Tobacco Documents Online

~ A ev-4 RJR Nabisco's Cartoon Camel Promotes SCamel Cigarettes to Children V Joseph R. DiFranza, MD; John W. Richards, Jr, MD; Paul M. Paulman, MD; Nancy Wolf-Gillespie, MA; Christopher Fletcher, MD; Robert D. Jaffe, MD; David Murray, PhD Objectives.-To determine if RJR Nabisco's cartoon-theme advertising is more effective in promoting Camel cigarettes to children or to adults. To deter- mine if children see, remember, and are influenced by cigarette advertising. Design.-Use of four standard marketing measures to compare the effects of Camel's Old Joe cartoon advertising on children and adults. Subjects.-High school students, grades 9 through 12, from five regions of the United States, and adults, aged 21 years and over, from Massachusetts. Outcome Measures.-Recognition of Camel's Old Joe cartoon character, product and brand name recall, brand preference, appeal of advertising themes. Results.-Children were more likely to report prior exposure to the Old Joe cartoon character (97.7% vs 72.2%; P<.0001). Children were better able to identify the type of product being advertised (97.5% vs 67.0%; P<.0001) and the Camel cigarette brand name (93.6% vs 57.7%; P<.0001). Chil en also found-ttie Carrrei ci arette advertisements more appeaiing f P<.000 i`. 1C-aMet's ~ ' s ctgare men as tncreased from 0 5% share o e i eqal children to 32.8%, representing sales estimated at $476 million per ear. Conclusfon Id Joe Camel cartoon advertisements are far more success- ful at marketing Camel cigarettes to children than to adults. This finding is con- sistent with tobacco industry documents that indicate that a major function of to- bacco advertising is to promote and maintain tobacco addiction among children. S WITH the number of US smokers de- clining by about 1 million each year, the tobacco industry's viability is critically dependent on its ability to recruit re- placement smokers. t Since children and teenagers constitute 90% of all new smokers, their importance to the indus- try is obvious.2 Many experts are con- vinced that the industry is actively pro- moting nicotine addiction amongyouth.3.4 See also pp 3145, 3154, and 3185. Spokespersons for the tobacco indus- try assert that they do not advertise to people under 21 years of age, the sole purpose of their advertising being to promote brand switching and brand loy- From the Department o`- Family Practice, University of Massachusetts Medica: School. Fitchburg (Dr Di- Franza); Department of Family Medicine, Medical Col- lege of Georgia. Augusta (Dr Richards); Department of Family Practice. University of Nebraska Medical Cen- ter. Omaha (Dr Paulman and Ms Wolf•Gillespie); De- partment of Family, Community and Emergency Med- icine. University of New Ltexico School of Medicine, Albuquerque (Dr Fletcher) Department of Family Prac- tice, University of Washiny:on. Seattle (Dr Jaffe); and Department of Epidemiology. University of Minnesota. • Minneapolis (Dr Murray). Reprint requests to U~sversity of Massachusetts Medical School, Departr.;nt of Family Practice, 47 Ashby State Rd, Fitchbur, MA 01420 (Dr DiFranza). (JAMA 1991;266:3149-3153) alty among adult smokers.5-1 However, industry advertising expenditures can- not be economically justified on this ba- sis alone.9 This study was therefore un- dertaken to determine the relative im- pact of tobacco advertising on children and adults. There is abundant evidence that to- bacco advertising influences children's images of smoking.10 In Britain, the pro- portion of children who gave "looks tough" as a reason for smoking declined after tough images were banned from ci arette advertisements.tt hi ren as young as t e age o years can reliably recall tobacco advertisements12 and match 9ersonality sketches with the I brands usinfr that imaf?ei _.10 n act; cigarette advertising estab ishes such imagery among children who are cog- nitively too immature to understand the purpose of advertising.13 Subsequently, children who are most attuned to ciga- rette