persuade, media literacy examined the immediate and delayed programs should cause the recipient to effects of a media literacy program on become a more effortful processor of the alcohol in 246 third-grade students. They media—for example, to be more skeptical proposed a model in which more critical of commercial messages and images.141 attitudes toward televised portrayals of An adolescent who is knowledgeable about alcohol use (less perceived realism, less the role of product placement in marketing identification, less desirability) would and the persuasive power of movie images affect alcohol expectancies and, ultimately, of smoking will be more resistant to behavior. Students were randomly assigned automatically accepting the positive to one of four groups according to two expectancies associated with the image. factors: pretest/no pretest and treatment/ no treatment. Outcomes were measured Media literacy has great appeal as theory. immediately and at three months posttest. However, only scant evidence suggests Children in the intervention group watched that these programs have short- or long- a 28-minute videotape Buy Me That, term effects on adolescents. One study was which Consumer Reports produced for identified that evaluated a youth tobacco use children and which discusses techniques prevention intervention that included media used by advertisers to make products look literacy skills among high school students.142 appealing. The videotape was followed by a Using a quasi-experimental design, the guided discussion of four advertisements investigators assigned 448 students in (two for beer and two for soda pop). 15 classes in three schools to receive the Outcomes surveyed included understanding intervention; 161 students in 5 classes of persuasive intent (“Ads on TV tell the in one school served as a control group. truth”), realism (“Real people act like The intervention curriculum included people in ads”), social norms (“Most teens health education (consequences of tobacco drink”), similarity (“I do things that people use, social norms, parental use of tobacco), in ads do”), desirability (“People in beer media literacy skills training (media analysis, ads are popular”), identification (“I want media production, product presentation, to have my life like people in beer ads”), and media advocacy), and skills training in and expectancies (“Drinking makes you resisting peer influence. The investigators happier”). Results of the experiment measured preintervention (one week before generally were very supportive of the notion intervention) and postintervention (one week that media literacy training has a strong

410 M o n o g r a p h 1 9 . T h e R o l e o f t h e M e d i a

immediate effect on raising skepticism The two published longitudinal studies toward advertising and decreasing show an independent link between participants’ intent to engage in the exposure to smoking in movies at baseline behaviors depicted in advertisements. and smoking initiation in the future; Some of these effects persisted, albeit to a estimates of the effect size are consistent lesser degree, at delayed posttest. with their cross-sectional counterparts. The experimental studies examine short- These studies suggest that media literacy term responses, generally supporting an may have a role in training children to effect of seeing movie stars smoking on resist entertainment messages. However, screen on attitudes such as favorable ratings this intervention area is still very little of smokers and intent to smoke in the studied, especially considering the extent future. The experimental studies suggest to which this practice already has been also that the findings among adolescents implemented in educational settings. may be applicable to young adult college Until better data are available regarding the students. As a whole, this rich research base long-term effectiveness of media literacy, provides strong support for the notion that emphasis—especially for young children smoking in entertainment media plays a and adolescents—should be directed at causal role in smoking initiation among reducing exposure. adolescents, and this role warrants action at the individual and societal levels.

Summary Still more research is needed on the important role of popular entertainment Content analyses of popular entertainment media, such as movies, in influencing young media indicate that portrayal of tobacco people to initiate smoking. Research has use is common in movies and is often not yet determined the role entertainment modeled by stars bearing favored social smoking may play in maintaining attributes. The negative health effects of experimental smoking or in prompting tobacco use are rarely depicted. Tobacco relapse among smokers who have quit. portrayal appears to be less common in In addition, no published intervention popular television and music than in studies have evaluated whether adolescents’ movies. Tobacco exposure in online media exposure can be decreased by motivating is an area for further study. parents to restrict access or by teaching adolescents to process depictions of smoking The results of cross-sectional and in movies with more skepticism. longitudinal studies assessing audience responses to portrayals of tobacco use Such research should continue to inform in movies are remarkably consistent in the ongoing effort to reduce exposure showing an association between seeing through media to tobacco use and/or smoking in movies and more positive counteract the effects of such exposure. attitudes toward smoking and adolescent Numerous efforts already have contributed smoking initiation. The population-based to reducing tobacco use in the media. These data include cross-sectional samples from efforts include policy interventions such as different regions of the United States, tobacco advertising and product placement Australia, New Zealand, and Asia, and restrictions, public education, and advocacy a nationally representative sample of efforts targeting entertainment providers. U.S. adolescents—all supporting a link In the future, research on trends—ranging between viewing smoking in movies and from encouraging increased parental adolescent smoking. responsibility to controversial initiatives

411 1 0 . R o l e o f E n t e r t a i n m e n t M e d i a

such as R ratings for movies featuring 4. Cross-sectional studies show that, tobacco use—will continue to build on this among adolescents, exposure to base of knowledge. Continued efforts to smoking in movies is associated with reduce exposure to tobacco through media initiation of smoking, independent of may potentially affect social attitudes and several other factors such as smoking behavior toward smoking, which in turn by friends and family. Cross-sectional may have a long-term effect on the public’s studies also indicate that among disease burden attributable to tobacco use. adolescent never smokers, exposure to smoking in movies is associated with more positive attitudes toward Conclusions smoking. 5. Two longitudinal studies demonstrate 1. Children and adolescents in the that adolescents with higher exposure United States have heavy exposure to to smoking in movies at baseline entertainment media, with an average are 2.0 to 2.7 times more likely to of 5.5 person-hours of media use per try cigarette smoking in the future. day. Tobacco use often is integrated into More studies are needed on the role entertainment media programming, exposure to smoking in movies plays especially in movies. in adolescents’ smoking beyond the 2. Portrayals of tobacco in movies include initiation phase. images of tobacco use and images of 6. Experimental studies show that tobacco product brand names and logos. images of cigarette smoking in film Depictions of smoking are pervasive in can influence adolescent and adult movies, occurring in three-quarters or viewers’ beliefs about social norms for more of contemporary box-office hits. smoking, beliefs about the function Cigar use also is commonly depicted in and consequences of smoking, movies, but use of smokeless tobacco and their personal intentions to is not. Smoking is more common in smoke. Protobacco movie content movies rated for adults (i.e., R-rated), (e.g., stars smoking, absence of health but depiction of smoking is not related consequences portrayed) appears to box-office success. Identifiable to promote prosmoking beliefs and cigarette brands appeared in about intentions. The effects observed for one-third of movies released during experimental studies of smoking in the 1990s. In contrast to its frequent movies on viewers’ smoking-related depiction in movies, tobacco use is found beliefs are of a similar magnitude in about 20% of television shows and as those observed in experimental 25% of music videos. media research on other health topics 3. Smoking prevalence among (e.g., effects of media violence on contemporary movie characters is viewers’ aggression). approximately 25%, about twice what it 7. Experimental studies indicate that was in the 1970s and 1980s. In contrast, antitobacco advertisements screened smoking in the general population before films can partially counter the has declined since the 1970s. Smokers impact of tobacco portrayals in movies. in movies differ from smokers in the general population: the former are 8. The total weight of evidence from more likely to be affluent and white. cross-sectional, longitudinal, and The health consequences of smoking experimental studies, combined with are rarely depicted in movies. the high theoretical plausibility from

