Tobacco use among adolescents in Scotland: profile and trends

6. PREDICTORS OF SMOKING

Logistic regression is used to predict an outcome using several predictor variables. In this case, it was used to predict factors which influence the likelihood of a 15 year old child being a regular smoker. The potential factors included in each of the logistic regressions were those analysed earlier in this report in Chapters 4 and 5. The factors included are detailed in Figure 6.1.

Figure 6.1 - Factors included in each logistic regression

The model identified several factors that were associated with regular smoking. Figure 6.2 summarises these results.^[40]

Despite clear evidence of a link between deprivation and smoking, SIMD was not a significant predictor of smoking, while free school meal entitlement and how well off the child thought their family is were only weak predictors. However, as with gender, many of these factors are interdependent, such as what the child expects to do after school and if they have been excluded. Thus, these results do not necessarily mean that deprivation is not an important factor in smoking, but that in terms of these results, other factors are more strongly associated with being a regular smoker. As before, it is important to bear in mind that these are potentially complex associations with, rather than direct causes, of smoking.

Figure 6.2 - Key drivers of smoking among 15 year olds