advertising have the most positive attitudes toward smoking, whether or not they already smoke. tt Children who are more aware of, or who approve of, cigarette advertisements are more likely to smoke,10•"•t1-tr and those who do smoke buy the most heavily advertised brands.ra.r7 ~Historicall,y, one;brand that children ! / have not bought is Camel. In seven sur- veys, involving 3400 smokers in the sev- enth through 12th grades, conducted be- tween 1976 and 1988 in Georgia, Loui- siana, and Minnesota, Camel was given as the preferred brand by less than 0.5% (Saundra MacD. Hunter, PhD, Weihang Bao, PhD, Larry S. Webber, PhD, and Gerald S. Berenson, MD,e unpublished data, 1991; D.M., unpublished data, 1991).t'•ts•19 In 1986, Camels were most popular with smokers over the age of 65 years,-of whom 4.4% chose Camels, and least popular among those 17 to 24 years of -age, of -whom -only -2.7%--prPfprred Camels.' In 1988, RJR Nabisco launched the "smooth character" advertising cam- paign, featuring Old Joe, a cartootl camel modeled after James Bond and Don Johnson of ` Miami Vice."21 Manv ilans- try analysts believe that the goal of this campaign is to reposition Camgl to c90- pete with Philip Mol-ris' Marlboro brand for the ille al children's market se nt. To detelmine t e relative impact of Cam- el's Old Joe cartoon advertising on chil- dren and adults, we used four standard marketing measures. 1. Recognition. We compared the pro- portlons o~teenagers and adults aged 21 years and over who recognize Camel's Old Joe cartoon character. 2. Recall. We compared the ability of 2 GC teenag` ers -and adults to recall from a masked Old Joe advertisement the type of product being advertised and the brand name. 3. Appeal. We compared how inter- esting an3 appealing a series of Old Joe cartoon character advertisements were to teenagers and adults. 4. rand pr Pr.nce. We compared brand preferences of teenaged smokers prior to the Old Joe cartoon character campaign with those 3 years into the campaign to determine if the campaign had been more effective with children or with adults, and to determine if Camel had been repositioned as a children's brand. METHODS Subjects Since adolescent brand preferences may vary from one geographic location to another (Saundra 141acD. Hunter, Promotion of Camel Cigarettes to Children -DiFranza et ai 3149 JAMA, December 11, 1991-Vol 266, No. 22 f! Q m J k.0 00 m ~ tT I i

PhD, WeihangBao, PhD, LarryS. Web- ber, PhD, and Gerald S. Berenson, MD, unpublished data, 1991; D.M., unpub- lished data,1991),14.1$•'s we selected chil- dren from Georgia, Massachusetts, Ne- braska, New Mexico, and Washington, representing five regions. One school in each state was selected based on its ad- ministration's willingness to participate. Schools with a smoking prevention pro- gram focused on tobacco advertising were excluded. A target of 60 students in each grade, 9 through 12, from each school was set. In large schools, classes were selected to obtain a sample representative of all levels of academic ability. Students were . told that the study concerned advertis- ing and were invited to participate anon= ymously. Since adult brand preferences are. available from national surveys, adult subj ects were recruited only at the Mas- sachusetts site. All drivers, regardless of age, who were renewing theirlicenses at the Registry of Motor Vehicles on the days of the study during the 1990-1991 . school year were asked to participate. Since licenses must be renewed in per- son, this is a heterogeneous population. Materials Seven Camel Old Joe cartoon char- acter advertisements were obtained from popular magazines during the 3 years prior to the study. One ad was masked to hide all clues (except Old Joe) as to the product and brand being ad- vertised (Fig 1). The survey instrument collected de- mographic information and information on past and present use of tobacco, in- cludingbrand preference. Childrenwere Fig 1.-Masked Old Joe Camel cartoon advertisement. ~ Fig 2.