Farm Road, Henley-on-Thames, Oxon RG9 1EJ, UK
Tel: +44 (0)1491 411000 Fax: +44 (0)1491 571188
|
Published by NTC Publications Ltd Farm Road, Henley-on-Thames, Oxon RG9 1EJ, UK Tel: +44 (0)1491 411000 Fax: +44 (0)1491 571188 |
Vol 12 No 3 (1993) |
Calls for the banning of tobacco advertising rely heavily on the belief that advertising contributes towards children starting or continuing the habit. A number of research studies which claim to demonstrate this connection have been reviewed. Some of these studies have been extensively quoted in anti-smoking literature. This paper is based on a detailed review.
The review found that the studies tended to show a degree of correlation between awareness/approval of advertising and propensity to smoke, but failed to establish a causal link, relying instead on emotive language to convey the impression of cause and effect. From evidence quoted in the studies themselves, it can be demonstrated that, even if advertising can lead children to smoke (which can neither be proved nor disproved), it is likely to be an influence of only minor importance, and much less relevant than other factors such as family environment and peer pressure.
There is no evidence in any of the studies to suggest that, if advertising were banned, it would make the least difference to the propensity of children to smoke. The major weakness of the studies reviewed is that they pay virtually no attention to the more important question of what, in fact, motivates the minority of children who start to smoke. It is hard to resist the conclusion that advertising is selected or attacked mainly because it is a relatively easy target.
When a strong body of opinion wishes to stop a product being used, either for health or other reasons, the first point of attack is always advertising. In the case of tobacco, this strategy has had considerable success. The advertising of tobacco has been totally banned in some countries and partially banned (eg on television), or otherwise restricted, in many others. In the European Community, a draft Directive on tobacco advertising, which would have the effect of banning all forms of advertising for tobacco products throughout the EC with the minor exception of certain types of point-of-sale advertising, was adopted by the Commission in May 1991 and approved with amendments by the European Parliament in February 1992. The EC Health Ministers, at their May meeting, delayed voting on this Directive, which is opposed by three countries (Britain, Germany and the Netherlands), with reservations from Denmark and Greece; the vote has been postponed to at least November 1992. Thus tobacco advertising and sponsorship might easily have already been made illegal throughout the European Community; as it is, it remains on the cards.
What is the logic of banning advertising, while the product itself remains legally available (in some countries, indeed, it is produced as a state monopoly) and on display in the shops which sell it? It is difficult to avoid the impression that people attack advertising at least in part because it is an easy target: one can stop advertising simply by passing a law, whereas there are much more serious results if one starts shutting down farms and factories, not to mention the loss of revenue to national exchequers. Furthermore, advertising is visible. If you take it away it is immediately obvious that you have achieved something, and it puts the problem out of mind.
Those who argue for the banning of tobacco advertising rely heavily on the belief that advertising contributes towards the decision taken by children and young people to start smoking. This is their major counter to the tobacco industry's claim that the purpose and function of their advertising is competition between brands and that it does not work to expand the market (which indeed in many places is declining, even where advertising is relatively unrestricted). The opposition says that whether the cigarette market is mature or not is irrelevant, because the industry needs to keep acquiring new smokers to sustain itself at all, and these naturally are young people. This case is often put with some passion, by people who see tobacco advertisers as immoral exploiters of the vulnerable. For example, in a Commons debate about the subject on 2 December 1991, members spoke in these terms:
Some people more cynical than I say that it is a cold-blooded and calculated effort... to recruit sufficient under-age smokers to replace the 300 men and women who die every day from smoking-related diseases.
That is how the tobacco industry looks upon our young people - 'Give them to us young, and we will keep them throughout their adult life'.
The fact is that the whole marketing strategy is to aim at children from the age of 12 upwards, especially girls of that age.
Is there any evidence for or against these views, which are expressed with such emotional intensity? Has it been shown that cigarette advertising plays an important part, whether intended or not, in causing children to start or continue smoking, or is it, as the industry claims, relevant only to brand choice and insignificant as a cause of the habit?
