Validity of smokers' information about present and past cigarette brands--implications for studies of the effects of falling tar yields of cigarettes on health.

Four hundred and twenty nine current smokers and ex-smokers who had provided details 12 years previously completed a self administered questionnaire about their present and past smoking habits, and two weeks later current smokers supplied an empty cigarette packet. The tar group and brand name of the current cigarette given on the questionnaire were compared with details on the packet, and the brand alleged to have been smoked 12 years ago was compared with that actually recorded at that time. Only 55% of "low middle" tar cigarettes as indicated by returned packets had been correctly identified in the questionnaire. The brand name was the same in the questionnaire and on the packet in 74% of cases. The recalled brand was confirmed by past records in only 49% of cases. It is concluded that current smokers should be asked to return an empty packet or packets of the cigarette brand or brands usually smoked with a self administered questionnaire and that follow up studies of populations for which brands of cigarette smoked were previously recorded might be more valid than studies relying on recall.