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 13 No 2 (1994) |
G.B. Hastings, P.P. Aitken and A.M. MacKintosh
In a recent article in this journal, Colin McDonald (1993)
criticized our research on cigarette advertising and children's
smoking. This paper refutes each of his criticisms and shows that
our research and that of others in this field provides a
convincing argument that cigarette advertising does influence
children's smoking. These arguments are now widely accepted, with
the only prominent disagreement coming from those with most to
lose - the tobacco industry.
Colin McDonald (1993) argues that our own and other people's research into the relationship between tobacco advertising and children's smoking is flawed and cannot live up to its claims. Specifically he alleges that this research claims, but fails, to:
Furthermore, he asserts that emotive language is used to disguise these shortcomings. This article takes each of these criticisms in turn and demonstrates that none of them can withstand close analysis.
Many surveys have now established that there are positive associations between awareness and approval of cigarette advertising and children starting and continuing to smoke. Some accounts of this research - mainly journalistic descriptions summarizing major findings - have used forceful language in its interpretation, pointing to possible causal links. However, no research reports in refereed journals that we are aware of have taken 'correlations as proving that advertising causes children to smoke', as McDonald claims. We, along with most other researchers, would of course agree with McDonald that correlations can never prove a causal link. This is a basic principle of social research and is why those investigating the link between tobacco advertising and smoking, in common with social researchers in other fields, have proceeded by building a plausible case on the basis of circumstantial evidence, rather than seeking an illusive proof.
With tobacco advertising the case is as follows. First, if advertising has an effect on the young we would expect children to be very aware of cigarette advertising. Several studies have shown this to be so. A recent report by the Economics and Operational Research Division of the UK Department of Health (1992) concluded that 'The evidence about awareness is beyond dispute'. Furthermore, recent studies indicate that children are often more aware and appreciative of cigarette advertising than adults are (Rombouts and Fauconnieu, 1988; DiFranza et al.,1991).
Second, if advertising has a rewarding or reinforcing effect on under-age smoking - and encourages uptake - then we would expect under-age smokers - and children who take up the habit - to be more aware and appreciative of cigarette advertising compared with non-smokers and those who remain non-smokers. In support of this view, a number of studies have reported positive associations between awareness/approval of cigarette advertising and children starting or continuing with the habit (eg Alexander et al., 1983; Aitken and Eadie, 1990; Armstrong et al., 1990; Goddard, 1990).
McDonald's feelings about this approach to social research are unclear. On the one hand, he argues that awareness/approval of cigarette advertising are not 'always and exclusively related to a child becoming or being a smoker' and dismisses correlations as 'the only positive evidence'. On the other hand, he accepts that 'we can never prove that advertising has never influenced a child, any more than we can prove that it actually has done so', and proceeds to accept and quote extensively from correlations in Goddard's (1990) work, which in no way meet his 'always and exclusive' criteria.
Despite this inconsistency, McDonald remorselessly applies his 'always and exclusively' criteria to our research, pointing out that many children who are aware or who approve of cigarette advertising do not smoke. This is hardly surprising. A little thought shows that the idea of only accepting correlations which meet the 'always and exclusively' criteria would restrict social research to the description of banalities. A parallel can be drawn with the evidence about the carcinogenic properties of tobacco smoke. The link between tobacco smoking and lung cancer was established using statistical associations that did not meet McDonald's criteria. Some smokers die from lung cancer, but many do not. The majority of people dying from lung cancer are smokers, but some are not. If we followed McDonald's logic, we would not pronounce any link between smoking and lung cancer until all smokers died of the disease and all its victims were smokers.
So yes, the evidence does show that many non-smokers are sensitive to cigarette advertising and many smokers are not. But the real point is that smokers are significantly more sensitive to it than non-smokers and that, as the next section indicates, these differences are independent of other known influences on smoking.
McDonald claims that our research failed to consider the possible importance of other influences on under-age smoking. This is simply not true. Our most recent quantitative research examined a variety of factors known to be associated with under-age smoking (Aitken and Eadie, 1990; Aitken et al., 1991).
