edge and behavior in school age children. J Fam Pract 30:677-680 (1990).

9. Kiburz, D., Jacobs, R., Reckling, F., and Mason, J.:Bicycle accidents and injuries among adult cyclists. Am JSports Med 14: 416-419 (1986).

10. Runyan, C. W., Earp, J. A. L., and Reese, R. P.: Helmetuse among competitive cyclists. Am J Prev Med 7: 232-236(1991).

11. Wasserman, R. C., et al.: Bicyclists, helmets and headinjuries: a rider-based study of helmet use and effective-ness. Am J Public Health 78: 1220-1221 (1988).

12. Cote, T. R., et al.: Bicycle helmet use among Marylandchildren: effect of legislation and education. Pediatrics89: 1216-1220 (1992).

13. Dorsch, M. M., Woodward, A. J., and Somers, R. L.: Dobicycle safety helmets reduce severity of head injury in realcrashes? Accid Anal Prev 19: 183-190 (1987).

14. Wasserman, R. C., and Buccini, R. V.: Helmet protection

from head injuries among recreational bicyclists. Am JSports Med 18: 96-97 (1990).

15. Vulcan, A. P., Cameron, M. H., and Watson, W. L.:Mandatory bicycle helmet use: experience in Victoria, Aus-tralia. World J Surg 16: 389-397 (1992).

16. DiGuiseppi, C. G., Rivara, F. P., and Koepsell, T. D.:Attitudes toward bicycle helmet ownership and use byschool-age children. Am J Dis Child 144: 83-86 (1990).

17. Williams, A. F.: Observed child restraint use in automo-biles. Am J Dis Child 130: 1311-1317 (1976).

18. Runyan, C. W., and Runyan, D. K.: How can physiciansget kids to wear bicycle helmets? A prototypic challenge ininjury prevention. [Editorial] Am J Public Health 81:972-973 (1991).

19. Cushman, R., Down, J., MacMillan, N., and Waclawik,H.: Helmet promotion in the emergency room following abicycle injury: a randomized trial. Pediatrics 88: 43-47(1991).

Smoking Behavior of Adolescents Exposedto Cigarette Advertising

GILBERT J. BOTVIN, PhDCATHERINE J. GOLDBERG, PhDELIZABETH M. BOTVIN, PhDLINDA DUSENBURY, PhD

The authors are with Cornell University Medical College,Department of Public Health, Institute for Prevention Research.

Tearsheet requests to Gilbert J. Botvin, PhD, Cornell MedicalCollege, Department of Public Health, 411 E. 69th St., room201, New York, NY 10021; tel. (212) 746-1270; fax (212)746-8390.

Synopsis....................................

The purpose of the study was to explore therelationship between the exposure of adolescents inthe seventh and eighth grades to cigarette advertis-

ing and their being smokers. A survey question-naire given to 602 adolescents assessed their expo-sure to cigarette advertising and provided measuresof their smoking behavior, demographic character-istics, and some psychosocial variables.

The results indicated that exposure to cigaretteadvertising and having friends who smoked werepredictive of current smoking status. Adolescentswith high exposure to cigarette advertising weresignificantly more likely to be smokers, accordingto several measures of smoking behavior, than werethose with low exposure to cigarette advertising.The findings extend previous research identifyingfactors that may play a role in the initiation andmaintenance of smoking among adolescents.

CIGARETTE SMOKING is the leading preventablecause of mortality and morbidity in the UnitedStates and has been described as the "most impor-tant public health issue of our time" (1). Morethan 40,000 studies have provided evidence ondeleterious effects of cigarette smoking. A growingbody of clinical and epidemiologic research demon-strates cigarette smoking to be associated withcancer, cardiovascular disease, and chronic obstruc-tive lung disease (1-3). Cigarette smoking has beenidentified as the most widespread form of drugdependence in our society (2).

