Smoking in France: an age-period-cohort analysis

Céline Goffette, Institut National d'Etudes Démographiques (INED)

Analysis of smoking trends have mainly focused on age and period effects. On one hand, it is now well known that tobacco use is a life-cycle phenomenon: smoking prevalence follows an inverse U-shape, increasing until the end of adolescence and then decreasing continuously. On the other hand, total smoking prevalence has reduced for decades, but men and women have known different evolutions: while male smoking has been declining for long, female smoking has been rising until recently in Europe and is still rising in southern European countries. The goal of this study is to add a cohort perspective to the analysis of smoking patterns, because tobacco use is undoubtedly a cohort phenomenon. Smoking initiation occurs during a fairly limited age range. It is presumably largely determined by the current attitudes towards tobacco (acceptability, advertising, bans, etc.) and by the prevalence which is observed at that time for young people’s significant others (either peers, parents, colleagues). But as tobacco is addictive, there is a dependence of smoking at older ages on smoking at younger ages. The objective is to assess the relative contribution of age, period and cohort effects on smoking trends in France. For this purpose, French Health Surveys from 1980-1981, 1991-1992 and 2002-2003 will be analysed. This individual-level cross-sectional data allows to study various cohorts.