In accordance with the intentions of health promotion initiatives, recent Australian research suggests that fewer adults and secondary school students are smoking these days. This decline is stronger for men than women (Hill, White and Gray 1991), and for students younger than 16 years (Hill, White, Williams and Gardner 1993).

This article looks at characteristics distinguishing adolescent smokers and non-smokers. Findings are based on data for Box Hill and Berwick families derived from the Institute's Australian Living Standards Study. Although both municipalities contained a mix of classes, Box Hill families generally had a higher socio-economic status than families in Berwick (Brownlee and McDonald 1993).

When Does Smoking Start?

Box Hill parents were less likely to be smokers than Berwick parents, with the prevalence of smokers in Berwick matching that for Australian adults in general. Thirty per cent of fathers and 26 per cent of mothers in Berwick were smokers, while 19 per cent of fathers and 16 per cent of mothers in Box Hill smoked cigarettes. The Institute study found that such differences were due to differences between the two localities in socio-economic status (Weston 1993).

Age at which smoking begins is an important issue, given that the longer the period of smoking, the heavier was the smoking. Of the parents who smoked in both localities, at least 60 per cent took up cigarettes as teenagers - a period when the longer-term, or delayed, effects of smoking seem of little importance to the present (Pierce 1990). The heaviest smokers were Berwick fathers who had been smoking for more than 20 years; the lightest were Box Hill mothers and fathers who had been smoking for a shorter period.

Given that tobacco smoking has been linked with serious health problems including many forms of cancer, coronary heart disease, stroke and chronic lung disease (ABS 1991; AIHW 1992) and that deaths due to tobacco smoking outstrip those related to the use of all other types of drugs (Hill, White Pain and Gardner 1990), identifying those adolescents who are most likely to take up smoking is an important issue for health policy.

Young Smokers

School leavers aged less than 20 years and secondary school students indicated whether or not they smoked and, if so, how many cigarettes they smoked a day. In total, 255 boys (140 from Box Hill and 115 from Berwick) and 275 girls (169 from Box Hill and 103 from Berwick) completed the Institute's questionnaire. The school leavers made up 22 per cent of these young people, and the secondary students 78 per cent.

While Box Hill parents were less inclined to smoke than their Berwick counterparts, the same was not apparent for young people. Of the secondary school students, smoking was reported by similar proportions of boys in Box Hill and Berwick (13 per cent and 11 per cent respectively), while significantly higher proportions of Box Hill girls than Berwick girls indicated smoking (17 per cent and 6 per cent respectively).

In Box Hill, girls who had left school were twice as likely as mothers to smoke (31 per cent and 16 per cent respectively), and in Berwick girls who had left school also seemed more likely than mothers to smoke (42 per cent and 26 per cent respectivley) - however, the 42 per cent is derived from only a very small sample of 19 girls.

Conversely, in Box Hill, smoking patterns were similar for boys who had left school (22 per cent) and fathers (19 per cent), while in Berwick, boys who had left school were less likely than fathers to smoke (18 per cent and 30 per cent respectively).

Overall, higher proportions of girls than boys who had left school smoked (35 per cent and 20 per cent respectively), while only about 12 per cent of boys and girls attending secondary school were smokers, most of these being at least 16 years old.

There were insufficient numbers of young people who smoked to compare daily cigarette intake according to age, but whereas Hill, White, Williams and Gardner (1993) found that boys who smoked at least one cigarette in the previous week were heavier smokers than their female counterparts, no significant gender differences in number of cigarettes smoked per day were apparent in the Institute's sample of self-defined smokers (on average, nine cigarettes for boys and ten for girls). However, our sample of smokers is small; a better indication of gender differences in daily cigarette intake will be possible when the data for residents in all the localities in the Australian Living Standards Study are available.

Of all young people who smoked, 46 per cent of boys and 31 per cent of girls smoked up to five cigarettes per day, while 23 per cent of boys and 13 per cent of girls said they smoked more than 15 cigarettes per day, with 6 to 7 per cent smoking more than 20 cigarettes per day.

Figure 1 shows the prevalence of smokers amongst Box Hill and Berwick young people (school students and school leavers combined) according to their age. Numbers in each of these groups defined by age, sex and residential location are quite small (21 to 46), and this should be kept in mind in interpreting these trends.

By age 14 - 15 years, smoking was reported by 20 per cent of Box Hill girls, and only 7 per cent of Box Hill boys, and 10 per cent of each sex in Berwick. By age 15 - 16, smoking was reported by 11 per cent of girls in Berwick, and 23 to 27 per of the other three groups. Finally, for those aged 18P19 years, most of whom had left school, smoking was reported by 33 per cent of girls and 21 per cent of boys in Box Hill, and by 29 per cent of girls and 19 per cent of boys in Berwick.

