Title:
Contextualizing smoking: the influence of household factors on smoking habits

Abstract:
Objectives Traditional approaches studying smoking focused primarily on individuals. Nowadays there is a shift from individual-level models towards models that incorporate contextual effects. This research project takes place in this trend. The purpose of this specific analysis is to investigate the influence of household on smoking practices, in order to determine if the household is a relevant unit to study smoking. Is there evidence for an household effect on smoking? In other words, do household factors (both observable and unobservable) affect individual probabilities of smoking, all other individual characteristics being equal? Is the household effect sensitive to the national context? Materials and methods Data from the European Community Household Panel (ECHP) were used. All individuals aged more than 16 years are interviewed within the selected households. The probability to be a daily smoker is explained by variables at the individual level and at the household level. We take into account a possible specific effect of the household by implementing a random effects probit model. Results This study brings to light a phenomenon of clustering of smoking habits within households, which is modulated by the composition of the household. Preliminary results seem to show that the longer the sitory of tobacco consumption in the country, the more sensitive the concordance. Conclusions This research gives evidence for clustering of smoking practices in household, the effect being different according to the context of the household and the country. Research about health concordance overwhelmingly suggests evidence for clustering of health status and health behavior. This offers room for a deeper understanding for the causes of health concordance. Next step would be to determine the reason for concordance (is it due to ex-ante correlation or ex-post convergence?).