He also claims that associations between awareness/approval of cigarette advertising and smoking can be explained in terms of family and peer influences, suggesting that children may become interested in smoking when they have family members and/or peers who smoke and subsequently start to notice cigarette advertising and to show less disapproval of such advertising.
While this is possible, the Aitken and Eadie research mentioned above suggests this is unlikely to be the only explanation. This found that under-age smokers' greater awareness/appreciation of cigarette advertising was independent of age, sex, socio-economic group, appreciation of advertising in general, attitudes towards smoking and smokers, perceived parental attitudes towards under-age smoking, and smoking by parents, siblings and peers. Furthermore, other studies indicating that cigarette advertising influences uptake of the habit also controlled for a range of variables, such as smoking by parents, siblings and peers (Alexander et al., 1983; Armstrong et al., 1990).
Thus although these studies do not prove that cigarette advertising influences under-age smoking, they do not indicate that associations between perceptions of cigarette advertising and smoking cannot be explained entirely in terms of family and peer influences, as McDonald suggests.
Even if we were to accept McDonald's argument, it would still leave questions about what messages children obtain from cigarette advertising. Do they, for example, get any kind of reward or reassurance from it? In order to answer such questions we should remember that smokers tend to be ambivalent about cigarettes. For example, most young smokers agree that smoking is harmful to health. However, the very presence of cigarette advertising can be taken to negate this view. Indeed, Marsh and Matheson (1983) found that 44 per cent of smokers (but only 26 per cent of non-smokers) agreed that 'Smoking can't be really dangerous or the Government would ban cigarette advertising'. Thus it seems likely that many young smokers are getting some kind of reassurance from cigarette advertising. Recent qualitative research, which contributed to the ASA's decision to ban Regal's 'Reg campaign', also supports this view (Hastings et al., 1993). As young smokers in this study said of cigarette adverts:
Most other adverts are saying like 'Don't smoke, it kills you'. These are like saying 'if you want to - just smoke - it's up to you'.
You go to the TV and it says, 'Don't smoke, Don't smoke, Don't smoke' and then you go outside and say you're on a bus and you look and that says 'smoke'. So you feel better about smoking - well when you see these adverts you don't care - feeling guilty.
You feel as if you're not doing something wrong, that you shouldn't be - it's (the cigarette ad) like saying to you, 'you can smoke, nothing's wrong with it'.
McDonald agrees that several studies have shown that under-age smokers tend to show enhanced or heightened preferences for heavily advertised brands. However, he claims that these studies 'were based on relatively homogeneous local areas or groups'. Again this is not true. Some were, but some were not. For example, the OPCS study (Goddard, 1990) was based on a wide sample of schools in England and Wales.
He also suggests that it is difficult to believe that brand choice among children is influenced by advertising, suggesting instead that it is due to parental and peer influences, local availability and the availability of packets of ten. These factors - especially peer influences - probably do influence brand choice. However, they do not provide the whole answer. If children based their choice on parental preferences then the gradation of children's and adults' preferences would be very similar. The evidence shows that this is not the case. The suggestion that choice is influenced by peers does not explain why peers tend to show an enhanced preference for heavily advertised brands.
Nor does the notion of 'availability' provide an adequate explanation. Why would the range available to parents differ so markedly from that available to children? Most top brands are available in packets of ten. For example, according to Goddard's (1990) study the three most popular brands among young smokers were widely available in tens, yet there were marked differences between the children's preferences for these brands: Benson and Hedges, 52 per cent; Embassy, 11 per cent; Silk Cut, 10 per cent.
It is worth noting that McDonald, despite his links with the tobacco industry(1), is perilously close to undermining its public position here - a position which he uncritically repeats elsewhere in his article. The industry claims that cigarette advertising, unlike advertising for almost any other product category, does nothing else but influence brand choice. If advertising affects brand choice among adult smokers, why should it not affect brand choice among under-age smokers? Given the growing body of evidence that children are often more aware of cigarette advertising than adults are, is it so surprising that they tend to show enhanced preferences for more heavily advertised brands?