Thus, at age 14 - 15 years, Box Hill girls were more likely than other groups to be smokers. At age 16 - 17 years, Berwick girls were less likely than other groups to be smokers, but had nearly caught up with the Box Hill girls by age 18 - 19 years.

Table 1 suggests there is no link between adolescent smoking and equivalent household income, and no link between smoking and whether or not their own fathers (or sole mothers) had a tertiary degree, or were in paid work. Boys with blue collar family backgrounds were more likely than other boys to smoke, while girls whose fathers (or sole mothers) spoke a language other than English at home were less likely than others to smoke. Girls whose mothers smoked were more likely than other girls to smoke, while boys' smoking was unrelated to their parents' smoking.

It was expected that young people with a paid job would be more likely to smoke than others of much the same age, because those with work would be better able to meet the costs. Although Table 1 suggests that girls with a job were more likely to smoke than those without a job, such trends were not apparent when the effects of age were controlled. Of the school students who were at least 16 years old, around 22 per cent of those with and without a job smoked (owing to small number of cases, boys and girls are here combined). Of school leavers, around 34 per cent of girls and 18 to 23 per cent of boys with and without paid work smoked cigarettes.

Table 1 also suggests that compared with non-smokers, young people who smoke have lower self-esteem and tend to argue with parents about issues other than smoking more frequently than non-smokers. However, such trends do not take into account the fact that the smokers included a disproportionate number of older adolescents.

The measure of self-esteem used here represents average ratings of satisfaction with: 'the extent to which you are the kind of person you would like to be'; 'the respect or recognition you get'; 'what you are accomplishing in life'; and 'your personal, emotional life'. The rating scales ranged from 1 'extremely dissatisfied' to 9 'extremely satisfied', with the mid-point (5) being 'mixed feelings'. Since most people tend to view themselves positively (therefore recording high scores), scores close to the mid-point indicate problems with self-esteem. In fact, such relatively low scores may be taken as a symptom of depression. (This is supported by trends for parents who completed a measure of depression in addition to self-esteem.)

Young people rated how frequently (never, rarely, sometimes, often, very often) they argued with their parents about a range of issues, such as general behaviour, parents' not knowing their whereabouts, their clothes and appearance, and smoking (Weston and Millward 1992). The measure used in Table 1 is based on each person's average rating on all issues apart from smoking. Arguments were not particularly frequent, however. The average of all ratings for smokers was between 'rarely' and 'sometimes', while that for non-smokers was between 'never' and 'rarely'. (For smokers, average ratings concerning arguments about smoking were 2.40 for boys and 2.68 for girls.)

The Overall Picture

Multiple regression analysis was used to develop the simplest model predicting smoking behaviour in young people, based on factors presented in Table 1. The final model contained nine predictors: the young person's age, the young person's residential location, frequency of arguments with parents about issues other than smoking, self-esteem, whether mother only or two parents live in the home, whether or not a parent in their household smoked, whether of not the father or sole mother had paid work, whether or not either held a blue collar job (most recently, if unemployed), and whether either father or sole mother spoke a language other than English at home.

This set of factors better predicted whether or not girls smoked than whether or not boys smoked (explaining 21 per cent and 15 per cent of the variances respectively). Age was a significant predictor of smoking for both sexes. For boys, however, the most significant predictor was parents' blue collar work status, while frequency of arguments with parents was also significant. Thus, boys from blue-collar backgrounds, older boys and those who argued relatively frequently with their parents were particularly likely to smoke.

The tendency for male smokers to experience relatively low self-esteem was no longer significant when the other factors in the model linked with self-esteem were controlled. There were two such factors significantly related to self- esteem for boys and girls: self-esteem tended to be relatively low in those whose father or sole mother did not have paid work and in those who reported relatively frequent arguments with parents. Any causal connection between low self-esteem and arguments may be bi-directional, for low self-esteem may underlie behaviours of which parents disapprove, while parental criticism may dampen self- esteem.

Girls who smoked tended to be older than girls who did not smoke, to live in Box Hill rather than Berwick, and to have low self-esteem. Other trends for girls which were significant when considered alone as shown in Table 1 (arguments with parents, language spoken at home, and having a parent in the household who smoked) only approached significance (p<.10) when these other factors were controlled.

While the multiple regression method assumes that smoking results from these factors, causal relationships may be reciprocal or in a direction opposite from that outlined here. For example, smoking may act as a trigger for other arguments with parents, and may represent one of many conflict-provoking behaviours reflecting a young person's determination to achieve independence, challenge parental authority, and/or impress peers who are important to them.