Cigarette smoking has declined in prevalencefrom more than 50 percent of the populationduring the late 1940s and early 1950s to about 25percent (4). The reductions, in large measure, arethe result of widespread public health campaignson the national, State, and local levels by govern-ment and voluntary organizations. Other contribut-ing factors have been legislative actions, such asbanning cigarette advertising on electronic media,requiring warning labels on cigarette packages andcigarette printed advertising, and more recently,legislation restricting cigarette smoking in public

March-April 1903, Vol. 108, No. 2 217

Table 1. Periodicals read by respondents to a survey of 602adolescents, showing total number of pages, total cost ofcigarette advertising in the periodical, and periodical ranking

by dollar amount of cigarette advertising, 1986

Total TotalPeldical pages dolla Rank

TV Guide ................... 327.48 32,252,868 1Sports Illustrated ............ 290.74 27,698,692 2People ..................... 324.53 23,576,525 3Time ..................... 156.92 21,560,209 4Newsweek .................. 132.28 12,821,946 5Better Homes ............... 117.40 11,654,745 6Family Circle ................ 124.58 10,115,265 7Women's Day ............... 128.29 9,575,703 8Cosmopolitan ............... 141.94 6,910,964 12U.S. News and World Report. 100.49 6,688,071 14Field and Stream ............ 95.77 5,026,825 18Life ..................... 71.00 4,406,555 19Rolling Stone ............... 132.50 4,166,607 20Ebony ..................... 96.31 3,279,056 21U.S. ..................... 139.86 2,832,560 24Outdoor Life ................ 83.85 2,831,335 25Home Mechanix ............. 61.60 1,903,289 39Esquire ..................... 50.70 1,502,996 48Psychology Today........... 51.02 1,324,067 53House and Garden .......... 38.00 921,433 60Discover .................... 31.70 671,644 66Harpers .................... 17.00 131,175 93

Table 2. Newspapers read by respondents of a survey of 602adolescents, showing the number of column inches andpages devoted to cigarette advertising, January-September

1986

Newspaper Column inches Pages

New York Times ............ 9,464 75.11New York Post .............. 5,960 85.14New York Daily News ....... 9,582 114.07Newsday ................... 2,520 30.00

places and on commercial passenger aircraft withinthe United States.However, despite such encouraging develop-

ments, concerns about adverse consequences ofcigarette smoking and little meaningful reduction insmoking onset rates among adolescents (5, 6) haveled to growing interest in the passage of otherlegislation that would impose a total ban oncigarette advertising.

Cigarettes are the most heavily promoted productin the country (7). Currently, $352 million is spentannually on advertising cigarettes in periodicals (8).More than $119 million was spent in 1988 toadvertise one brand of cigarettes, Marlboro (8).Since the 1971 ban on cigarette advertising inelectronic media, advertisers have dramatically in-creased their expenditures for cigarette advertisingin periodicals and other types of promotion (9).

Although cigarette manufacturers have maintainedthat the sole purpose of advertising is to competewith other brands among current smokers, thetobacco industry aggressively recruits new smokers.Tye and coworkers (10) estimated that more than5,000 children and adolescents would need to beginsmoking every day to maintain the current size ofthe smoking population. Since tobacco companiesare not permitted to market cigarettes to childrendirectly, advertising appeals are made indirectly.Tobacco companies sponsor sporting events; sellsouvenirs; display brand-associated cartoon charac-ters on billboards; show cigarette brands in moviesdesigned for young people; and ignore the sale bycandy manufacturers of candy cigarettes havingcigarette brand logos, overlooking copyright in-fringements (11-14).

Research on the relationship between cigaretteadvertising and adolescent smoking has generallyfocused on two factors: awareness of cigaretteadvertising and cigarette brand identification. Ingeneral, research has shown that there is a relation-ship between awareness of cigarette advertisingmessages and adolescent smoking (15-17) and thatcigarette advertising may impact on children asyoung as 3 years of age. Aitken and coworkers (18)found that children as young as 6 years were awareof the presence of cigarette advertisements and thatprimary school children were able to identify cor-rectly cigarette brands in advertisements whenbrand identifications were deleted (19). In a morerecent study, Fisher and coworkers (20) found that30 percent of 3-year-olds surveyed were able tomatch correctly Old Joe, the cartoon camel, withthe tobacco brand.

