Variation in Social Smoking Behavior Among U.S. Young Adults: Different Targets for Intervention?

Introduction: Young adulthood (ages 18-24) is a critical developmental transition and provides an enormous opportunity to alter trajectories of smoking behavior for a large public health impact. Of note, there is a high prevalence of “social smoking” in this age group. The goal of this study is to describe, through latent class analysis, groups of social smoking young adults (SSYAs) and other young adult smokers.

Methods: Baseline data from a nationally representative cohort of young adults were collected in July 2011 using an online survey. Latent class analysis was applied to a subsample of participants (n=854) who were aged 18-24 and reported ever smoking a cigarette at baseline.  Variables in the latent class model included all available measures of social smoking (i.e., self-identified smoking status, smoking alone vs. with others, confidence resisting smoking in social situations, alcohol use, and smoking at parties or social events, bars or restaurants).  Latent class models with two through eight classes were run to find the model that best fit the data.

Results:  The optimal model selected based on BIC, entropy, and the odds of correct classification identified five latent classes of young adult ever smokers: 1) social smokers (n = 83; 10%), 2) smokers (n=192; 22%), 3) non-smokers (n=452; 53%), 4) social smokers-more dependent (n=86; 10%), and 5) loner smokers (n=41; 5%). Class 1 (social smokers) was characterized by self-identifying as “social smokers,” smoking mainly or only with others, and smoking at parties or social events.  Class 2 (smokers) largely self-identified as “smokers,” reported smoking equally alone and with others, and smoked at parties or social events and bars or restaurants. Class 4 (social smokers – more dependent) largely identified as “social smokers," but smoked equally alone and with others. Class 5 (loner smokers) mainly identified as “smokers” and reported smoking alone.  All classes used alcohol, with the greatest proportion of alcohol users in Class 1 (social smokers).

Conclusions: This study highlights a unique class of social smokers that comprises 10% of all young adult ever smokers.  It also identifies a class of social smokers that may have escalated their smoking behavior from a purely social activity to smoking equally alone and with others. Future research is needed to examine the profiles and smoking behaviors of young adult smokers and to determine how to better target smoking prevention and cessation interventions to subgroups of young adult smokers.