Tobacco Documents Online

& .~///33 WHY DO JUVENILES START S44OKING?. AN INTERNATIONAL STUDY OF THE ROLE OF ADVERTISING & OTHER CONTRIBUTORY FACTORS IN ARGENTINA, AUSTRALIA, CANADA, HONG KONG, NORWAY, SPAIN, SWEDEN, SWITZERLAND, TURKEY, & THE UNITED KINGDOM. EDITED AND INTRODUCED BY: PROFESSOR J.J. BODDEWYN, Ph.D BARUCH COLLEGE, CITY UNIVERSITY OF NEW YORK Publ~ed by. INTONATIONAL ADVERTISING ASSOCIATION, New York Noveml~er 1987 O~, C~ C~ L~ O BatCo document for PFSFC 1 March 1999

A Report Prepared by: CHILDREN'S RESEARCH UNIT (CRU), London Sponsored by:. INFOTAB, Brussels Published by: INTERNATIONAL ADVERTISING ASSOCIATION, New York Cr', r~ O', BatCo document for PFSFC 1 March 1999

PREFACE The IAA is most interested in the eff~'ts of advertising on children, a so-called vulnerable group. For this reason, we pubLish this study as a valuable addition to the literature.. It is based on very careful research by a most q~ organization employing methodology which wc have examined thoroughly and found to be of the highest standards. It is also rese.a~h which has generated a comparable international data base. The IAA believes in the freedom to advertise all products and services which are legally sold and legally consumed. Further, the IAA beli~es that this freedom is indivis~le in the seine that restrictiom applied to one group of products inevitably lead to erosion of the freedom to advertise other products. These beliefs led us to publish"Tobacco Advertising Bans and Consumption in 16 Countries'(in 1983 and 1986), which clearly showed that the implementation of advertising bans was generally not followed by decreases in overall tobacco consumption. It would be convenient for those against advertising to indict it as a main cause in how and why people buy and consume products. But advertising is only one of the many variables affecting consumer choices. This ten-nation comparative study bears on all of these issues and should be an important addition to a growing literature on them. International Advertising Association World Headquarters 342 Madison Avenue, New York NY 10017, USA (212) 557-1133 0 Cr~ BatCo document for PFSFC 1 March 1999

CONTENTS 1. EDITOR'S INTRODUCTION 2. THE ISSUE 3. BASIC RESEARCH APPROACH 4. RESULTS OVERVIEW 5. KEY FINDINGS 6. JUVENILE SMOKING INCIDENCE 7. THE DIRECT IMPACT OF ADVERTISING ON JUVENILE SMOKING INITIATION 8. STARTING TO SMOKE: KEY FACTORS 9. CONCLUSIONS APPENDICES 9 I1 13 A : The Children's Research Unit (CRU) 18 B : Interviewing Children: G-~neral Comments 19 C : CRU's Research Methods Used in This Study 21 D : Smoking Frequency 23 E : Description of the Reported Surveys 24 F : Comparison of Restrictiom on Tobacco Advertising in Countries Covered in This Report 26 G : Review of the Literature 28 H : Refe~ncas 30 O', G', Lm ¢_m BatCo document for PFSFC 1 March 1999

