Actors and smoking susceptibility, p. 19
The third hypothesis examined the premise that, controlling for traditional
demographic and risk factors, younger non-smoking adolescents who regularly saw
actors smoking would be more susceptible to becoming smokers. Table 3 further
summarizes the analyses comparing younger (ages 14 – 15) and older (ages 16 –18) sub-
samples. The model for the younger sub-sample was significant [X
2
(10, N = 2,507) =
76.15, p < .001]; the model for the older sub-sample was not statistically significant.
Among the younger adolescents, the patterns in the multiple regression model
matched the those in the total sample. Once again, younger never-smoker adolescents
who regularly saw TV and movie actors smoking were 2.26 (p < .05) times more likely
to be susceptible to smoking. When controlling for the other predictor variables, there
were no significant relationships among the ages 16 – 18 sub-sample. However, once
again, the 95% confidence intervals for the younger and older respondents overlap
appreciably.
The final hypothesis proposed an interaction such that females and younger
adolescents who regularly viewed actors smoking would be more susceptible to smoking.
Table 3 summarizes the results for the multiple logistic regression analyses testing this
final hypothesis. The models largely repeat the total sample results in Table 2 with
notable changes -- three interactions between respondents’ sex, age, and exposure to
actors smoking were introduced. Specifically, after statistically adjusting for the other
predictor variables, (Model 1) females who regularly saw TV and movie actors smoke
were 2.47 (p < .001) times more likely to be susceptible to smoking. Model 2 indicates
that younger adolescents (14 – 15) who regularly saw actors smoking on television and in
the movies were 2.26 (p < .001) times more likely to be susceptible to smoking. Finally,
Convention