A number of pieces of research have been published in various papers and called in evidence by the opponents of cigarette advertising. Many of them have been conducted by the Advertising Research Unit at the University of Strathclyde with financial help from the Cancer Research Campaign, and there are some studies from other countries. Members of the opposition rely on these studies as apparently demonstrating that advertising and sponsorship encourage children to start, and continue, smoking; a report based on some of them (From the Billboard to the Playground) was widely quoted during the Commons debate referred to above (Hastings et al., 1991).
This paper is an independent review of 12 reports based on a number of these research studies.
All the studies involved interviews with children. In most cases, the sampling and interviewing procedures appear to have been conducted with care and attention to good practice: for example, the Strathclyde researchers conducted interviews at home, with parental permission, using professional interviewers who had been carefully briefed not to show any bias for or against smoking. In some other cases, questionnaires were administered to classes by teachers, in which one has to have less confidence. Two of the studies reported were longitudinal studies: that is, the same children were interviewed a second time when they were a little older to see whether changes in their behaviour could be related to their attitudes.
All the studies produced similar findings which can be summarized as follows.
More specifically, it was found in various studies that:
- likely to have strong brand preferences;
- more aware of the brand preferences of others, such as parents, older siblings or peers;
- more aware of cigarette advertisements;
- more appreciative or less disapproving of cigarette advertising;- more likely to have parents, siblings or friends who smoke;
In summary, what we have is a correlation between awareness/approval of advertising (and knowledge of cigarette brands) and propensity to smoke. Two questions arise: what does such a correlation mean (how do we interpret it), and how important is it anyway?
Having observed correlations of the kind indicated above, the research reports themselves, as well as certain other papers which have used the results to argue the anti-advertising case, have not hesitated to take them as proving that advertising causes children to smoke. The relationship is often expressed in somewhat emotional terms. For example, the ability of most children to recall some cigarette advertising, to recognize examples when shown them, and to understand what the advertisements are trying to suggest (cigarettes are macho, glamorous, colourful, etc.) is taken to show that cigarette advertising is 'getting through to children'. The finding that children who smoke are more likely than the non-smokers to recognize advertisements, to identify the brands from their imagery, and to express a liking for the leading brand's advertising, is interpreted as meaning that cigarette advertising is 'encouraging' children to smoke, that under-age smokers pay more attention to the advertising and 'get more pleasure and reward from it', and that 'this reward helps reinforce their smoking'. And the finding that children who approve of advertising were twice as likely to become smokers in the future as children who disapprove of it is quoted as evidence that cigarette advertising is 'recruiting' children to smoking, 'helping them to start' (Hastings et al., 1991, pp. 5-7).
The short answer to all this is that, statistically, correlations of this kind can never prove a causal link. Even if the most extreme assumptions were fulfilled, so that recognition and favourable attributes to advertising were always and exclusively related to a child becoming or being a smoker, it would still be no more than an inference that the one probably caused the other. We are thrown straight back on the question, is it a reasonable hypothesis? How likely or plausible is it to infer from such findings that advertising causes children to smoke?
When the research studies are perused with this in mind, it becomes apparent that the correlations are the only 'positive' evidence, and much less striking than they seem at first sight. For one thing, they are by no means 'always and exclusively' present; for another, they can easily be accounted for by other explanations which are at least as plausible, if not more so.
To take the 'always and exclusively' point first, most of the research studies present comparisons between smokers and non-smokers in a way which ignores their relative numbers, and therefore imposes a false perspective on the finding. Take, for example, the statement that '85 per cent of young smokers recognized five or more cigarette ads compared with only 63 per cent of non-smokers' (Hastings et al., 1991, p.6). The statement does not mention what can be discovered from the original report of the study, that these figures are based on 568 non-smokers and only 60 smokers: in a randomly-selected sample of young children the smokers are very much a minority. From this information it is easy to calculate that 51 smokers and 358 non-smokers in the sample recognized five or more advertisements. This puts the greater 'sensitivity' of smokers to advertising into quite a different perspective. The proportion of 'advertisement recognizers' who smoke is only 14 per cent: admittedly higher than the 10 per cent of the whole sample who smoke, but not by much; and more than four-fifths of the 'recognizers' were not smokers.