Children and adolescents are responsive to thevisual images and messages of cigarette advertise-ments. Advertisers present images of smoking thatdownplay health concerns and instead associatesmoking with positive attributes, such as beautyand youth. The majority of advertisements portrayhealthy, enthusiastic, young people engaged inoutdoor or social activities, sports, or feats ofpersonal achievement (21). Since experimentationwith new social behaviors often begins with theimitation of attractive models, who appear to berewarded for their behavior (22), carefully craftedadvertisements using attractive models are likely toincrease the possibility that children and adoles-cents will try cigarettes.

Little empirical evidence concerns the actualprocess of advertisement exposure and behavioralimitation. Atkin and coworkers (23), in a studydesigned to assess teenage exposure to alcohol

218 Public Healh Reports

advertising and its impact on teenage drinking,developed an overall index of their exposure. Theindex consisted of a measure of the reading habitsof teenagers reading various periodicals, weightedby the average number of Advertisements in eachperiodical, the number of advertisements typicallynoticed, the proportion of advertisements viewedfor at least 5 seconds, the number of advertise-ments remembered in six categories of advertise-ments of alcoholic beverages, an estimate of thenumber of times the advertisements were viewed,the amount of attention paid to these advertise-ments, and the number of times other advertise-ments had been viewed for the particular beverage.A dose-response relationship was found betweenexposure to alcohol advertising and adolescentdrinking behavior, with the higher exposure groupreporting the most drinking. In addition, drinkingbehavior was found to be less related to suchfactors as peer influence, social status, communitysize, viewing television characters consuming alco-hol, or viewing public service announcements.Our study was designed to extend our under-

standing of the relationship between cigarette ad-vertising and smoking among junior high schoolstudents. The majority of research studies assessingthe impact of advertising on adolescent tobacco usehave focused primarily on the relationship betweensmoking and cigarette brand identification, or havebeen statistical studies correlating the relationshipbetween demand and promotional expenditures.The focus of this research was the exposure ofadolescent-aged persons to tobacco advertising inpopular periodicals and the effect that periodicaladvertising has on their smoking habits.

Methods

Subjects. In the spring of 1986, 602 junior highschool students from grades seven (28 percent) andeight (72 percent) participated in our survey. Thesample was 48 percent male, 82 percent white, andincluded students from four middle-class, suburbanschools. Most lived in intact families (75 percent).More than one-third of the students indicated thatboth their mother and father had attended college.

Measures. Four measures of smoking behaviorwere on the questionnaire. The 7-point currentsmoking measure assessed how often the subjectsmoked currently, with possible responses rangingfrom never to more than a pack a day. Studentswere asked questions to determine the number ofcigarettes they had smoked the day before, in the

past week, and in the past month. Their intentionto smoke in the future was assessed by asking thestudents if they intended to smoke 2 years later.Demographic characteristics were assessed using

items structured in a multiple choice format. In-cluded were items concerning race or ethnicity,grade, sex, age, and family structure. To assesssocial influences to smoke, data were collected onthe smoking habits of others, such as friends,siblings, and parents; attitudes of parents and peerstoward smoking; and prevalence of smoking bypeers and adults. Smoking by friends was measuredusing a 5-point scale for the number who smoked,ranging from none to all. Smoking by older sib-lings was measured on a 4-point scale for thenumber who smoked, ranging from none to threeor more. Smoking by parents was assessed byseparate questions for the respondents' father andmother, with response categories including have nofather (mother), no, used to but quit, and yes. Theperceived attitudes of respondents' parents towardtheir smoking was assessed on a 5-point scaleranging from strongly against to strongly in favor.Perceived smoking by peers and adults was mea-sured with six categories for each item rangingfrom none to all.Ten true or false items were used to assess

smoking-related knowledge. Included were itemsconcerning the immediate and short-term conse-quences of smoking, smoking prevalence amongadults and peers, and social acceptability. Asser-tiveness was measured using a shortened, 14-itemversion of the Assertion Inventory (24). Items wereincluded on substance refusal assertiveness, saying"no" when someone tries to get you to smoke, andsocial assertiveness skills, such as complimentingfriends. Responses were measured on a 5-pointLikert-type scale ranging from never to always.Locus of control was assessed using five items fromthe Norwicki and Strickland Locus of ControlScale for Children (25).