1. EDITOR'S INTRODUCTION Whether tobacco advertising affects overall tobacco consumption is a complex problem. The bulk of research on this topic points to little or no relationship between the two. A subset of this issue concerns the impact of cigarette advertising on smoking initiation by thcyoung: do juveniles start to smoke because they have been exposed to print and broadcast advertisements? Important Evidence The 10-country comparison* report~ here provides strong evidence that adverr~ingpi~ys a miniscule role in the initiation of smoking by theyoung. Instead, parents, siblings and friends appear to be the determining factors when children start to smoke. New Evidence Such a point has been made and proved before. However, this recent study (1984-1987) provides not only corroborative evidence but also a new an#e by-foeusslng on nine countries where the control of cigarette advertising ranges from a ban (Norway) to rath~ Limited restrictions (Argentina, Hong Kong and Spain), with Australia, Canada, Sweden, Switzerland, Turkey and the United Kingdom standing in between. It establishes that family and peer influences appear to be the determining factors, irrespective of whether the young are exposed to cigarette advertising or not, with all nine countries reporting the similar overwhelming impact of social and cultural influences on juvenile smoking initiation. New Methodology This study also breaks new methodological ground in that, for the fwst time, an established Smoking Prevalence Estimator has been applied internationally to produce a comparable measure among countries with differing tobacco-advertising controls. Besides, instead of using diaries or impersonal questionnaires administered at school, as is common in this field, the present survey used personalinterviews conducted at home.** Again, this is a In'st international methodological breakthrough which has generated a comparable international database about juvenile smoking initiation and incidence. Great care was also exercised in adapting some of the questions to the particular locales, since customs vary from country to country. Finally, afar broader age range (7 to 15/16 years old) of respondems were interviewed than in most other studies, in order to provide a more comprehensive understanding of the factors involved. Implications The findings would seem to challenge the validity of fairly common assertions that the young start to smoke because they have been exposed to cigarette advertising. They also raise questions about the effectiveness oftobac, co advertising bans.*** In Norway, the subjects of the study were too young to have been influenced by cigarette advertising before a ban was imposed in 1975; indeed, some of the subjects ofthe study had not even been born. By contrast, all of the subjects of the study in Spain and Hong Kong had grown up in the presence of cigarette advertising, yet the incidence of smoking among the juveniles studied in Spain and Hong Kong was lower by far than the incidence of smoking among juveniles in Norway. Clearly, factors other than advertising are at play, and they even predominate, so that advertising should not be made into a scapegoat for juvenile smoking. Is the Evidence Believable? This study was initiated and financed by the tobacco industry. No one should question its right to engage in research, any more than research by the antismoking movement should be considered suspicious a priori. The test, instead, should be: "Is it good research?" "Ten countries are compared. 9 of which wet~ suaveyed by the CRU. British government data collected separately ate also included for comparati~ purpose~ ** In Canada, however, interviews were conduc"~-d in shopping-mall locations. **" See Tobacco Advertising Bans and Consumption in 16 Countries. New York: International Advertising Association. 1986. cr~ (=3 d2~ C", BatCo document for PFSFC 1 March 1999

The findings reported here were provided by the London-based Children's Research Unit (CRU), which is an experienced research house that has been employed by both business and govemmem (see Appendix A). CRU relied in part on the methods and findings of studies commissioned by the British Government in 1982 and 1984, but it also drew on its own extensive international research experience to improve on the British Government study's methodology (Appendices B and C detail the research methodology used in the nine country studies). I think that the methodology used by CRU was appropriate, and that the findings are credible - after all, other studies have reached similar conclusions. Particularly relevant in this respect arc the conclusions of a recent study of schoolchildren's smoking in four counmes, sponsored by the World Health Orgnnisation: "The lack of clear differences in smoking habits between countries probably reflects the selection of countries involved in the study in 1983-84. However, since Norway and Fh'dand am countries with restrictive legislation [actually, a ban] on advertising of tobacco products, and the other two countries [Austria and England] are not, a difference might have been expected. No suchsystematic d~fferences are found (emphasis added)." L.E. Aaro ¢t al., "Health Behaviour in School Children: A WHO Cross- national Survey," Health Promotion, I, l (May 1986"), p.32. In any case, readers should reach their own conclusions, with the new data presented here -- collected internationally -- and in a comparable manner, now available for discussing the issue of juvenile smoking initiation. Limitations This study emphasizes the differences among national tobacca>advertising controls. However, the present report also highlights various cultural habits, particularly in Hong Kong, Spain and Turkey. Whilst we need further studies of the role of other cultural factors in the initiation of smoking by the young, this very requirement also applies to those who advocate tobacco-advertising bans and other restrictions all over the world. They ignore or play down the varying impact of cultural values and customs when they propose the same solutions -- bans and restrictions -- everywhere. The findings reported here deal mainly with juvenile smoking initiation, and the survey did not investigate factors accounting for the continuation of smoking behaviour. Only additional research can explain the latter but, meanwhile, the Children's Research Unit's study can be considered to have thoroughly investigated factors influencing the initiation of smoking by youngsters on a cross-national basis. In my editorial role, I have asked the authors of this report to clarify their methods and findings, and to limit their interpretations to what can be reasonably inferred from the data. As such, the following study provides valuable evidence for researchers, policy-makers, advertising practitioners and concerned citizens. Jj. Boddewyn Professor of Marketing/International Business Barueh College, City University of New York 17 I.axing, ton Avenue, New York I0010, USA Tel. (212) 725-3295 j... BatCo document for PFSFC 1 March 1999