412 M o n o g r a p h 1 9 . T h e R o l e o f t h e M e d i a

the perspective of social influences, 10. Efforts to reduce media exposure indicates a causal relationship between to tobacco include restrictions on exposure to movie smoking depictions tobacco advertising and product and youth smoking initiation. placements, advocacy targeted to entertainment providers, media literacy 9. One longitudinal study indicates that interventions aimed at the general parental steps to reduce the exposure public, continued dialogue with key of never smokers (aged 10–14 years) stakeholders in the entertainment to R-rated movies, which have higher industry, and proposed self-regulation numbers of smoking events, produced by the movie industry (e.g., tobacco- a corresponding reduction in their related ratings). smoking initiation.

413 1 0 . R o l e o f E n t e r t a i n m e n t M e d i a

Appendix 10A. Statement by Attorney General Curran of Maryland on Role of the State Attorneys General

EFFORTS OF STATE ATTORNEYS GENERAL TO SEEK MOVIE INDUSTRY COOPERATION IN REDUCING YOUTH EXPOSURE TO SMOKING IN MOVIES

By

J. JOSEPH CURRAN, JR. ATTORNEY GENERAL OF MARYLAND TOBACCO LITIGATION & THE 1998 MASTER SETTLEMENT AGREEMENT (MSA)

When I filed Maryland’s lawsuit in 1996 against the nation’s largest tobacco companies, as did my fellow Attorneys General from across the country, we sought restitution for the billions of dollars paid by our states to treat tobacco related illnesses. Just as important, we also sought to stop the tobacco companies’ marketing campaigns that target and encourage children to purchase and consume tobacco products.

In November 1998, I was one of the 46 state Attorneys General who signed the historic Master Settlement Agreement (MSA) which settled our state suits. Under the MSA, the tobacco companies are required to pay the settling states more than $200 billion over 25 years. Equally important, tobacco companies are restricted from targeting youth or making tobacco brand names ubiquitous through apparel or other merchandise, billboard and bus ads, sponsorships or product placements in the media, including movies.1

The MSA states in part:

No participating tobacco manufacturer may…make, or cause to be made, any payment or other consideration to any person or entity to use, display, make reference to or use as a prop any Tobacco Product, Tobacco Product package, advertisement for a Tobacco Product, or any other item bearing a Brand Name in any motion picture, television show, theatrical production or other live performance, live or recorded performance of music, commercial film or video, or video game (“Media”).…

1The MSA prohibits, generally and with exceptions not listed here: any action to target youth in the advertising or marketing of tobacco products; cartoons in cigarette advertising or packaging; outdoor and transit ads; brand name sponsorships of concerts or sporting events and naming rights to sports venues; tobacco brand name merchandise; free samples of tobacco products; tobacco coupons or credits to children; and payment for use of tobacco products in the media.

414 M o n o g r a p h 1 9 . T h e R o l e o f t h e M e d i a

MSA, Section III (e).2 Moreover, the MSA also prohibits the participating manufacturers from authorizing any third party to use a brand name in a way in which a participating manufacturer may not.3

IN SPITE OF THE MSA PROHIBITIONS, DEPICTIONS OF SMOKING AND BRAND APPEARANCES PERSIST IN THE MOVIES

In spite of these express prohibitions, smoking in movies—particularly in youth rated movies—remains as prevalent today as it was before the MSA—and by some measures has increased. Since the MSA, movie stars continue to smoke on-screen.

Most films portray smokers and smoking in a positive or neutral light and few films appear to contain negative statements about tobacco use. Moreover, even after the MSA, movies continue to show tobacco brand names.

THE TOBACCO COMPANIES DENY A ROLE IN MOVIE BRAND APPEARANCES

In March 2003, California Attorney General Bill Lockyear wrote to each of the four major tobacco companies to express concern over depictions of smoking and tobacco brand appearances since the MSA. In light of the MSA’s express prohibitions, General Lockyear asked each manufacturer whether it had played a role in the appearance of its cigarette brands in post-MSA movies identified in his letters. All four companies denied any role in the appearances of their products in movies. Indeed, at General Lockyer’s urging that the tobacco companies take commercially reasonable steps against brand appearances, Philip Morris, Lorillard and R.J. Reynolds have sent letters notifying movie studios that they do not want their products to appear in the movies. Most recently, we are pleased that Philip Morris and R.J. Reynolds have taken commercially reasonable steps to ask studios to remove references to their tobacco brand names from two particular movies before the films are released on DVD or video or licensed for broadcast.