Another example, quoted in the paper Pushing Smoke (Chapman et al., 1988), is taken from the Australian longitudinal study: 'Children who approved of cigarette advertising were twice as likely to become smokers during a follow-up year than were children who disapproved'. The figures on which this is based are as follows: 370 approved of advertising, and 27 per cent of these (ie 100 children) had taken up smoking a year later; 3,712 (ten times as many) disapproved of cigarette advertising, and of these 12.1 per cent (ie 449 children) took up smoking; in addition, there were 946 children who did not know whether they approved or not, and 19.3 per cent of these, ie 183 children, took up smoking. Thus, four times as many disapprovers of advertising as approvers took up smoking, and of those who eventually smoked, nearly three-quarters disapproved of cigarette advertising.
These examples illustrate the danger of only looking at part of the picture, and particularly of comparing percentages without being aware of the relative numbers involved. Seen in perspective, it hardly looks as if advertising is doing much to manipulate children into taking up the habit; on the contrary, only a small minority at best seem to be sensitive to it at all.
Can the correlation, such as it is, be accounted for otherwise than by the causal hypothesis? Certainly it can, and very plausibly. A child who has become interested in smoking for quite different reasons (eg his parents, brothers, sisters or friends smoke), might naturally be expected to notice advertising for cigarettes in just the same way that a man who has just bought a new car will tend to notice other people driving the same make. It is at least as credible to say that children who are interested in, or tolerant of, smoking are likely to notice cigarette advertising because they have already become interested, as it is to say that they become interested because they have seen advertising. The 'cause' is whatever interested them in the first place.
A major weakness of all the research studies is that, having been designed simply to reveal correlations between awareness, etc of advertising and propensity to smoke, they have little to say about children's motivations for smoking: why they find it attractive or acceptable. It is known to be a complex and still little understood area. A recent OPCS survey (Goddard, 1990) made a serious attempt to investigate and record child-smoking habits. The author wrote in her report:
The survey results show that the onset of smoking in children aged 12-14 is seldom a single distinct event, and that children's smoking behaviour is much more erratic than adult smoking...
It seems equally likely that children's attitudes towards smoking are intrinsically different from those of adults - comparatively more intuitive, less rational, less stable, and probably even less closely related to behaviour.
The most important factors which the OPCS survey found to be associated with starting to smoke were:
These results from the OPCS survey are consistent with findings from the other research studies, that the smoking habits of parents, friends and siblings are primary influences on whether children start smoking (and indeed on their choice of brand). Advertising does not appear in the OPCS list as a key factor. The other studies fail to show convincingly that the partial correlation they find with advertising is independent of the primary influences.
There has yet to be a sufficiently complete study of the different reasons why children want to take up a habit against which there are such strong warnings. But it is clear even from the research evidence we have that it has much to do with family example, peer pressure and the environment in which they live. Perhaps adolescent rebellion comes into it, or (as the OPCS results appear to suggest) the desire to compensate for a lack of self-confidence. The minority of children who take up smoking seem to be attracted for reasons like these. It is not in the least surprising if children who have become interested in smoking, for these reasons, start to notice advertised brands and show less disapproval of advertising. It is a much more credible story than the opposite one of advertising as a powerful tempter.
A number of the research studies reported that children appear to adopt a preferred brand even more strongly than do adults, and that this is often the brand most advertised in the area. This has been quoted as another example of advertising being powerfully effective.
The phenomenon appears to be a general one, since it has been reported from different countries and social groups. It is difficult to accept the explanation that it is due to advertising, since, if that were so, one would expect to find brands chosen in proportion to advertising weight; in fact, it is reported that other brands which are substantial advertisers do not obtain a preference share anything like what one would have expected from their advertising share.
We do not know the reason, but there is a probable clue in the fact that the studies reporting this finding were based on relatively homogeneous local areas or groups. Children will pick up the brand most often smoked by their families, most available in their local shops and, perhaps most important of all, the brand which becomes fashionable among their peers. In one example, a study in Glasgow showed that the children who smoked came almost exclusively from working-class homes and tended to prefer Kensitas Club, a predominantly working-class brand (sold locally in packs of ten, which children could more easily buy). Advertising for Kensitas Club was, not surprisingly, particularly strong in the local press in Glasgow. A different study conducted among public-school girls showed Benson & Hedges as the 'fashionable' brand.
