METHODS: A nationwide data set was examined for this secondary data analysis. The dependent variable was the degree of risk, which was measured based on the number of high-risk behaviours in which adolescents participated. Age, gender, ethnicity, self-rated academic performance, family size, parental marital status and parental academic attainment were included as independent variables. Analyses stratified by educational level were conducted. Odds ratios (ORs) were calculated using ordered logit.
RESULTS: The most common high-risk behaviour among Malaysian adolescents was physical inactivity (35.97%), followed by smoking (13.27%) and alcohol consumption (4.45%). The majority of adolescents had low risks (52.93%), while only a small proportion had high risks (6.08%). Older age was associated with increased odds of having high risks (OR: 1.26). Male adolescents had higher odds of being in a high-risk category compared to female adolescents (OR: 1.28). Compared to Malays, Chinese adolescents had higher odds of being in a high-risk category (OR: 1.71), whereas Indian adolescents had lower odds (OR: 0.65). Excellent academic performance was associated with reduced odds of participating in high-risk behaviours (OR: 0.41).
CONCLUSION: Personal factors are important determinants of high-risk behaviours. This study provides a better understanding of those adolescent groups that are at greater risk.
PRACTICAL IMPLICATIONS: An intervention directed towards reducing participation in high-risk behaviours among adolescents who have both poor academic performance and less-educated parents may yield promising outcomes.
METHODS: Cigarette and alcohol use was assessed in a large cross-sectional national sample aged 50 years and above from the Irish Longitudinal Study on Ageing (TILDA) (n = 6,576). The Brief Ageing Perceptions Questionnaire (BAPQ) assessed individual's views of their own ageing across five domains. Study hypothesis that stronger beliefs on each of the BAPQ domains would be related to drinking and smoking was examined using multinomial logit models (MNLM). Regression parameter estimates for all variables were estimated relative risk ratios (RRR).
RESULTS: More women were non-drinkers (30 % vs. 20 %) and men displayed significantly higher alcohol use patterns. One in five older Irish adults was a current smoker (16.8 % of women, 17 % of men), and smoking and harmful drinking were strongly associated (P