IN LIGHT OF THE SCIENTIFIC EVIDENCE AND POLICY RECOMMENDATIONS, ATTORNEYS GENERAL SEEK COOPERATION OF THE MOVIE INDUSTRY

In August 2003, compelled by the strength of the research linking seeing smoking in movies with teen smoking, I wrote a letter, joined by the Attorneys General of 27 other states and jurisdictions, to Mr. Jack Valenti, President of the Motion Picture Association of America (MPAA), seeking cooperation of the motion picture industry to reduce the depiction of smoking in movies. Mr. Valenti promptly responded by extending an invitation to my

2Under the MSA, “Tobacco Products” means cigarettes and smokeless tobacco products. Section II (vv). 3MSA Section III(i) provides that “no Participating Manufacturer may license or otherwise expressly authorize any third party to use or advertise within any Settling State any Brand Name in a manner prohibited by this Agreement if done by a Participating Manufacturer itself.… Following such written notice, the Participating Manufacturer will promptly take commercially reasonable steps against any such non-de minimis third party activity.” In other words, tobacco companies are required to take affirmative steps such as written demands to third parties to cease and desist the unauthorized activity.

415 1 0 . R o l e o f E n t e r t a i n m e n t M e d i a

colleagues and me to meet and share with him the details of the study. Mr. Valenti further proposed setting up a round-table in discussion in Los Angeles with representatives of the creative guilds and movie production companies.

My colleagues and I have followed up on Mr. Valenti’s offer, several times over. In October 2003, Connecticut Attorney General Richard Blumenthal. former Pennsylvania Attorney General Mike Fisher, Utah Attorney General Mark Shurtleff, Vermont Attorney General Bill Sorrell, and I met with Mr. Valenti and his staff in Washington, D.C. After presenting the research, Dr. James D. Sargent, a pediatrician and lead investigator of the Dartmouth study, handed Mr. Valenti the following “prescription” which mirrors the policy recommendations endorsed by a growing number of our leading major medical and public health organizations:

n Give smoking movies an R-rating;

n Eliminate brand identification;

n Certify that no consideration was received for smoking in the movie; and

n Run antismoking messages before any movie that depicts smoking.

Although Mr. Valenti unequivocally rejected the R-rating for movies that depict smoking, he proposed a series of round table discussions with other members of the movie industry. Since that initial October 2003 meeting, my colleague attorneys general and I have taken our message, accompanied by Dartmouth scientists Dalton and Sargent, to Hollywood. As proposed by Mr. Valenti, on December 17, 2003, we spent a morning in Los Angeles at the Directors Guild of America (DGA) with their executive staff and directors who serve on the DGA’s Social Responsibility Task Force. Later that same day, we met and discussed the research and its implications for movies and youth smoking with senior production executives of the MPAA studios: Metro-Goldwyn-Mayer Studios, Paramount Pictures, Universal Pictures, Warner Bros., Sony Pictures Entertainment, Walt Disney Pictures, and 20th Century Fox Film Corporation. Representatives of the Screen Actors Guild and the Writers Guild of America also participated in the afternoon discussion. In these two sessions, after Dr. Dalton presented her findings, the attorneys general voiced our concerns directly to these directors, writers, actors and movie studio executives that depictions of smoking in their youth rated films and the persistence of cigarette brand names in any movie works against the goals of the MSA. We encouraged them to adjust and enhance their voluntary movie ratings system—designed to provide America’s parents with the information necessary to make informed and responsible decisions about their childrens’ movie-going choices—so that parents can be as informed about smoking in movies as they currently are about foul language. Given the state attorneys general’s responsibility to enforce the MSA prohibition against cigarette brand placements in the media by tobacco companies, we also asked for the opportunity to learn more from the MPAA studio executives about the circumstances surrounding appearances of cigarette brands in movies.

We also have taken our message to the National Association of Theatre Owners. In April 2004, Vermont Attorney General Bill Sorrell, Dr. Dalton and I had the opportunity to address the NATO Board of Directors at its annual meeting in Washington, D.C. In addition to the Dartmouth research, Dr. Dalton also reviewed the promising findings that antismoking PSAs

416 M o n o g r a p h 1 9 . T h e R o l e o f t h e M e d i a

run before movies can “inoculate” youth to depictions of smoking in films.4 Given NATO’s joint power with the MPAA over the movie ratings system and its members’ exclusive control over their movie screens, NATO has a unique opportunity to protect our youth from smoking by making smoking a criterion in movie ratings (equal to foul language) and by running antismoking PSAs before movies.

And, because we believe that educating the movie industry is a crucial first step toward achieving the changes we seek in reducing youth exposure to smoking depictions and eliminating cigarette brand appearances, we are very pleased to report that the DGA has agreed to feature an article on this important subject in the June issue of its widely circulated magazine. We are hopeful that this message will be communicated most effectively by directors to directors and other movie makers and will guide their creative decisions.

With regard to the MPAA and its member studios, we will continue our educational efforts by seeking mutually agreeable ways to sensitize these individuals and organizations to the public health benefits of reducing youth exposure to smoking depictions and eliminating cigarette brand name appearances.

Most recently, on May 11, 2004, I presented the concerns and efforts of the state attorneys general at a hearing before the United States Senate Committee on Commerce, Science & Transportation which was convened to consider the impact of smoking in movies on children. At the hearing, at which Senator John Ensign presided, Mr. Valenti testified on behalf of the MPAA and Mr. LeVar Burton testified on behalf of the Social Responsibility Task Force of the Directors Guild of America. Dr. Madeline Dalton reviewed the method and compelling findings of the Dartmouth research. Dr. Stanton Glantz argued for the adoption by the movie industry of the four policy recommendations. I was very pleased that Senators Ensign, Ron Wyden and Bill Nelson agree that the movie industry should avail itself of its unique opportunity to eliminate cigarette brand appearances, reduce or eliminate smoking depictions in movies and run antismoking public service announcements in theaters. Moreover, Mr. Burton announced at the hearing that he and other colleagues would donate their time and talent to create antismoking public service announcements.

CONCLUSION

Reducing youth exposure to depictions of smoking and eliminating tobacco brand appearances in movies will require bold, voluntary action by the entire movie industry. The DGA’s pledge to feature this issue in their magazine and Mr. Burton’s willingness to create antismoking PSA’s to be run in theaters are very important and positive steps. I am hopeful that such leadership will prompt similar commitments of resources from the entire movie industry—studios, actors, writers and theater owners—to become part of the solution to the nation’s deadliest preventable problem of smoking.