-A portion of an eight-page Camel advertising supplement. considered to be smokers if they had smoked one or more cigarettes during the previous week. Previously validated questions were used to determine chil- dren's intentions regarding smoking in the next month and year' and their at- titudes toward the advertised social ben- efits of smoking.21,14 Subjects rated the ads as "cool or stu- pid" and `Interesting or boring." Sub- el cwere asked if theLthought Oloe was cool' ad if thev would like to be ends with him. Each positive response to these four questions was scored as a one, a negative response as a zero. The "appeal score" was the arithmetic sum of the responses to these four questions, with the lowest possible score per re- spondent being a zero and the highest a four. Procedure Subjects were first shown the masked ad and asked if they had seen the Old Joe character before. They were then asked to identify the product being ad- _vertised and the brand name of the prod- uct. Subjects who could not answer these questions were required to respond "Don't know" so they would not be able to write in the correct answer when the unmaske4 advertisements were shown. The subjects were then shown, one at a time, the six unmasked advertisements and asked to rate how the advertise- ments and the Old Joe cartoon charac- ter appealed to them. Subjects then com- pleted the remainder of the survey in- strument. Adolescent brand preference data from this study were compared with the data obtained by seven surveys com- pleted prior to the kickoff of Camel's Old Joe cartoon character campaign early in 1988 (Saundra MacD. Hunter, PhD, WeihangBao, PhD, Larry S. Web- ber, PhD, and Gerald S..Berenson, MD, unpublished data, 1991; D.M., unpub- lished data, 1991).1a•1s.19 Tests of significance were made using the Two-tailed Student's t Test for con- tinuous data and the XZ and Fisher's Exact Test for discrete data. A P value of less than .05 was used to define sta- tistical significance. The study was conducted during the 1990-1991 school year. RESULTS A total of 1060 students and 491 sub- jects from the Registry of Motor Vehi- cles were asked to participate. Usable surveys were obtained from 1055 stu- dents (99%) and 415 license renewal ap- plicants (84.5%). Seventy drivers were under 21 years of age, leaving 345 adults aged 21 years or older. Students ranged in age from 12 to 19 years (mean, 15.99 years) and adults from 21 to 87 years (mean, 40.47 years). Females repre- sented 51.0% of the students and 54.8% sented of the adults. Children were much more likely than adults to recognize Camel's Old Joe car- toon character (97.7% vs 72.2%; P<.0001) (Table). It is not plausible that the children were simply saying they had seen Old Joe when they had not, since they also demonstrated a greater familiarity with the advertisement on the two objective measures. When shown the masked advertise- ment, the children were much more suc- cessful than the adults in identifying the product being advertised (97.5% vs 67.0%; P<.0001) and the Camel brand 3150 JAMA, December 11, 1991-Vol 266, No. 22 Promotion of Camel Cigarettes to Children-DiFranza et al

Comparison of Student and Adult Responses to Camel's Old Joe Cartoon Character Advertisements Georgia Students Massachusetts Students Nebraska Students New Mexico Students Washington Students Total Students' Total Adultst Ne k,bjocts$ 212 224 232 210 177 1055 345 H• n Old Joe, % 98.1 99.6 96.6 95.7 98.9 97.7§ 72.2§ Know product, % 95.3 100 97.8 95.7 98.9 97.5§ 67.0§ Know brand, % 92.9 97.3 91.8 90.0 96.6 93.6§ 57.7§ niink ads took cool, % 62.4 54.1 57.4 61.2 55.1 58.0§ 39.9§ Ads are interesting, % 83.4 73.9 77.3 62.3 69.9 73.6§ 55.1§ Like Joe as friend, % 46.2 31.1 33.9 31.4 32.6 35.0§ 14.4§ rhink Joe is coot, % 51.0 38.6 44.1 40.9 40.0 43.0§ 25.7§ Mean appeat scoreji 2.4 2.0 2.1 1.9 2.0 2.1 § 1.4§ Smoke Camel (%)9 29/76 (38.2) 12/55 (21.8) 13/52 (25.0) 23/43 (53.5) 9/35 (25.7) 86/261 (33.0)§ 8/92 (8.7)§ •Age range, 12 to 19 years. tAge range, 21 to 87 years. $This is the total number of subjects in each category; due to incomplete questionnaires, respondents for some questions may be fewec §P<.0001. pSee text for explanation. 