It is clear that none of the research evidence proves that advertising causes children to smoke. The circumstantial evidence quoted above from the studies themselves show that it is not a particularly likely hypothesis. If advertising does 'cause' smoking it clearly fails to work in most cases, since most children, even when they enjoy advertising, seem to be unaffected by it. It seems much more likely that children become interested in smoking through family background, peer pressure, etc and naturally, having become interested, they notice advertising more.
We can of course never prove that advertising has never influenced a child, any more than we can prove that it actually has done so. It is therefore unhelpful to set up the argument in terms of proof. What is far more useful is to discuss the importance, or otherwise, of advertising; and the true implication of the findings from these research studies is that advertising is at best of very minor importance, and quite possibly of no importance at all, as a factor in persuading children to start or continue smoking. Remember that we are discussing here only whether advertising can influence the adoption of the habit; we are not concerned with brand information and choice, which is the declared sole purpose of tobacco advertising.
Some of the research study authors argue that tobacco advertising and sponsorship should be banned because they enable children, before they are old enough to know what lies behind it, to absorb a link between cigarettes and desirable sporting events, racing cars, etc, which they enjoy. One qualitative study found that there is a shift in children's response to advertising after they reach secondary school age (at 11 upwards). Before that time, children view advertising and sponsorship at face value and interpret what is shown literally; from the age of 11 they start to have more sophisticated perceptions and to 'interpret' the messages behind the advertising in the same way as adults, making links with other things in their experience.
The authors build on such findings to argue that advertising and sponsorship suggest associations to young children who at the time are unconscious of them and who are vulnerable because they lack the necessary critical awareness. Young children, it is suggested, will see that the motor racing that they enjoy watching has something to do with a cigarette called Embassy, and this will work away in their unconscious minds until, when they are old enough to start thinking about smoking, it resurfaces to provide a background of Embassy being 'all right'. 'It is for these reasons that we have suggested that cigarette brand sponsorship of sports has an insidious quality' (Chapman et al., 1988).
It is easy to make this kind of argument emotively appealing by the use of terms such as 'vulnerable' and 'insidious'; but the suggestion that advertising messages are somehow working subliminally to twist children's minds before they are old enough to know better is a complete invention, for which there is no evidence whatever. The facts, as revealed by the studies, are simply these: young children who enjoy motor racing are able to recognize sponsoring brand names, colours, packs, etc without being able to say what the association means, and older children also recognize the association and are much better able to explain the marketing purpose. It is not surprising that children, with their lively curiosity, will notice brand names and colours, or that older children have more ability to express verbally the meaning of what they have seen. It simply does not follow from these expected and obvious findings that recognizing the brands has anything to do with starting to smoke, either at the time or later.
By the time children are old enough to become interested in smoking, they are quite capable of understanding what advertising and sponsorship are about. They are also exposed to many other influences, depending on where they live and go to school; many of these influences are negative ones which will suggest social disapproval of smoking (school anti-smoking lessons, not being allowed to smoke in trains or public places, etc). It seems likely, although one would like more evidence, that teenage rebellion against 'social acceptability' may sometimes be part of starting to smoke, as well as peer group pressure and anxiety to conform. As we have seen, most children decide to remain non-smokers whether or not they enjoy advertising and sponsorship.
Thus, of all the influences which might help to make smoking attractive to a child, the one suggesting that a particular brand name has advertised or sponsored motor racing seems, even on the evidence shown, to be the most feeble. There is not the slightest indication from any of this research that, if advertising were stopped, there would be any effect whatever on the important factors of peer pressure and the rest. Argument, education, and controlling the sale and public use of the product are far more likely to reduce child smoking than a ban on advertising or sponsorship, because they bear directly on the influences which we have reason to believe are the important ones.
 http://www.warc.com | © Advertising Association |