Measurement and latent class typologies of smoking patterns in college students

Abstract

Recently, there has been a reported decrease in smoking in many populations, but the prevalence of cigarette smoking in college samples is still high. Before effective prevention and intervention programs can be developed to address this issue, there was a need for more qualitative and quantitative research on what types of patterns of smoking are present on campus. Therefore, the goal of this present study was to investigate different types of smokers on campus, in particular different types of occasional smokers, using both survey and focus groups methods. The survey was completed by 335 smokers (M=18.58 years; 63.9% female; 85.7% White/Caucasian) from Colorado State University. The survey included measures of smoker identity, age of initiation, current smoking behavior, reasons for smoking, desire to quit, addiction level, and alcohol use. To test whether reasons for smoking could produce different classes of smokers in the college population, a LCA was conducted using the modified reasons for smoking scale (n=327). The results revealed that a four class model (Logliklihood= -2362.594; BIC=5136.275; AIC=4867.188; LRT= p < .05) was a good fit. The classes included addicted smokers (23.55%), non-endorsing smokers (18.04%), stress smokers (33.03%), and social smokers (25.38%). An advanced LCA with covariates was used to examine whether these classes differed on specific characteristics (n=303). Overall, the results revealed age of initiation, current smoking patterns, smoker type, and smoking cessation predicted class membership while current age and alcohol use did not. To augment these findings with qualitative data, 41 individuals taken from the larger sample participated in focus groups based on their current smoking patterns (i.e., social smoker, regular smoker, and occasional smoker groups). The focus groups indicated that there were light, regular, heavy, and nondaily/occasional smokers on campus which included, stress, social, and "drunk smokers". These findings as well as the findings from the survey support the notion that there are different types of nondaily smokers with distinct smoking patterns in the college population. More specifically, both data sources revealed that stress smokers and social smokers were occasional smoker typologies that emerged as distinct classes and differed on key predictors. Two relatively new typologies also emerged: non-endorsing smokers and drunk smokers. More research is needed to make further conclusions about these groups of occasional smokers. The typologies revealed in the present study should be kept in mind when designing interventions for the college population. Additional implications and future directions are also discussed.