9Percentage of smokers who identify Camel as their favorite brand. name (93.6% vs 57.7%; P<•0001)• Even when the analysis was limited to those subjects who were familiar with the Old Joe cartoon character, children were still more likely than adults to remember the product (98.6% vs 89.6%; P<.0001) and the Camel brand name (95.0% vs 79.1%; P<.0001). This confirms that Old Joe cartoon advertisements are more effective at communicating product and brandname'information to children than to adults. Because Massachusetts adults may not be representative of adults in the four states where children were .,yed, the above analyses were re- peated comparing only Massachusetts children and adults. In all cases the dif- ferences between adults and children were significant and of even greater magnitude (P<•0001), excluding the pos- sibility that the above findings were due to a lighter level of advertising expo- sure in the Massachusetts area. On all four measures, the children foundthe Camel cartoon advertisements more appealing than did the adults. Chil- dren were more likely to think the ad- vertisements looked "cool" (58.0% vs may be more familiar with the Old Joe tion of smokers under 18 years of age Camel campaign than adults in general. who choose Camels has risen from 0.5% Camel cigarettes are now most popular to 32.8%. Given that children under 18 with children and progressively less pop year ccount for 3.3% of all cigarette ular with older smokers. sale~d given a national market share About equal proportions of adults of 4.4% for Camel,'~ we compute that s adult market share is actually (28.2%) and children (29.0%) reported Camel's current cigarette use, making it 3.4%. Given a current average price of unlikely that this factor influenced any 153.3 cents per pack,n the illegal sale of of the above findings. Although there Camel cigarettes to children under 18 were some statistically significant dif- years of age is estimated to have risen ferences in the responses of children from $6 million per year prior to the from different regions, these were not cartoon advertisements to $476 million the focus of this study (Table). per year now, accounting for one quar- When c d with nonsmokers, ter of all Camel sales. children ho wer currently smoking From both a legal and moral perspec- gave highe oval ratings to the ad- (Z tive, it is important to determine if the vertisements (mean approval score of , tobacco industry is actively promoting 2.8 for smokers vs 1.8 for nonsmokers; nicotine addiction among youngsters. P<,0001). Approving attitudes to d However from a public health ers ec- cigarette advertisements seem t re- tive it is irrelevant w ether the effects ~ cede actual mokin . Among the - of tobacco advertising on children are smo ng c i ren, t ose who either were intentional. If tobacco a vertismg is a ~ ambivalent about their future smoking proximate cause of disease, it must be intentions or expressed a definite in- addressed accordingly. In the following tention to smoke were more approving discussion we will examine the evidence of the advertisements than those chil- produced by this study, the marketing dren who intended not to smoke (mean practices of the tobacco industry as a approval scores of 2.6 and 1.8, respec- whole as revealed in industry docu- tively; P<.001). ments, and the marketing practices used Children were more likely to smoke if by RJR Nabisco, in particular, to pro- they believed that smoking is pleasur- mote Camel cigarettes. The quotations able (relative risk [RR], 6.6; P<•0001) cited below are from tobacco industry and that it makes a person more popular personnel and from documents obtained ~(RR, 2• 0; P<• 0001), and attractive (RR, during litigation over Canada's ban of 1~ 2.5; P<,0001), all common themes in tobacco