4Pechmann, C., Shih, C-F. Smoking scenes in movies and antismoking advertisements before movies: effects on youth. J. Marketing. 1999; 63(3) 1-13.

417 1 0 . R o l e o f E n t e r t a i n m e n t M e d i a

Appendix 10B. Letter from 28 State Attorneys General to Jack Valenti and Response

418 M o n o g r a p h 1 9 . T h e R o l e o f t h e M e d i a

419 1 0 . R o l e o f E n t e r t a i n m e n t M e d i a

420 M o n o g r a p h 1 9 . T h e R o l e o f t h e M e d i a

421 1 0 . R o l e o f E n t e r t a i n m e n t M e d i a

Appendix 10C. Letter from Lorillard to California Assistant Attorney General Dennis Eckhart Regarding Brand Appearance of Newport in the Movie City by the Sea

422 M o n o g r a p h 1 9 . T h e R o l e o f t h e M e d i a

11. Wagner, S. 1971. Cigarette country: Tobacco References in American history and politics. New York: Praeger. 1. Roberts, D. F., U. G. Foehr, V. J. Rideout, 12. U.S. Department of Health and Human and M. Brodie. 1999. Kids and media @ the Services. 2001. Women and smoking. A new millennium: A comprehensive national report of the Surgeon General. Rockville, analysis of children’s media use. Menlo Park, MD: U.S. Department of Health and Human CA: Kaiser Family Foundation. http://www Services, Public Health Service, Office of .kff.org/entmedia/1535-index.cfm. the Surgeon General. http://www.cdc.gov/ 2. Woodard IV, E. H., and N. Gridina. 2000. tobacco/data_statistics/sgr/sgr_2001/index Media in the home: The Fifth Annual .htm#full. Survey of Parents and Children. Survey 13. O’Keefe, A. M., and R. W. Pollay. 1996. series 7. Philadelphia: Univ. of Pennsylvania, Deadly targeting of women in promoting Annenberg Public Policy Center. cigarettes. Journal of the American Medical 3. Dalton, M. A., J. D. Sargent, M. L. Beach, Women’s Association 51 (1–2): 67–69. L. Titus-Ernstoff, J. J. Gibson, M. B. Ahrens, 14. Pollay, R. W. 1993. Cigarette advertising J. J. Tickle, and T. F. Heatherton. 2003. and life (1937–1947): The history of Effect of viewing smoking in movies on advertising and archives. Vancouver: adolescent smoking initiation: A cohort Univ. of British Columbia. study. Lancet 362 (9380): 281–85. 15. Tye, L. 1998. The father of spin: Edward L. 4. Distefan, J. M., J. P. Pierce, and E. A. Gilpin. Bernays and the birth of public relations. 2004. Do favorite movie stars influence New York: Crown Publishers. adolescent smoking initiation? American 16. Pautz, M. 2002. The decline in average Journal of Public Health 94 (7): 1239–44. weekly cinema attendance, 1930–2000. 5. Goldberg, M. E. 2003. American media and Issues in Political Economy. http://www the smoking-related behaviors of Asian .elon.edu/ipe/Vol%2011%202002.htm. adolescents. Journal of Advertising Research 17. Ingram, B. 2004. Video vault: Cigarette 43 (1): 2–11. advertising on TV. http://www.tvparty.com/ 6. Goldberg, M. E., and H. Baumgartner. 2002. vaultcomcig.html. Cross-country attraction as a motivation for 18. Mekemson, C., and S. A. Glantz. 2002. How product consumption. Journal of Business the tobacco industry built its relationship Research 55 (11): 901–6. with Hollywood. Tobacco Control 11 Suppl. 7. Sargent, J. D., M. L. Beach, M. A. Dalton, 1: i81–i91. L. A. Mott, J. J. Tickle, M. B. Ahrens, and 19. Lambert, A., J. D. Sargent, S. A. Glantz, T. F. Heatherton. 2001. Effect of seeing and P. M. Ling. 2004. How Philip Morris tobacco use in films on trying smoking unlocked the Japanese cigarette market: among adolescents: Cross sectional study. Lessons for global tobacco control. Tobacco British Medical Journal 323 (7326): 1394–97. Control 13 (4): 379–87. 8. Sargent, J. D., M. L. Beach, M. A. Dalton, 20. Schudson, M. 1984. Advertising, the uneasy L. T. Ernstoff, J. J. Gibson, J. J. Tickle, and persuasion: Its dubious impact on American T. F. Heatherton. 2004. Effect of parental society. New York: Basic Books. R-rated movie restriction on adolescent smoking initiation: A prospective study. 21. Terre, L., R. S. Drabman, and P. Speer. Pediatrics 114 (1): 149–56. 1991. Health-relevant behaviors in media. Journal of Applied Social Psychology 21 9. Tickle, J. J., J. D. Sargent, M. A. Dalton, (16): 1303–19. M. L. Beach, and T. F. Heatherton. 2001. Favourite movie stars, their tobacco use in 22. Mekemson, C., D. Glik, K. Titus, A. Myerson, contemporary movies, and its association A. Shaivitz, A. Ang, and S. Mitchell. 2004. with adolescent smoking. Tobacco Control Tobacco use in popular movies during the 10 (1): 16–22. past decade. Tobacco Control 13 (4): 400–402. 10. Pierce, J. P., and E. A. Gilpin. 1995. 23. American Lung Association of Sacramento. A historical analysis of tobacco marketing 2004. Ten year report on tobacco in the and the uptake of smoking by youth in movies: 1994–2003. the United States: 1890–1977. Health 24. MacKinnon, L., and L. Owen. 1998. Psychology 14 (6): 500–508. Smoking in films: A review. London: Univ.