McDonald claims that there is no evidence in our research that a ban on tobacco advertising would have any effect whatsoever on under-age smoking. This is certainly true. Our research did not examine bans on tobacco advertising. This does not mean, however, that other researchers have not done so. In fact, a recent review of studies of advertising bans in four countries conducted by the Economics and Operational Research Division (EORD) of the UK Department of Health (1992) came to this conclusion: 'In each case the banning of advertising was followed by a fall in smoking on a scale which cannot reasonably be attributed to other factors'.
The EORD also reviewed findings from studies examining associations between tobacco consumption and two other kinds of variation in advertising: differences in controls on tobacco advertising between countries and temporal fluctuations in tobacco advertising expenditure within countries. It concluded that the lower levels of tobacco advertising controls and higher levels of tobacco advertising expenditure are associated with higher levels of tobacco consumption.
These investigations were mainly concerned with overall levels of tobacco consumption, because suitable data referring to short-term fluctuations in under-age smoking are not available. The higher awareness and appreciation of cigarette advertising among children compared with adults suggests that a ban will have its greatest influence on under-age smoking.
On a number of occasions throughout his article McDonald takes exception to what he sees as emotive language being used by us and other researchers in this field. Some of the words he objects to seem fairly innocuous. Thus he suggests that words like 'reward' and 'reinforcement' when describing how advertising works, illustrate our use of emotive language, yet such terms are widely used in accounts of advertising practice (Kleinman, 1977; Day, 1978). For example, Barry Day (1978), one of the best-known names in international advertising, described the important characteristics of good advertisements in these words: 'I believe that all advertising must in some way reassure... I believe an ad should be a small reward'.
McDonald also dislikes our referring to children as 'vulnerable'. This seems odd. Most people, including the Independent Television Commission, accept that children are vulnerable and need special protection. Why, otherwise, is a whole section in the ITC advertising code devoted to children's advertising?
On other occasions we would agree with McDonald. Emotive language has been used. Indeed, McDonald himself uses phrases like 'most feeble' and 'not the slightest indication from any of this research'; and describes conclusions drawn as 'a complete invention'.
So it is true that passions have led to the use of emotive language during this debate. Both sides are guilty of this. Our passions have been raised by the 300 unnecessary deaths that smoking causes each day in the UK alone (Callum et al., 1992) and the hypocritical and cynical behaviour of the tobacco industry. Tobacco industry documents, and statements by its personnel, indicate its public denial that advertising influences young people, but this stance is not taken seriously in private. Whether or not McDonald believes the public pronouncements, the industry clearly does not (see DiFranza et al., 1991, for detailed references).
For example, documents laid before a Canadian court and later passed to the UK Observer newspaper show that children as young as 12 years old were targeted in the marketing strategy of a subsidiary of BAT Industries (Ferriman and Macrae, 1991). If the tobacco industry genuinely believes that cigarette advertising does not influence the young, it would not engage in such activities.
In summary, notwithstanding McDonald's forthright attack on our and other researchers' work, there is now clear evidence that tobacco advertising does influence levels of smoking, especially among children.
This evidence does not, and never will, amount to proof. Proof is simply not possible in social research of this kind. However, it is evidence nonetheless that has been accepted by eminent medical organizations in Britain such as the BMA and the Royal College of Surgeons; by academic researchers the world over; by both the Conservative government and the Labour opposition; by leading figures in the advertising industry, and even (covertly) by the tobacco industry itself. Despite this apparently overwhelming consensus, some will continue to avow the opposite view. Why they do so is unclear. Whatever the reason, the fact remains that the evidence of cigarette advertising's effect on children's smoking is there for those of us who wish to see it.
Much of the research discussed in this article was supported by the Cancer Research Campaign.
(1): A press release by Imperial Tobacco in December 1993 referred to their employing McDonald's company to do work in this area.
 http://www.warc.com | © Advertising Association |