The greater likelihood for girls with relatively low self- esteem to smoke is consistent with results of a 1992 survey of Victorian school children in Years 7, 9 and 11, which suggested that for each year level, girls with depressive symptoms were two to three times more likely than other girls to smoke (Centre for Adolescent Health 1993). For boys, a significant relationship between depressive symptoms and smoking was apparent at the Year 9 level only. As noted in this report, some young people who are depressed may smoke in an attempt to feel better, or they may join peer groups in which smoking is common. Box Hill students were more likely than Berwick students to attend private schools. Thus one possible explanation for the greater tendency for Box Hill rather than Berwick girls to smoke is that smoking is more common among girls attending private schools. However, this explanation was not supported. In total, smoking was reported by around 17 per cent of boys and girls in government schools, no more than 5 per cent of students in Catholic schools, and 10 per cent in other private schools. So reasons for Box Hill girls starting to smoke at an earlier age remains to be explained.

Implications

Adolescence marks a period when issues of responsibility and autonomy are negotiated with parents, when values of the adult world, particularly parents, are challenged, and when self-esteem is closely associated with issues of peer acceptance, body image and the like. Smoking is one means of challenging values concerning healthy lifestyles, but at the same time experimentation with cigarettes often represents the beginnings of a habit which is not only dangerous to health in the long term but also very difficult to give up. If prevention of smoking can be achieved during this period, then it is likely that smoking will never be taken up. How, then, can such prevention be achieved?

As Chesterfield-Evans (1988) notes, conservative methods of health promotion presuppose rational behaviour and a healthy self-esteem, while cigarette companies have exploited sexuality and insecurity in their efforts to get young people to take up cigarettes. While their efforts have been eroded somewhat by the ban on advertising first on television and later in the sponsorship of sport, manufacturers can remain confident that future addicts are in the making.

Smoking is most likely to begin when people are too young to be persuaded not to smoke by information about the delayed serious effects of smoking on health. Indeed, taking up smoking can be seen as brave, risky, exciting, and 'adult' - and particularly attractive to those with low self- esteem.

As well as highlighting the variety of fatal illnesses linked with smoking, one means of reducing cigarette uptake has been to use methods similar to those adopted by cigarette advertisers, but making healthy lifestyles appear positive and exciting and encouraging young people to see non-smokers as positive role models and smokers as negative role models.

Whatever anti-smoking approaches are adopted they need to take into account and capitalise upon the special needs and concerns marking the period of adolescence. The results of this analysis suggest that some approaches would not necessarily focus on smoking, but rather on helping families manage conflict more constructively, and helping vulnerable young people to overcome depression or low self- esteem and associated problems such as poor interpersonal skills. Indeed, it makes sense to consider effective family and adolescent mental health programs as tools in the prevention of smoking and other substance abuse.

References

• Australian Bureau of Statistics (1991), 1989-90 National Health Survey Users' Guide, Catalogue No. 4363.0, Canberra.
• Australian Institute of Health and Welfare (1992), Australia's Health 1992: The Third Biennial Report of the Australian Institute of Health and Welfare, AGPS, Canberra.
• Brownlee, H. and McDonald, P. (1993), 'In search of poverty and affluence: an investigation of families living in two Melbourne municipalities', Working Paper No.10, Australian Institute of Family Studies, Melbourne.
• Centre for Adolescent Health (1993), 'Adolescent health Survey 1992: Depression', Information Note, No.4, March.
• Chesterfield-Evans, A. (1988), 'Towards a new model in adolescent health promotion', in D. Bennett and M. Williams (eds) New Universals: Adolescent Health in a Time of Change, Brolga Press for the Australian Association for Adolescent Health, pp.82-84.
• Hill, D.J., White, V.W. and Gray, N.J. (1991), 'Australian patterns of tobacco smoking in 1989', Medical Journal of Australia, Vol. 154, pp.797-801 .
• Hill, D.J., White, V.M., Pain, M.D. and Gardner, G.J. (1990), 'Tobacco and alcohol use among Australian secondary schoolchildren in 1987', The Medical Journal of Australia, Vol.152, pp.124-130.
• Hill, D.J. White, V.M., Williams, R.M. and Gardner, G.J. (1993), 'Tobacco and alcohol use among secondary school students in 1990', The Medical Journal of Australia, Vol.158, pp.228-234.
• Pierce, J.P. (1990), 'Time to ban cigarette advertising and continue the "Quit. For Life"', The Medical Journal of Australia, Vol.152, p.113.
• Weston, R. (1993), Smoking across the generations, Unpublished paper, Australian Living Standards Study, Australian Institute of Family Studies, Melbourne.
• Weston, R. and Millward, C. (1992), 'Adolescent children and their parents', Family Matters, No.33, December, pp. 36-39.