March-April 1993, Vol. 108, No. 2 219

Relationship between exposure to cigarette advertising and currentsmoking status, respondents to a survey of 602 adolescents,

March-September 1986

Advertising exposure

20

10

Never Less than 1 More than 1smoked pack a week pack a week

Current smoking status

Exposure to cigarette advertising was assessedusing a composite measure to determine how manycigarette advertisements that adolescents reportedreading, defined as "looked at for 5 seconds ormore," weighted by how frequently they wereexposed to each of the periodicals listed, shown intables 1 and 2, in the period of March throughSeptember, 1986. The number of cigarette adver-tisements that adolescents looked at in each period-ical was assessed on a 5-point Likert-type scaleranging from none to most. The frequency ofexposure to these periodicals was assessed on a5-point Likert scale ranging from none to almostalways.

In this way, exposure to cigarette advertising wasassessed both in terms of overall level of exposureto periodicals containing cigarette advertisementsand level of exposure to cigarette advertisementswithin each periodical. Scores were obtained foreach periodical and summed across all periodicalsincluded in the study. One-week test-retest reliabil-ity for the advertisement exposure measure was0.84.

Magazine and newspaper selection criteria. A re-view of the literature did not provide a basis fordetermining which periodicals adolescents are likelyto read with any regularity and which of those con-tain cigarette advertising. Forty-five subjects in 6focus groups of about 5 to 12 students were con-ducted with both urban and suburban junior highschool students (grades 6 to 10). Information ob-tained in preliminary focus groups showed that al-

though the main sources of cigarette advertisingwere in printed material directed to adults, suchmaterial was readily available to adolescents and ofinterest to them. From the details gathered in thefocus groups, a list of 23 periodicals and 5 newspa-pers was constructed, limited to periodicals foundto have the highest percentage of cigarette advertis-ing.

Table 1 lists 22 periodicals read by respondents,with the total number of pages and the total dollarcost of the cigarette advertising in each publicationin 1986. Information on an additional periodicalread by respondents was not available. Table 2presents similar information concerning the amountof cigarette advertising contained in four of thefive newspapers read by respondents, measured inthe number of column inches and pages devoted tocigarette advertising. Information was not availablefor the local newspaper.

Table 3 shows the periodicals included in thestudy and two measures of the amount of cigaretteadvertising contained in each magazine for theperiod of the study, March through September1986. Cigarette advertising is measured both interms of (a) the percentage of all advertisementsthat are for cigarettes, which is the total number ofcigarette advertisements divided by the total num-ber of all advertisements in each magazine and (b)the percentage of cigarette advertisement pages, thetotal number of cigarette advertising pages dividedby the total number of advertising pages.The first measure provided an indication of the

number of cigarette advertisements relative to thetotal number of advertisements. The second pro-vided an indication of the number of cigaretteadvertisement pages relative to the total number ofall advertising pages. The two measures yield simi-lar percentages; differences between the measuresare the result of differences in the size of theadvertisements. Table 1 ranks the periodicals by thetotal cost of cigarette advertising carried by theperiodical in 1986.

Procedure. Data were collected using group admin-istration of the study questionnaire in regularschool classes. Students were told that their an-swers would be confidential and would only beseen by members of the project staff. For the pur-poses of the study, subjects were told that theirdefinition of reading (for the readership survey)should include "browse, look through, or pickup." Subjects who had questions were told by theproject staff to "do the best you can" or "answerto the best of your knowledge."

220 Public Health Reports

Results

Relationship between smoking and advertising. Theresults indicate that exposure to cigarette advertis-ing is significantly correlated with reported smok-ing behavior. The number of cigarettes smoked perday produced the highest correlation with exposureto cigarette advertising (r = 0.50, P = < 0.001).The figure shows the relationship between exposureto cigarette advertising and level of cigarette smok-ing. The seven categories of smoking were col-lapsed into three categories: never (those who hadnever smoked), less than a pack a week (those whoindicated they smoked from about one cigarette perweek to about a pack a week), and more than apack a week (those who indicated that they smokedfrom more than a pack a week to more than apack a day). As the figure shows, subjects whosmoked more than a pack a week had the highestexposure to cigarette advertising, measured in thenumbers of cigarette advertisements in the publica-tions the respondents reported reading.