423 1 0 . R o l e o f E n t e r t a i n m e n t M e d i a

of North London and Health Education JAMA: The Journal of the American Medical Authority. http://www.globalink.org/tobacco/ Association 281 (12): 1131–36. docs/misc-docs/0003films.shtml. 37. Thompson, K. M., and F. Yokota. 2001. 25. Ng, C., and B. Dakake. 2002. Tobacco Depiction of alcohol, tobacco, and other at the movies: Tobacco use in PG-13 substances in G-rated animated feature films. http://static.masspirg.org/reports/ films. Pediatrics 107 (6): 1369–74. TobaccoattheMovies.pdf. 38. Omidvari, K., K. Lessnau, J. Kim, 26. Polansky, J. R., and S. A. Glantz. 2004. First- D. Mercante, A. Weinacker, and C. Mason. run smoking presentations in U.S. movies, 2005. Smoking in contemporary American 1999–2003. http://repositories.cdlib.org/ cinema. Chest 128 (2): 746–54. cgi/viewcontent.cgi?article=1047&context 39. McIntosh, W. D., D. G. Bazzini, S. M. Smith, =ctcre. and S. M. Wayne. 1998. Who smokes in 27. Roberts, D. F., L. Henriksen, and Hollywood? Characteristics of smokers in P. G. Christenson. 1999. Substance use in popular films from 1940 to 1989. Addictive popular movies and music. http://ncadi Behaviors 23 (3): 395–98. .samhsa.gov/govstudy/mediastudy/default 40. Escamilla, G., A. L. Cradock, and I. Kawachi. .aspx. 2000. Women and smoking in Hollywood 28. Hazan, A. R., H. L. Lipton, and S. A. Glantz. movies: A content analysis. American 1994. Popular films do not reflect current Journal of Public Health 90 (3): 412–14. tobacco use. American Journal of Public 41. Omidvari, K., K. Lessnau, C. Mason, and Health 84 (6): 998–1000. A. Weinacker. 2006. Smoking in movies. 29. Glantz, S. A., K. W. Kacirk, and C. McCulloch. Chest 129 (2): 495–96. 2004. Back to the future: Smoking in movies 42. Pechmann, C., and C.-F. Shih. 1999. in 2002 compared with 1950 levels. American Smoking scenes in movies and antismoking Journal of Public Health 94 (2): 261–63. advertisements before movies: Effects on 30. Stockwell, T. F., and S. A. Glantz. 1997. youth. Journal of Marketing 63 (3): 1–13. Tobacco use is increasing in popular films. 43. Glantz, S. A., and J. Polansky. 2006. Smoking Tobacco Control 6 (4): 282–84. in movies. Chest 129 (2): 495–96. 31. Everett, S. A., R. L. Schnuth, and 44. McGuire, W. J. 1985. Attitudes and attitude J. L. Tribble. 1998. Tobacco and alcohol use change. In Handbook of social psychology, in top-grossing American films. Journal of vol. 2, 3rd ed., ed. G. Lindzey and E. Aronson, Community Health 23 (4): 317–24. 233–346. New York: Random House. 32. Dalton, M. A., J. J. Tickle, J. D. Sargent, 45. Bandura, A. 1994. Social cognitive theory M. L. Beach, M. B. Ahrens, and of mass communication. In Media effects: T. F. Heatherton. 2002. The incidence and Advances in theory and research, ed. context of tobacco use in popular movies J. Bryant and D. Zillmann, 61–90. Hillsdale, from 1988 to 1997. Preventive Medicine NJ: Lawrence Erlbaum. 34 (5): 516–23. 46. King 3rd, C., and M. Siegel. 2001. The 33. Sargent, J. D., J. J. Tickle, M. L. Beach, Master Settlement Agreement with the M. A. Dalton, M. B. Ahrens, and tobacco industry and cigarette advertising T. F. Heatherton. 2001. Brand appearances in magazines. New England Journal of in contemporary cinema films and Medicine 345 (7): 504–11. contribution to global marketing of 47. Watson, N. A., J. P. Clarkson, R. J. Donovan, cigarettes. Lancet 357 (9249): 29–32. and B. Giles-Corti. 2003. Filthy or 34. Dozier, D. M., M. M. Lauzen, C. A. Day, fashionable? Young people’s perceptions of S. M. Payne, and M. R. Tafoya. 2005. Leaders smoking in the media. Health Education and elites: Portrayals of smoking in popular Research 18 (5): 554–67. films. Tobacco Control 14 (1): 7–9. 48. McCool, J. P., L. D. Cameron, and K. J. Petrie. 35. Stern, S. R. 2005. Messages from teens on 2001. Adolescent perceptions of smoking the big screen: Smoking, drinking, and imagery in film. Social Science and Medicine drug use in teen-centered films. Journal of 52 (10): 1577–87. Health Communication 10 (4): 331–46. 49. McCool, J. P., L. D. Cameron, and K. J. Petrie. 36. Goldstein, A. O., R. A. Sobel, and 2003. Interpretations of smoking in film by G. R. Newman. 1999. Tobacco and alcohol older teenagers. Social Science and Medicine use in G-rated children’s animated films. 56 (5): 1023–32.