2. THE ISSUE There is a growing body ofnationalsmdics identifying factors accounting for "juvenile smoking initiation," that is, the conditions, motivations and pr~pitar.ing circumstances associated with children starting to smoke (see Appendix G for a review of that lit=mmre). Still, whether substantiated or not, there is a fairly widespread belief around the world that advertising constitutes a major factor in this initiation process. In contrast, there has been a lack of systematic cross-national r~earch to compare juvenile smoking initiation under different country conditions. Nations, of course, vary considerably in such factors as values, economic development, political systems and social stratification. Controlling for all of these factors is a daunting task indeed for any rtscaxchcr. However, considering that one common remedy has been proposed to combat juvenile smoking, namely, to ban or severely restrict tobacco advertising, it was highly desirable to compare countries that d~ffer aignfzcantly in terms of public policies, towards tobacco advertising, in order to determine the relative impact of tobacco advert~f~zg on why ~n#es s:an smol~g. . . - To this end, the tobacco industry's international information organisafion (INFOTAB) commissioned the Children's Research Unit (CRU- sec Appendix A) to determine the extent to which tobacco advertising influenced juvenile smoking initiation in a sample of countries selected for their diffetumt regulatory systems regarding the advertising of tobacco products (see Appendix F for further details about national regulatory systems). i: ,J! • iI li i ,i il !O'x r... !c-, G'x BatCo document for PFSFC 1 March 1999

3. BASIC RESEARCH APPROACH The programme of research reported here has involved independent investigations in nine countries, conducted during 1984 (Australia), 1985 (Norway), 1986 (Spain, Hong Kong, Canada and Argentina) and 1987 (Switzerland. Sweden and Turkey). However, UK data, collected by the British government, are also used for the purpose of comparison. Approximately I000 interviews were conducted in each of these nine countries with boys and girls aged 7-15/16 years according to a quota sample. In eight countries, personal interviews were conducted in-home with parental permission and the willing participation of the child. In Canada, respondents were recruited (via their parents) and personally interviewed in private locations within shopping malls (see Appendices C and E for further details). Fast of alL it was important to apply a reliable estimator of smoking prevaltnce so that countries with varying polities towards tobacco-advertising control could be meaningfully compared. CRU therefor¢ decided to apply Bcwley's Smoking Prevalence Estimator (see Appendix D), widely accepted in this field, throughout this interna- tional research project. Since the UK Office of Popular/on Cereuses and Surveys (OPCS) studies (Dobbs & Marsh I and 2* had also applied Bewley's modified estimator in 1982 and 1984, this enabled comparisons to be made between data from the United Kingdom (OPCS) and other countries (CRU studies). The nine countries were chosen as providing important comparisons in relation to the degree of media exposure allowed for tobacco advertising at the time of each survey, as summariscd below (further details are provided in Appendix F): U.K. 1984 Australia 19M Norway 1985 1986 No cigarette or roll-your-own advertising permitted on "IV or radio. Voluntary controls or restrictions exist for cinema, press, sponsorship and posters. No spe~c restrictions on point-of- sal= advertising. No'IV or radio advertising since 1976. Cinema advertising is restricted, but there are no specific restrictions on press, spomorship, posters and point-of-sale advertising. Complete ban on all kinds of tobacco advertising and sponsorship since 1975. HongKong 1986 Canada 1986 Some restrictions apply to "IV and radio advertising. No speci~c restrictions on press advertis- ing and sponsorship. Cinema unx,stricted except in Catalonia. There are also restrictions on posters, which are not allowed in Catalonia. Point-of-sale advertising is only allowed for domestic brands. Some restrictions apply to "IV and radio advertising, but there axe no specific restrictions on cinema, press, posters and point-of-sale advertising, or on sponsorships. No "IV or radio advertising since 1972. There am no restrictions on cinema or pi'ess advertising. Sponsorship is limited to the non-electronic media, and is being phased out of amateur sports. Restrictions apply to posters, but not to point-of-sale advertising. Argemina Some restrictions apply to TV, radio and cinema advertising. There are no rtstrictions on 1986 sponsorship, posters and point-of-sale advertising. Switzerland No TV advertising sinca 1964. Advertising expenditures for tobacco advertising on foreign "IV 1987 and radio stations broadcasting to Swiss audiences have been disallowed since 1982, by voluntary agreement. No specific restrictions apply to dmema advertising although self- imposed restrictiom by cinema distributors do exist. There are restrictions applying to press advertising, sponsorship and posters, but none on point-of-sale advertising. Sweden No commercial advertising for any product on TV and radio. Cinema tobacco advertising has 1987 been prohibited since 1979. There am restrictions on press advertising for tobacco products. Sponsorship and posters here have bccn prohibited since 1979. Restrictions also apply to point-of-sale advertising. Turkey No "IV or radio advertising. No restrictions on cinema advertising for tobacco products. Press 1987 advertising of tobacco products is allowed, as is sponsorship, except for football. No restrictions apply to posters and point-of-sale advertising. The OK OPCS studies provided the baseline essential for conducting this international research programme, and the rmults from all ten coumries (including the United Kingdom) are prtsented herr (in the United Kingdom, Ox there were three saparate studies which are treated here as one - see Appendices D and E). CZD * The numbers in pattmtheses t'd'er to entries in the bibliography (Appendix H) at the ~ of this report. Ox ""4 BatCo document for PFSFC 1 March 1999