Factors associated with exposure to cigarette adver-tising. Cor-relations were computed between theadvertisement exposure index and the demographic,psychosocial, social influence, and knowledge vari-ables. A parental smoking index was constructedby combining the reported smoking behavior of thesubject's mother and father. Strong associationswere found between exposure to cigarette advertis-ing and smoking by the subjects' friends (r = 0.28,P = < 0.001). Similarly, subjects' estimates ofsmoking prevalence among their peers (r = 0.18, P= < 0.001) and among adults (r = 0.17, P = <0.001) were associated with exposure to cigaretteadvertising. Exposure to cigarette advertising wasnot found to be associated with smoking by stu-dents' fathers, mothers, or older siblings. Finally,an association was found between locus of controland exposure to cigarette advertising (r = 0.13, P= < 0.01).

Concurrent predictors of cigarette smoking. In ad-dition to the correlational analyses presented, sev-eral multiple regression analyses were used to ex-amine the role of cigarette advertising in promotingand supporting cigarette smoking. All of the vari-ables included in this data set, along with the ad-vertisement exposure measure, were used to predictcurrent smoking. Table 4 shows separate regres-sions that were computed for each of the smokingvariables. The two most important predictors ofcigarette smoking were friends' smoking status and

Table 3. Periodicals read by respondents to a survey of 602adolescents, ranked by percentages of all advertisements thatare for cigarettes, and percentages of all advertising pagesthat have advertisements for cigarettes, March-September

1986

PercentPercent cigarettecigarette advertising

Periodial advertising pages

TV Guide ......................... 19 18U.S. .......................... 17 21Psychology Today ................. 13 15Life .............................. 12 11Field and Stream .................. 9 9People .......................... 8 19Better Homes ..................... 8 9Women's Day ..................... 8 9Rolling Stone ..................... 8 8Sports Illustrated .................. 7 11Ebony .......................... 7 9Family Circle ...................... 7 9Outdoor Life ...................... 6 9Discover .......................... 6 10Home Mechanix ................... 6 10Time .......................... 6 6Harpers .......................... 5 7U.S. News and World Report ....... 5 6Cosmopolitan ..................... 5 5House and Garden ......... ....... 4 5Esquire .......................... 3 5Newsweek ........................ 3 5

Table 4. Results of regression analysis of variables as predic-tors of cigarette smoking by respondents to a survey of 602

adolescents, 1986

Variable beta t

Current smoking status:Friends smoking ................ 0.48 17.88Exposure to advertising .......... 0.17 13.38

2........................... 0.37 ...

Smoked yesterday:Friends smoking ................ 0.25 13.75Exposure to advertising .......... 0.30 15.51

2 . .......................... 0.24...Smoked last week:

Friends smoking ................ 0.23 13.48Exposure to advertising .......... 0.37 17.03

2 . .......................... 0.28...Smoked last month:

Friends smoking ................ 0.38 16.10Exposure to advertising .......... 0.24 14.76Siblings smoking ................ 0.10 22.02

2........................... 0.36 ...

'p < o.oo1, 2p < 0.05.

exposure to cigarette advertising.Table 5 shows the results of the regression using

the same set of independent variables and theintention to smoke in the future as the dependentvariable. Exposure to cigarette advertising andfriends' smoking status emerged as the most signif-icant predictors. Other less significant predictors of

March-April 193, Vol. 108, No. 2 221

Table 5. Results of regression analysis of variables as predic-tors of intention to smoke in the future by respondents to a

survey of 602 adolescents, 1986

Variable befta

Friends smoking .................. 0.40 16.99Exposure to advertising ............ 0.18 '3.96Sibling smoking ................... 0.10 22.06Locus of control ................... 0.12 32.61Assertiveness ..................... -0.15 3- 3.22Single parent household ........... 0.11 22.41

2........................... 0.44 ...

1p< 0.001,2p< 0.05,3P< 0.01.