424 M o n o g r a p h 1 9 . T h e R o l e o f t h e M e d i a

50. Plano Clark, V. L., D. L. Miller, J. W. Creswell, 61. Burton, D., S. Sussman, W. B. Hansen, K. McVea, R. McEntarffer, L. M. Harter, C. A. Johnson, and B. R. Flay. 1989. Image and W. T. Mikelson. 2002. In conversation: attributions and smoking intentions among High school students talk to students about seventh grade students. Journal of Applied tobacco use and prevention strategies. Social Psychology 19 (8): 656–64. Qualitative Health Research 12 (9): 1264–83. 62. Grube, J. W., I. L. Weir, and S. Getzlaf. 51. Crawford, M. A. 2001. Cigarette smoking 1984. Own value system, value images, and and adolescents: Messages they see and hear. cigarette smoking. Personality and Social Public Health Reports 116 Suppl. 1: 203–15. Psychology Bulletin 10: 306–13. 52. Nichter, M., M. Nichter, and D. Van Sickle. 63. Barton, J., L. Chassin, C. C. Presson, and 2004. Popular perceptions of tobacco S. J. Sherman. 1982. Social image factors products and patterns of use among male as motivators of smoking initiation in early college students in India. Social Science and middle adolescence. Child Development and Medicine 59 (2): 415–31. 53 (6): 1499–1511. 53. Shields, D. L., J. Carol, E. D. Balbach, and 64. Aloise-Young, P. A., K. M. Hennigan, and S. McGee. 1999. Hollywood on tobacco: J. W. Graham. 1996. Role of the self-image How the entertainment industry understands and smoker stereotype in smoking onset tobacco portrayal. Tobacco Control 8 (4): during early adolescence: A longitudinal 378–86. study. Health Psychology 15 (6): 494–97. 54. MacFadyen, L., A. Amos, G. Hastings, and 65. Sargent, J. D., M. L. Beach, A. M. Adachi- E. Parkes. 2003. “They look like my kind of Mejia, J. J. Gibson, L. T. Titus-Ernstoff, people”—Perceptions of smoking images C. P. Carusi, S. D. Swain, T. F. Heatherton, in youth magazines. Social Science and and M. A. Dalton. 2005. Exposure to movie Medicine 56 (3): 491–99. smoking: Its relation to smoking initiation 55. World Health Organization. 2003. among US adolescents. Pediatrics 116 (5): Bollywood: Victim or ally? A WHO study on 1183–91. the portrayal of tobacco in Indian cinema. 66. McCool, J. P., L. D. Cameron, and K. J. Petrie. http://www.who.int/tobacco/wntd/2003/en/ 2005. The influence of smoking imagery on bollywood-exesum.pdf. the smoking intentions of young people: 56. Distefan, J. M., E. A. Gilpin, J. D. Sargent, Testing a media interpretation model. and J. P. Pierce. 1999. Do movie stars Journal of Adolescent Health 36 (6): 475–85. encourage adolescents to start smoking? 67. Distefan, J. M., and J. P. Pierce. 2004. Evidence from California. Preventive Smoking in movies influences adolescents Medicine 28 (1): 1–11. to start smoking: A longitudinal study. 57. Sargent, J. D., M. A. Dalton, M. L. Beach, American Journal of Public Health 94 (7): L. A. Mott, J. J. Tickle, M. B. Ahrens, and 1239–44. T. F. Heatherton. 2002. Viewing tobacco 68. Pierce, J. P., W. S. Choi, E. A. Gilpin, use in movies: Does it shape attitudes that A. J. Farkas, and R. K. Merritt. 1996. mediate adolescent smoking? American Validation of susceptibility as a predictor of Journal of Preventive Medicine 22 (3): which adolescents take up smoking in the 137–45. United States. Health Psychology 15 (5): 58. Dixon, H. G. 2003. Portrayal of tobacco 355–61. use in popular films: An investigation 69. Sargent, J. D., M. A. Dalton, T. Heatherton, of audience impact. PhD thesis, Univ. of and M. Beach. 2003. Modifying exposure Melbourne. to smoking depicted in movies: A novel 59. Swann, W. B. 1983. Self-verification: approach to preventing adolescent smoking. Bringing social reality into harmony with Archives of Pediatrics and Adolescent the self. In Psychological perspectives on the Medicine 157 (7): 643–48. self, vol. 2, ed. J. Suls and A. G. Greenwald, 70. Gibson, B., and J. Maurer. 2000. Cigarette 33–66. Hillsdale, NJ: Lawrence Erlbaum. smoking in the movies: The influence of 60. Chassin, L., C. C. Presson, S. J. Sherman, product placement on attitudes toward E. Corty, and R. W. Olshavsky. 1981. smoking and smokers. Journal of Applied Self-images and cigarette smoking in Social Psychology 30 (7): 1457–73. adolescents. Personality and Social 71. Dixon, H. G., D. J. Hill, R. Borland, and Psychology Bulletin 7 (4): 670–76. S. J. Paxton. 2001. Public reaction to the