advertising. , cigarette advertising. Our data show that children are much 'ting children, those who Among believed nonsmok that - C~ more familiar with Camel's Old Joe car )- smoking would make them more attrac- toon character than are adults. This may t~ iti be because children have more expo- ~veress an were eight intention to times more smoke in likely the to ex next - sure to these a ve isements, orbecause year ° (P<,001). t~ iT advertisements are inherently more 1 appealing to youngsters. The tobacco COMMENT industry has long followed a policy of ~ data demonstrate that in just 3 preferentially placing selected advertise- ~years Our Camel's Old Joe cartoon charac- ments where children are most likely to had an astounding influence on chil- ~ter see them•~',m,~ For example, print ad- %tlren's smoking behavior. The propor- vertisements are placed in magazines 39.9%; P<.0001) or "interesting" (73.6% ~ vs 55.1%; P<.0001). More of the chil- dren thought Old Joe was "cool" (43.0% 1 ~ vs 25.7%; P<.0001) and wanted to be friends with him (35.0% vs 14.4%; P<.0001). The brand preference data revealed a dramatic reversal in the market seg- ment pattern that existed prior to Cam- el's Old Joe cartoon character campaign. Camel was given as the preferred brand by 32.8% of children up dto the age of 18 years who smoked, 23.1% of Massachu= setts adult smokers aged 19 and 20 years, J)and 8•7% of those 21 years of age and o r• The figures for the Nlassachusetts ts were significantly higher than the ~onal market share for Camel, 4.4%,L' suggesting that Massachusetts adults JAMA, December 11, 1991-Vol 266, No. 22 ~ Promotion of Camel Cigarettes to Children-DiFranza et al 3151 I

'specifically l "~ ~ e. ~ peop tiqv_s-4Mear esigned to reach young igarette brand promo- dozens of teen movies.30 ~ C mels e featured in the Walt Disney mo ' Who Framed Roger Rabbit? and Honey I Shrunk the Kids. The industry targets poster adver- tisements for "key youth locations/meet- ing places in the proximity of theaters, records [sic] stores, video arcades, etc." It is common to see Old Joe poster ad- vertisements in n1g Iis, an obvious gath ering spot for young teens. Billboards T-shirtsl baseball caps, posters orting events and entertainment events such as the Camel "Mud and Mon- ster" series are all used to promote Cam- els. All are effective marketing tech- niques for reaching children.3,ze•31s' The fact that children are much more attracted to the themes used in the Old Joe cartoon character advertisements may also explain why they are more familiar with them. The themes used in tobacco advertising that is targeted at children are the result of extensive re- search on children condu~+p- -" eo- bacco in us ry to "learn everything there was to learn about how smoking begins."*'-~ Their research identifies the. major sycholo 'ca1 vulnerabilities of children which c~ en e exnloitedw advertisin to foster and ~~ 'c- otine a on. The m'~et_irlg plan for "Export A" cigarettes describes their "psychologi- cal benefits": "Export smokers will be perceived as ... characterized by their self-confidence, strength of character an m vl ualityw 1 makesthempop- ular anacm1 ed by their peers." Consider a child's vulnerability to peer pressure. According to one industry study, "The goading and taunting that exists at the age of 11 or 12 to get non- smokers to start smoking is virtually gone from the peer group circles by 16 or 17."35'11 If peer influence is virtually gone by the age of 16 years, who is the intended target group for RJR-MacDon- ~ ald's Te~m1o brand, described as indi- • viduals who are "[e]xtremely influenced by their peer group"?40 (RJR-MacDon- ald is a wholely owned subsidiary of RJR Nabisco.) The recommended strategy for promoting this brand is the "[m]ajor usage of imagery which portrays the positive social appeal of peer group ac- ceptance."40 In one Camel advertise- ment, a cowboy (a Marlboro smoker?) is being denied admission to a party be- cause "only smooth characters [ie, Camel smokers] need apply" (Fig 2). It appears that Camel advertisements are also tar- geted at individuals who are influenced by their peer group. Children use tobacco, quite simply, because they believe the benef1,t's out- weigh the risks. To the insecuiKe child, the benefits are the "psycholodical ben- efits" promised in tobacco ments: confidence, an impt; and popularity."