Table 6. Likelihood of survey respondents being cigarettesmokers, by degree of exposure, 602 adolescents, 1986

Number of respondents

Low HighVariable exposure exposure Rate ratio' 95 percent C/

Ever smoked...... 95 173 1.44 1.19-1.75Smoked in pastmonth ........... 42 93 1.77 1.27-2.45Smoked in pastweek ............ 26 54 1.67 1.08-2.59Smoked in pastday ............ 22 44 1.60 0.98-2.60

Currently smoke .. 32 78 1.93 1.32-2.82Intend to smoke... 47 90 1.52 1.11-2.08

1 Low exposure group is the referent.

intention to smoke in the future included anexternal locus of control, low assertiveness, smok-ing status of older siblings, and living in a single-parent household. The data show that exposure tocigarette advertising is an important predictor ofcurrent smoking and intention to smoke in thefuture.

Smoking risk and cigarette advertising. To examinemore fully the role played by exposure to cigaretteadvertising, we performed a median split on the ad-vertisement exposure index; adolescents who scored14 or lower on the index were placed in a low ex-posure group, and adolescents who scored greaterthan 14 were placed in a high exposure group. Fol-lowing the median split, rate ratios and confidenceintervals were calculated to evaluate differences onthe various smoking measures between high andlow exposure groups. Rate ratios for the smokingmeasures were calculated using the low exposuregroup as the reference group.

Table 6 shows the proportion of adolescents inthe high and low exposure groups who were likelyto smoke. Adolescents in the high exposure groupwere 1.44 times more likely than adolescents in the

low exposure group to have tried cigarette smok-ing, 1.93 times more likely be a current smoker,1.77 times more likely to have smoked in the pastmonth, 1.67 times more likely to have smoked inthe past week, and 1.52 times more likely to reportthat they intended to smoke in the future.

Discussion

Many environmental and psychosocial risk fac-tors predispose adolescents towards starting tosmoke cigarettes, such as parental smoking and lowself-esteem (26). The results of this study point toan additional risk factor: the level of exposure tocigarette advertising. Exposure to cigarette advertis-ing was found to be correlated with several mea-sures of adolescent smoking behavior, and it wasfound to be an important concurrent predictor ofcurrent smoking and intention to smoke in thefuture. Adolescents who reported a high level ofexposure to cigarette advertising were between 1.44and 1.93 times more likely to be smokers thanthose reporting a low level of exposure. They were1.5 times more likely to indicate that they intendedto smoke in the future.The findings are consistent with social learning

theory concerning the power of attractive models inpromoting the adoption of specific behaviors, suchas smoking, which are repeatedly presented asfacilitating the acquisition of desired characteristicsor goals (22). The results support ideas resultingfrom research in self-enhancement that suggest thata modeled behavior is adopted if one identifieswith the image portrayed by the model (27). If themodels depicted in cigarette advertisements portraythe personal ideal of an adolescent, the self-image,as well as the self-esteem of the adolescent, can beenhanced by adopting the smoking behavior.Some support for the modeling hypothesis can be

shown through statistics on cigarette sales. In 1968,when Philip Morris launched the first major pro-motional campaign directed to women (14), only8.4 percent of teenage women smoked (28, 29).Other tobacco companies followed with their ownversion of cigarettes for women. Soon after, ciga-rette products for women became one of the mostheavily advertised products in periodicals for youngwomen. Associated with this phenomenon was asignificant increase in the proportion of teenagewomen smoking cigarettes, which nearly doubledbetween 1968 and 1974 to 15.3 percent (29). Today,the prevalence of cigarette smoking among teenagewomen is slightly higher than it is among teenagemen (6). Young women may be particularly suscep-