425 1 0 . R o l e o f E n t e r t a i n m e n t M e d i a

portrayal of the tobacco industry in the film Journal of Public Health Policy 5 (2): The Insider. Tobacco Control 10 (3): 285–91. 257–70. 72. Edwards, C. A., W. C. Harris, D. R. Cook, 84. Gerbner, G., L. Gross, M. Morgan, and K. F. Bedford, and Y. Zuo. 2004. Out of N. Signorielli. 1981. Health and medicine the Smokescreen: Does an anti-smoking on television. New England Journal of advertisement affect young women’s Medicine 305 (15): 901–4. perception of smoking in movies and their 85. Sone, T. 1999. Tobacco-related scenes intention to smoke? Tobacco Control 13: in television dramas for young Japanese 277–82. audiences. Tobacco Control 8 (3): 350. 73. Jones, B., and M. Carroll. 1998. The effect 86. Fernandez-Collado, C. F., B. S. Greenberg, of a video character’s smoking status F. Korzenny, and C. K. Atkin. 1978. Sexual on young females’ perceptions of social intimacy and drug use in TV series. Journal characteristics. Adolescence 33 (131): of Communication 28 (3): 30–37. 657–67. 87. Greenberg, B. S., C. Fernandez-Collado, 74. Hines, D., R. N. Saris, and L. Throckmorton- D. Graef, F. Korzenny, and C. K. Atkin. Belzer. 2000. Cigarette smoking in popular 1984. Trends in use of alcohol and other films: Does it increase viewers’ likelihood to substances on television. In Drug abuse. smoke? Journal of Applied Social Psychology Foundation for a psychosocial approach, 30 (11): 2246–69. ed. S. Eiseman, J. A. Wingard, and 75. Anderson, C. A., and B. J. Bushman. 2002. G. J. Huba, 187–97. Farmingdale, NY: The effects of media violence on society. Baywood. Science 295 (5564): 2377–79. 88. Cruz, J., and L. Wallack. 1986. Trends in 76. Groesz, L. M., M. P. Levine, and S. K. Murnen. tobacco use on television. American Journal 2002. The effect of experimental presentation of Public Health 76 (6): 698–99. of thin media images on body satisfaction: 89. Hazan, A. R., and S. A. Glantz. 1995. Current A meta-analytic review. International Journal trends in tobacco use on prime-time of Eating Disorders 31 (1): 1–16. fictional television. American Journal of 77. Cohen, J. 1992. A power primer. Psychological Public Health 85 (1): 116–17. Bulletin 112: 155–59. 90. Christenson, P. G., L. Henriksen, and 78. Petty, R. E., and J. R. Priester. 1994. Mass D. F. Roberts. 2000. Substance use in media attitude change: Implications of the popular prime-time television. http://www elaboration likelihood model of persuasion. .mediacampaign.org/publications/primetime/ In Media effects: Advances in theory and tv_toc.html. research, ed. J. Bryant and D. Zillmann, 91. Byrd-Bredbenner, C., M. Finckenor, and 91–122. Hillsdale, NJ: Lawrence Erlbaum. D. Grasso. 2003. Health related content 79. Conrad, K. M., B. R. Flay, and D. Hill. 1992. in prime-time television programming. Why children start smoking cigarettes: Journal of Health Communication 8 (4): Predictors of onset. British Journal of 329–41. Addiction 87 (12): 1711–24. 92. Gidwani, P. P., A. Sobol, W. DeJong, 80. Balbach, E. D., and S. A. Glantz. 1998. J. M. Perrin, and S. L. Gortmaker. 2002. Tobacco control advocates must demand Television viewing and initiation of smoking high-quality media campaigns: The California among youth. Pediatrics 110 (3): 505–8. experience. Tobacco Control 7 (4): 397–408. 93. Gutschoven, K., and J. Van den Bulck. 2004. 81. Zucker, D., R. S. Hopkins, D. F. Sly, J. Urich, Television viewing and smoking volume in J. M. Kershaw, and S. Solari. 2000. Florida’s adolescent smokers: A cross-sectional study. “Truth” campaign: A counter-marketing, Preventive Medicine 39 (6): 1093–98. anti-tobacco media campaign. Journal of 94. Hancox, R. J., B. J. Milne, and R. Poulton. Public Health Management and Practice 2004. Association between child and 6 (3): 1–6. adolescent television viewing and adult 82. Federal Communications Commission. health: A longitudinal birth cohort study. 1934. Communications Act of 1934. Lancet 364 (9430): 257–62. http://www.cybertelecom.org/notes/ 95. DuRant, R. H., E. S. Rome, M. Rich, communications_act.htm. E. Allred, S. J. Emans, and E. R. Woods. 83. Breed, W., and J. R. De Foe. 1984. Drinking 1997. Tobacco and alcohol use behaviors and smoking on television, 1950–1982. portrayed in music videos: A content

426 M o n o g r a p h 1 9 . T h e R o l e o f t h e M e d i a

analysis. American Journal of Public Health 108. Hong, T., and M. J. Cody. 2002. Presence 87 (7): 1131–35. of pro-tobacco messages on the Web. 96. Whelan, E. M., M. J. Sheridan, K. A. Meister, Journal of Health Communication 7 (4): and B. A. Mosher. 1981. Analysis of coverage 273–307. of tobacco hazards in women’s magazines. 109. Ribisl, K. M., R. E. Lee, L. Henriksen, and Journal of Public Health Policy 2 (1): 28–35. H. H. Haladjian. 2003. A content analysis of 97. Warner, K. E., L. M. Goldenhar, and Web sites promoting smoking culture and C. G. McLaughlin. 1992. Cigarette advertising lifestyle. Health Education and Behavior and magazine coverage of the hazards of 31 (1): 64–78. smoking: A statistical analysis. New England 110. Interactive Digital Software Association. Journal of Medicine 326 (5): 305–9. 2003. Essential facts about the computer 98. Sciacca, J., and N. Antonucci. 2003. Tobacco and video game industry: 2003 sales coverage in popular magazines: 1996–1999. demographics and usage data. Washington, American Journal of Health Behavior 27 (1): DC: Interactive Digital Software Association. 25–34. 111. Haninger, K., and K. M. Thompson. 2004. 99. Amos, A., C. Bostock, and Y. Bostock. 1998. Content and ratings of teen-rated video Women’s magazines and tobacco in Europe. games. JAMA: The Journal of the American Lancet 352 (9130): 786–87. Medical Association 291 (7): 856–65. 100. Chapman, S., Q. Jones, A. Bauman, and 112. Perry, C. L., K. A. Komro, B. Dudovitz, M. Palin. 1995. Incidental depiction of S. Veblen-Mortenson, R. Jeddeloh, R. Koele, cigarettes and smoking in Australian I. Gallanar, K. Farbakhsh, and M. H. Stigler. magazines, 1990–1993. Australian Journal 1999. An evaluation of a theatre production of Public Health 19 (3): 313–35. to encourage non-smoking among elementary age children: 2 Smart 2 Smoke. 101. Amos, A., B. Jacobson, and P. White. 1991. Tobacco Control 8 (2): 169–74. Cigarette advertising policy and coverage of smoking and health in British women’s 113. Lippman, M. 1995. Smokers in Marvel magazines. Lancet 337 (8733): 93–96. comics. Tobacco Control 4 (2): 196. 102. Amos, A., G. Hastings, E. Parkes, and 114. Nakahara, S., M. Ichikawa, and S. Wakai. L. MacFayden. 2000. Smoking and youth 2005. Smoking scenes in Japanese comics: magazines: An exploration of images, a preliminary study. Tobacco Control perceptions and messages in magazines and 14 (1): 71. the implications for cessation among older 115. Doherty, T. 1999. Pre-code Hollywood: Sex, teens. Final report to the Department of immortality, and insurrection in American Health, London. Stirling, Scotland: Center cinema, 1930–1934. New York: Columbia for Tobacco Control Research. Univ. Press. 103. Amos, A. 1992. Style and image: Tobacco and 116. Adato, A. 2002. The education of Meathead. alcohol images. London: Health Education Los Angeles Times, January 20. Authority. 117. World Health Organization. 2003. WNDT 104. Wenger, L., R. Malone, and L. Bero. 2001. 2003: Tobacco free film, tobacco free The cigar revival and the popular press: fashion: Action! http://www.euro.who.int/ A content analysis, 1987–1997. American mediacentre/PR/2003/20030527_1. Journal of Public Health 91 (2): 288–91. 118. People’s Daily Online. 2001. Russian 105. Feit, M. N. 2001. Exposure of adolescent Parliament votes to ban smoking at work girls to cigar images in women’s magazines, and on film. http://english.peopledaily.com 1992–1998. American Journal of Public .cn/english/200106/22/eng20010622_73253 Health 91 (2): 286–88. .html. 106. Plumridge, E. W., L. J. Fitzgerald, and 119. Indlaw Communications Pvt. Limited. G. M. Abel. 2002. Performing coolness: 2006. Ban on smoking scenes to be deferred Smoking refusal and adolescent identities. till July 31. http://www.indlawnews.com/ Health Education Research 17 (2): 167–79. NewsDisplay.aspx?f008d022-5c4a-485c-91f9­ 2309f5670831. 107. Amos, A., C. Currie, D. Gray, and R. Elton. 1998. Perceptions of fashion images from 120. Hollywood Reporter. 2003. Thai censors youth magazines: Does a cigarette make delay film. Hollywood Reporter, February 7. a difference? Health Education Research 121. CTV.ca News. 2004. Keep teens from movies 13: 491–501. with smoking: Activists. http://toronto.ctv