•4°•41 Chil lieve that smoking will popular or more attractf times more likely to s Previous research children derive so images of smoki ing. "•13•'' Children advertise- ved image, ren who be- ke them more eareupto4.7 oke.2:'•'" akes it clear that of their positive g from advertis- vho are aware of to- bacco advertising; and those who ap- prove of it, are also more likely to be smokers.10•"•'3-'s Children's favorable at- titudes toward smoking and advertising precede actual tobacco use and corre- late with the child's intention to smoke, , suggesting that the images children de- rive from advertising encourage them to smoke.42 Our data confirm these ear- lier findings. Among nonsmoking chil- dren, those who were more approving of the Old Joe advertisements were more likely either to be ambivalent about their smoking intentions or to express a def- inite intention to smoke. Nonsmoking children who believed that smoking would make them more popular were eight times more likely to express an intention to smoke in the future. ictor of fu- . Since a child's intention to smoke is 7 ' considered to be a gop fiRd ture smoking beha ' `,'~ i ~eems rea- lief in the imagery of Export 'A' against young starter smokers.""" The average age for starter smokers is 13 years.r~' The industry also researches the best ways of keeping children from quitting once they are "hooked on smoking."A1° The purpose of one tobacco industry 7_ study was to assess the feasibility of~`~ marketing low-tar brands to teens as an alternative to quitting.:~0 The study found that for boys, "[tjhe single most com- monly voiced reason for quitting among those who had done so... was sports. "~a` The tobacco industry's sponsorship of sporting events, such as the Camel Su- percross motorcycle race, should be seen in relation to its need to discourage teen- age age boys from quitting. Similarly, its emphasis on slimness serves as a con- , stant stant reinforcement of teenage girls' fears of gaining weight as a sult Yf quitting. - {~, c tiv ~~rr Our study provides further evidenci~ that tobacco advertising promotes and maintains nicotine addiction among chil- dren and adolescents. A t ta9 ban of tobacco advertising and promotions, as part of an effort to protect children from the dangers of tobacco,'~•~ can be based on sound scientific reasoning. This project was supported by grants from the University of Massachusetts Medical Center, the Massachusetts chapter of the American Cancer Society, and Doctors Ought to Care. We would Hke to thank the participating schools and the following for their contributions to this study: the Massachusetts Registry of Motor Vehi- des, Bruce Churchill, MD, Della C. de Baca, Sharon DiFranza, Saundra MacD. Hunter, Melinda Raboin, MD; Mary Sherhart, and the Ca- nadian Council on Smoking and Health. References 1. PierceJP, Fiore MC, NovotnyTE, etal. Trends in cigarette smoking in the United States -projec- tions to the year 2000. JA.~YIA. 1989;261:61-65. 2. Kandel DB, Logan JA. Patterns of drug use from adolescence to young adulthood, I: periods of risk for initiation, continued use, and discontinua- tion. Am J Public Health. 1984;74:660-666. 3. Tye JB. RJ Reynolds targets teens with sophis- ticated marketing campaign. Tobacco Youth Rep. 1987;2(1):1-16. 4. Borsch B. How Madison Avenue seduces chil- sonable to conclude t - psychological benefits of smoking, de- rived from advertising, precedes, and contributes to, the adoption of smoking. There are other lines of evidence in- dicating that tobacco advertising in- creases the number of children who use tobacco. In countri'Ss .vhere advertising has been totally banned or severely re- stricted, the percentage of young peo- ple who smoke has decreased more rap- idly than in countries where to co pro- motion has been less restricte .''