The overall approach adopted in this report is to illustrate the picture which emerges from each national survey. It must be str~sed, however, that "r~ular smokers~ are always def'med in the same way as children who smoke at least one cigarette every week, and that'children" always means child~n of comparable age, when this appears to be important (see Appendix D). Against this background, this research report considers the national patterns of smoking, the ways in which smoking seems to start, and the part which advertising may play in this process, against the impact of the social and familial environments. With a study of this complex nature, ther~ are many possible analytical breakdowns of the statistics, such as boys versus girls, those who live in towns versus those who live in the country, and social and culmraJ factors. However, the purpose of the research was to examine differences among nine countries, and the data is therefore presented here by country of study, and, within this, by smokers against non-smokers, and often by age as wen. Further fragmentation would seem likely to bring confusion ra~er than enlightenment with.in the conte~ of this report. Further enquiries about this cross-national survey can be addressed to: Glen Smith, Chin ...... ". Children's Research Unit (CRLD - Albany House -- • Portslade Road London SW8 3DJ England Tel. 01-622 0286 Fax. 01-720 0537 Tlx. 8952387 .l I 'I I I I 5 BatCo document for PFSFC 1 March 1999

4. RESULTS OVERVIEW CRU's cross-national research concentrated on examining the smoking initiation process by juveniles. Particular attention was given to examining the role of advertising within and among countries with diffe~nt approaches to tobacco-advertising controls, ranging from countries with a few or many restrictions to one with a complete ~tver~ing ban (Norway). This comparative study found, in the face of the varying national patterns regarding the control of tobacco adverzising, that it was not possible to predict which country would have the lowest incidence of juvenile smoking. Conversely, juvenile smoking incidence statistics would not help predict which country has the strongest restrictions on tobacco advertising. Clearly, factors other than tobacco advertising and its regulator), control must have played a key role in juvenile smoking initiation and incidence. The research revealed key factors such as the circumstances in which children begin to experiment with smoking;, the role played by the smoking behaviour of parents, brothers, sisters and peers; the ch~enge of daredevilry, together with sodo-adtural factors. The data patterns which have emerged are remarkably similar on acountry-by-country basis, and they show that a combination of personal, family and social factors are the predominant reasons accounting for smoking initiation by juveniles. Such data patterns persist despite the presence or absence of tobacco advertishag. Advertising was also found to be an insignificant factor with respect to the list of reasons advanced by juvenile respondents for starting to smoke. [n all cases, it is apparent that tobacco advertising does not significantly influence the smoking initiation process as far as children and young people concerned Instead, the decision to start smoking involves mos@ a combination of personal, family and social factors. To summarise, the smoking initiation process and the role of advertising have been internationally examined within and between coumries with different approaches to tobacco-advertising controls. Advertising has bccn comistently found to bc irrelevant not only to the smoking initiation process by juveRilcs, but also regarding juvenile smoking incidence. 5. KEY FINDINGS I. Table I reveals that the proportions of 7-15-year-old children smoking at all, wcrc found to be fairly similar in all the countries surveyed, except for Hong Kong and Argentina, which have relatively few restrictions on tobacco advertising, and where a very high proportion of those children have never smoked. Overall, there were rather low levels of regular and occasional smoking (from I to 15 percent) amongst children aged 7-15 years. 2. In all countrias, "regular smokers" (those who smoked at least one cigareue per week) in the I 1-15-year age group ranged from 3% in Argentina and Hong Kong to 13% in England and Norway, and to [6% in Scotland (see Table 2). Again, the connection with tobacco-advertising controls is not evident st all. 3. Theinddenceofregularsmokingamong 15-year-olds (the older age group where comparable intemational dam arc available) was highest in Norway (36%), a country with a total advertising ban on tobacco products, and substantially lower in Hong Kong (1 I%), where there were relatively few restrictions on tobacco advertising. Switzerland had the lowest incidence of 15-year-old smokers (8%) of all the countries in this survey (refer to table on page 7). 4. The start of smoking was found to depend very much on the influence of family and friends, and the chances of a child smoking in a household where there were no other smokers were low (see Chapter 8). 5. The influence of advertising on smoking initiation was found to bc insignificant in relation to the overwhelming pressures of personal and social (family and friends) influences surrounding the potential smoker (see Chapter 7). BatCo document for PFSFC 1 March 1999