427 1 0 . R o l e o f E n t e r t a i n m e n t M e d i a

.ca/servlet/an/local/CTVNews/20040121/ 132. Smoke Free Movies. 2006. Web site. smoking_movies_040121?hub=Toronto http://www.smokefreemovies.ucsf.edu/ Home%5B/url. ourads/ad_inbedroom_nyt.html. 122. Motion Picture Association of America. 133. Smoke Free Movies. 2006. Bedroom 1. 2004. Classification and ratings. http://www http://www.smokefreemovies.ucsf.edu/ .mpaa.org/FilmRatings.asp. ourads/ad_inbedroom_nyt.html. 123. National Association of Attorneys General. 134. Sargent, J. D., T. F. Heatherton, M. B. Ahrens, 1998. Master Settlement Agreement and M. A. Dalton, J. J. Tickle, and M. L. Beach. amendments. Washington, DC: National 2002. Adolescent exposure to extremely Association of Attorneys General. http:// violent movies. Journal of Adolescent Health www.naag.org/backpages/naag/tobacco/msa. 31 (6): 449–54. 124. Harvard School of Public Health. 2007. 135. Dalton, M. A., M. B. Ahrens, J. D. Sargent, HSPH releases presentations made to L. A. Mott, M. L. Beach, J. J. Tickle, and Motion Picture Association of America T. F. Heatherton. 2002. Relation between on the depiction of tobacco smoking in parental restrictions on movies and movies. Press release. Boston: Harvard adolescent use of tobacco and alcohol. Univ. http://www.hsph.harvard.edu/news/ Effective Clinical Practice: ECP 5 (1): 1–10. press-releases/2007-releases/press04032007 136. Federal Communications Commission. 1996. .html. Telecommunications act of 1996. http://www 125. U.S. Senators. 2007. Letter to MPAA. .fcc.gov/vchip/legislation.html#vchip. http://smokefreemovies.ucsf.edu/pdf/Senate 137. Robinson, T. N. 1999. Reducing children’s %20Letter%20to%20MPAA%2006%2021 television viewing to prevent obesity: %2007.pdf A randomized controlled trial. JAMA: 126. Polansky, J. R., S. Glantz, and K. Titus. The Journal of the American Medical 2007. Six months later: Are MPAA’s tobacco Association 282 (16): 1561–67. ratings protecting movie audiences? Paper. 138. Jensen, J. A., N. J. Hickman 3rd, H. Landrine, Center for Tobacco Control Research and and E. A. Klonoff. 2004. Availability of Education. Tobacco Control Policy Making. tobacco to youth via the Internet. JAMA: The http://repositories.cdlib.org/ctcre/tcpmus/ Journal of the American Medical Association MPAA2007. 291 (15): 1837. 127. Council for Excellence in Government. 139. Reuters. 2007. Disney-branded films to Univ. of California, Norman Lear Center. ban depictions of smoking. http://www 2004. How pro-social messages make their .reuters.com/article/domesticNews/ way into entertainment programming: idUSN2535342320070725. A report to the Carnegie Foundation on 140. Brown, J. A. 2001. Media literacy and the Media, Citizens & Democracy. http:// critical television viewing in education. learcenter.org/pdf/MCDReport.pdf. In Handbook of children and the media, 128. Centers for Disease Control and Prevention. ed. D. G. Singer and J. L. Singer, 681–97. 2006. Entertainment: A powerful tool in Thousand Oaks, CA: Sage. communicating health issues. http://www 141. Wright, P. 1986. Schemer schema: .cdc.gov/tobacco/media_communications/ Consumers’ intuitive theories about countermarketing/entertainment/strategies marketers’ influence tactics. Advances in .htm. Consumer Research 13: 1–3. 129. Business Wire. 2001. STARS and Hollywood 142. Gonzales, R., D. Glik, M. Davoudi, and speak out: Premier and gala set to debut A. Ang. 2004. Media literacy and public new film; “Scene smoking: cigarettes, health: Integrating theory, research, and cinema and the myth of cool” takes a hard practice for tobacco control. American look at a controversial subject. http:// Behavioral Scientist 48 (2): 189–201. findarticles.com/p/articles/mi_m0EIN. 143. Austin, E. W., and K. K. Johnson. 1997. 130. Entertainment Industries Council. 2007. Effects of general and alcohol-specific media http://www.eiconline.org/aboutus/. literacy training on children’s decision 131. Smoke Free Movies. http://smokefreemovies making about alcohol. Journal of Health .ucsf.edu/actnow/encourage_ags.html. Communication 2 (1): 17–42.

428