222 Public Health Report

tible to cigarette advertising images that portraywomen smokers as slim, attractive, sophisticated,intelligent, and independent.While the findings of our study provide empirical

support for the hypothesis that cigarette advertisingplays a role in the smoking initiation process, thesefindings must be regarded as suggestive rather thanconclusive because of the nature of the study andits reliance on correlational data. An alternativeexplanation of these findings is that the observedassociation between smoking and advertising expo-sure might be the result of selective attention onthe part of smokers, rather than an indication thatexposure to cigarette advertising causes people tobecome smokers. If that were the case, that wouldsupport the consistency theory idea that peopleselectively attend to information that is consonantwith their personal behavior (30).A second limitation concerns the population

studied. Because the majority of the participants inthe study were white, middle-class adolescents, thefindings cannot be generalized to other popula-tions. Previous research has documented racialdifferences in cigarette advertising. Cummings andcoworkers (31) found that periodicals directed toAfrican American readers contained significantlymore cigarette advertising and more advertising formenthol cigarettes than periodicals similar in con-tent but directed to white readers. If a dose-response relationship does exist between cigarettesmoking and advertising exposure, African Ameri-can adolescents may now be at increased risk forbecoming cigarette smokers, since cigarette adver-tisers appear to have increased the amount ofadvertising directed to that population. However,although higher rates of smoking have been ob-served among adult African Americans relative tothat of whites, rates are currently lower amongAfrican American youth (3).A third limitation is that to some extent the

results of this study may have been confounded bypromotional efforts of cigarette companies otherthan advertising. As Blum indicated, cigarette com-panies spend a considerable amount of money onother types of promotional activities, such as spon-soring sports events, point-of-sale promotions, cou-pons, sample give-aways, and other activities de-signed to promote brand name recognition (11).However, evidence from this and other studies

suggests that there may be a causal relationshipbetween cigarette advertising and smoking initia-tion. In addition, a temporal association has beenobserved between cigarette advertising campaignsand increased adolescent smoking (29). Taken to-

gether, there would appear to be ample evidence towarrant more careful scrutiny by legislators of theimpact of cigarette advertising on children andadolescents.

Future research is needed to extend the findingsof this study to other populations to better under-stand the relationship between advertising andsmoking in general and among specific racial andethnic groups in particular. One issue concerns therelative impact of cigarette advertising on smokingbehavior at different points during the adolescentperiod. However, like many other types of phe-nomena, the strongest evidence concerning the roleof advertising in the promotion and maintenance ofsmoking can best be obtained from longitudinalresearch.

References..................................

1. Public Health Service: The health consequences of smo-king: chronic obstructive lung disease. A report of theSurgeon General. DHHS Publication No. (PHS) 84-50205.Office on Smoking and Health, Washington, DC, 1984.

2. Public Health Service: Cigarette smoking as a dependenceprocess. NIDA Research Monograph 23. DHEW Publica-tion No. (ADM) 79-800. National Institute on DrugAbuse. U.S. Government Printing Office, Washington,DC, 1979.

3. Public Health Service: Reducing the health consequencesof smoking: 25 years of progress. A report of the SurgeonGeneral. DHHS Publication No. (CDC) 89-8411. Centersfor Disease Control, Washington, DC, 1989.

4. Public Health Service: Strategies to control tobacco use inthe United States: a blueprint for public health action inthe 1990s. NIH Publication No. 92-3316. National CancerInstitute, Bethesda, MD, 1992.

5. Public Health Service: Healthy people 2000: nationalhealth promotion and disease prevention objectives. DHHSPublication No. (PHS) 91-50212. Office of the AssistantSecretary for Health, Office of Disease Prevention andHealth Promotion. U.S. Government Printing Office,Washington, DC, 1990.

6. Johnston, L., O'Malley, P., and Bachman, J.: Drug useamong high school seniors, college students, and youngadults, 1975-1990. Public Health Service, National Insti-tute on Drug Abuse, Rockville, MD, 1991.

7. Myers, M., et al.: Staff report on the cigarette advertisinginvestigation. Federal Trade Commission, Washington,DC, 1981.

8. Ad dollar summary: January-December, 1988. LeadingNational Advertisers, New York, NY, 1989.

9. Warner, K., and Goldenhar, L.: The cigarette advertisingbroadcast ban and magazine coverage of smoking andhealth. J Public Health Policy 10: 32-42, spring 1989.

10. Tye, J., Warner, K., and Glantz, S.: Tobacco advertisingand consumption: evidence of a causal relationship. JPublic Health Policy 8: 492-507, winter 1987.

11. Blum, A.: The Marlboro Grand Prix: circumvention ofthe television ban on tobacco advertising. N Engl J Med324: 913-917, Mar. 28, 1991.