~After a 24-year decline in smokeless otiacc sales, an aggressive youth-oriented mar' keting campaign has been followed by what has been termed "an epidemic" of smokeless tobacco use among children; with the average age for new users be- ing 10 years.'1'41Many of the tobacco industry docu- ments cited above provide abundant ev- idence that one purpose of tobacco ad- vertising is to addict children to tobacco. In the words ofgn_e advertisin consult- ant, "Where I worke we were trying very hard to influence kids who were 14 to start to smoke."6i Two marketing strategy documents for Export A also reveal that it is the youngest children they are after.'8•'9 "Whose behavior are we trying to affect?: new users."'s The goal is "[o]ptimizing product and user r' t7 f ~Qojx\dv",~ 3152 JAMA, December 11; 1991-VoI 266, No. 22 dren. Pediatr Management. March 1991:14-24. 5. Voluntcary Initiatives of a Responsible Indus- try. Washington, DC: The Tobacco Institute; 1983. 6. Cigarette Advertising Code. Washington, DC: The Tobacco Institute; 1964:1-8. 7. Code of Cigarette Sampling Practices. Wash- ington, DC: The Tobacco Institute; 1971:1-4. 8. RJ Reynolds Tobacco Company advertisement. Time. April 9, 1984:91. 9. Tye JB, Warner KE, Glantz SA. Tobacco ad- vertising and consumption: evidence of a causal relationship. J Public Health Policy. 1987;8:492- 508. 10. Aitken PP, Leathar DS, O'Hagan FJ, Squair SI. Children's awareness of cigarette advertise- ments and brand imagery. BrJAddict.1987;82:615- 622. 11. Charlton A. Children's advertisement-aware- ness related to their views on smoking. Health Educ J. 1986;45(2)75-78. 12. Aitken PP, Leathar DS, O'Hagan FJ. Chil- dren's perceptions ofadvertisements foccigarettesi Soc Sci d1ed. 1985:21:785-797. ; Promotio[~~Camel Cigarettes to Children-DiFranza et al. 2 0

13. Aitken PP, Leathar DS, Squair SI. Children's awarer+esa of cig•arette brand sponsorship of sports :md games in the U K. Ilealtlr. Educ Res. 1986;1:203- . Goldstein AO, Fischer PM, Richards.JW, Cre- 46 BA. Relationship between high school student smoking and recognition of cigarette advertise- mentn. J Pediatr. 1987;110:4t38-491. 15. Chapman S, Fitzgerald 13. Brand preference and advertising recall in adolescent smokers: some implications for health prqmotion. Ana J Pnblir Il ea l t b.. 1982 ; 72:4 91-4 94. 16. Alexander HM, Callcott R, Dobson AJ, et al. Ciqarette smoking and drug use in schoolchildren, I V: factors associated with changes in smoking be- havior. Int J EpirlenaioL. 198:3;12:59-66. 17. McNeill AD, Jarvis M.J, West RJ. Brand pref- crences among schoolchildren who smoke. Lancet. 15185;2:271-272. 18. Hunter SM, Webber LS. Berenson GS. Ciga- rette smoking and tobacco usage behavior in chil- dren and adolescents: Bogalusa Heart Study. Preo Med. 1980;9:701-712. 19. Hunter SM, Croft JB, Burke GL, Parker FC, Webber LS, Berenson GS. Longitudinal patterns of cigarette smoking and smokeless tobacco use in youth: the Bogalusa Heart Study. Am J Public Health. 1986;76:193-195. 20. Centers for Disease Control. Cigarette brand use among adult smokers-United States, 1986. MMWR. 1990;39:665-673. 21. Bird L. Joe Smooth for President. Adweek's Marketing Week. May 20, 1991:20-22. 22. Barton J, Chassin L, Presson CC. Social image factors as motivators of smoking initiation in early and middle adolescence. Child Dev. 1982;53:1499- 1511. 23. McAlpine KE, DiFranza J. Social attitudes of children to cigarette smoking. Am Assoc Adv Sci. 1988;88:215. Abstract 725. 24. DiFranza JR, Murphy PJ, Ellefsen K. Coun- teradvertising: an image oriented school smoking Presented at the annual meet- of the Society of Teachers of Family Medicine; .ventionprogram. y 8, 1986; San Diego, Calif. 25. PM keeps cig lead. Advertising Age. December 3, 1990:56. ® Franza .JI3„ Tye JI3. Who profits from to- co sales to children? JAMA.1990;263:2784-2787. 