March-April 1993, Vol. 108, No. 2 223

12. Ledwith, F.: Does tobacco sports sponsorship on televisionact as advertising to children? Health Educ J, 43: 85-88(1984).

13. Tye, J. B.: Cigarette ads in kid's movies. Tobacco YouthReporter 4: 1-2 (1989).

14. Davis, R. M.: Current trends in cigarette advertising andmarketing. N Engl J Med 316: 725-732, Mar. 19, 1987.

15. Goldstein, A., Fischer, P., Richards, J., and Creten, D.:Relationship between high school student smoking andrecognition of cigarette advertisements. J Pediatrics110: 88-491 (1987).

16. Chapman, S., and Fitzgerald, B.: Brand preference andadvertising recall in adolescent smokers: some implicationsfor health promotion. Am J Public Health 72: 491-494(1982).

17. Botvin, E., et al.: Adolescent smoking behavior and therecognition of cigarette advertisements. J Appl Soc Psychol21: 919-932 (1991).

18. Aitken, P., Leathar, D., and O'Hagan, F.: Children'sperceptions of advertisements for cigarettes. Soc Sci Med21: 785-797 (1985).

19. Aitken, P., Leathar, D., O'Hagan, F., and Squair, S.:Children's awareness of cigarette advertisements and brandimagery. Br J Addict 82: 615-622 (1987).

20. Fisher, P. M., et al.: Brand logo recognition by childrenaged 3 to 6 years. JAMA 266: 3145-3148, Dec. 11, 1991.

21. Warner, K. E.: Tobacco industry response to public healthconcern: a content analysis of cigarette ads. Health EducQ 12: 115-127 (1985).

22. Bandura, A.: Social learning theory. Prentice Hall, Engle-wood Cliffs, NJ, 1977.

23. Atkin, C., Hocking, J., and Block, M.: Teenage drinking:does advertising make a difference? J Communication34: 157-167, spring 1984.

24. Gambrill, E., and Richey, C.: An assertion inventory foruse in assessment and research. Behav Ther 6: 550-561(1975).

25. Norwicki, S., and Strickland, B. R.: A locus of controlscale for children. J Consult Clin Psychol 40: 148-154(1973).

26. Botvin, G. J., Schinke, S. P., and Orlandi, M. A.: Psy-chosocial approaches to substance abuse prevention: theo-retical foundations and empirical findings. Crisis 10: 62-77(1989).

27. Kaplan, H.: Patterns of juvenile delinquency: social ori-gins, continuities, and consequences. Sage Publications,Newbury Park, CA, 1984.

28. Albright, C., Altman, D., Slater, M., and Maccoby,N.: Cigarette advertisements in magazines: evidence for adifferential focus on women's and youth magazines.Health Edu Q 15: 225-233 (1988).

29. Public Health Service: The health consequences of smok-ing for women: a report of the Surgeon General. Office ofthe Assistant Secretary for Health, Office on Smoking andHealth. U.S. Government Printing Office, Washington,DC, 1980.

30. Abelson, R. P., et al., editors: Theories of cognitive con-sistency. Rand McNally, Chicago, IL, 1968.

31. Cummings, K. M., Giovino, G., and Mendicino, A. J.:Cigarette advertising and black-white differences in brandpreference. Public Health Rep 102: 698-701, November-December 1987.

OrderP1'essVCocd Superintendent of Documents Subscriptions Order Form* 6175L] YES, please send me subscriptions to PUBLIC HEALTH REPORTS (HSMHA) for $12.00 each per year

($15.00 foreign).

The total cost of my order is $ . Price includes regular domestic postage and handling and is subject to change.

May we ake your aMeladdre avaiUable to oblier milers? El YES L NOCustorces Nane and Address

Phi Csooe Method of Payment:LI (Ceck Payable to the Superintendent of DocumentsL VISA or MasterCard Account

LIJZW (Credt cad exl ration date)

Zip(AuthoxiagSieatue) 1192

Mail To: Superintendent of Documents, P.O. Box 371954, Pittsburgh, PA 15250-7954

224 Public Heath Reports