27. The Tax Bacrden rlf Totxu.co. Washington, DC: The Tobacco Institttte; 19tHh25:109. 28. Davis RM. Current trends in cigarette advcr- tising and marketing. N Engl J Med. 1987;316:725- 732 29. 1LJ K-MacUanald UocttntentLSFi, p 1980-1(lt./!t- MaaDonald Inc & Intperial Tobacco e The Atlor- ne,r! General of Cttnada). 30. Tye J B. Cigarette ads in kid's movies. Tobacca Youth Rep. 1989;4(l):1-'L. 31. Blum A. Medicine vs Madison Avenue:SJ;hting smoke with smoke. JAMA. 1980;244:739-740. 32. KleinJD, Forehand B, OliveriJ, Patterson GJ, Kupersmidt.Jl3, Strecher V. Candy cigarettes: do they encourage childreei s smoking? Pediatrics. In press. 33. Blum A. The Marlboro Grand Prix: circum- vention of the television ban on tobacco advertis- ing. N Engl J Med. 1991;324:913-917. 3•1. Ledwith F. Does tobacco sports sponsorship on television act as advertising to children? Health Edtcc J. 1984;43:85-85. 35. Report for Imperial Tobacco Ltd. Montreal, Quebec: Kwechansky Marketing Research Inc;1977, Project 16 (RJR-MacDonald Inc & Imperial To- bacco v The Attorney General of Canada). 36. Reportfor Imperial Tobacco Ltd. Kwechansky Marketing Research Inc; 1982. Project PLUS/ MINUS (RJR-MacDonald Inc & Imperial Tobacco v The Attorney General of Canada). 37. Youth 1987. The Creative Research Group Ltd; 1987. Prepared for RJR-MacDonald (RJR-MacDon- ald Inc & Imperial Tobacco v The Attorney Gen- eral of Canada). 38. Report to RJR-MacDonald Inc: Third Family Qualitative Concept Test. Toronto, Ontario: Co- gemie Marketing; 1981 (RJR-MacDonald Inc & Imperial Tobacco v The Attorney General of Can- ada). 39. RJR-MacDonald. Export Family Strategy Doc- ument, 1982 (RJR-MacDonald Inc & Imperial To- bacco v The Attorney General of Canada). 40. J. Walter Thompson. Report to RJR-MacDon- ald: Third Family Creative Direction Recommen- dation, 1984. (RJR-MacDonuzld Inc & Imperial To- bacco o The Attorney General (iCanada). 41. Chassin L, Presson CC, Sherman S.J, Olshav.ky RW. Self-images and cigarette smoking in adoles- cence. Personality Snc Psyr.hol Bnll. 1981;7a370- 676. 42. Aitken Pl', Leathar DS, SquairSl. Children's opinions on whether or not cigarette advertise- mentw should be banned. Healtla Edrrr J. 1986 . 43. Teenage Srnakiug, Imarsedirtte and Long-term. Patterrr^s. Washington, DC: National Institute of Educaticn; 1979. US Dept of Health, Education, and Welfare, contract 400-79-0010. 44. Toxic Substances Board. Health or Tobacco. An End to Tobacco Alrerti.ving and Prmuntion. Wellington. New Zealand: Dept of Health: 1989. 45. US Dept of Health and Human Services, Pub- lic Health Service. The Health Con: equence:e of U.sinq Smokeless Tobacco: A Report qj'tlre Advi- sory Cmnmittee to the Surgeon General. Bethesda. Md: National Institutes of Health; 1986. NIH 86- 2874. 46. Connolly GN, Whm DM, Hecht SS; Henning- field JE, Walker B, Hoffman D. The reemergence of smokeless tobacco. N Engl J Med.1986;314:1020- 1027. 47. Report to Congress Pursuant to the C.igarettrl Smoking Act 1978: An. Action Oriented Research Program for Discocering and Creating the Best Possible Image for Viceroy Cigarettes, Ted Bates Advertising,1975. Washington, DC: FederallYade Commission; 1975. Document AD11345. 48. RJR-MacDonald. ExportABrand, LongTerm Strategy (RJR-MacDonald Inc & Imperial Tobacco v The Attorney General of Canada). 49. RJR-MacDonald. Export Family Strategy Doc- ument (RJR-MacDonald Inc & Imperial Tobacco v The Attorney General of Canada). 50. Johnston LD, O'Malley PM, Bachman JG. Use of Licit and Illicit Drugs by America's High School Students: 1975-1984. Rockville, Md: National In- stitute on Drug Abuse; 1985. 51. Mintz M. Marketing tobacco to children. The Nation. May 6, 1991:cover, 591, 592, 594, 596. 52. Media advertising for tobacco products. JA1fA. 1986;255:1033. 53. Blasi V, Monaghan HP. The First Amendment and cigarette advertising. JAMA.1986;256:502-509. 9 i JAMA, December 11, 1991 -Vol 266, No. 22 Promotion of Camel Cigarettes to Children-DiFranza et al 3153