We estimated the incidence of erectile dysfunction (ED) in a population-based sample during 5-y follow-up and determined how the rate was affected by sociodemographic and life-style factors. The target population comprised all men aged 50, 60 or 70 y residing in the city of Tampere or 11 surrounding municipalities in Finland at the start of follow-up. A questionnaire was mailed to 3143 men in 1994 and to 2864 in 1999. The follow-up sample consisted of 1442 men who responded to both baseline and follow-up questionnaires. We estimated the effect of sociodemographic and life-style factors on the incidence of ED among the 1130 men free of ED at baseline. We found no differences in the incidence of ED by the level of education, marital status, urban/rural place of residence, amount of alcohol and coffee consumption. Obesity (rate ratio (RR)=1.7, 95% confidence interval (CI): 1.1-2.5) and current smoking (RR=1.5, 95% CI: 0.9-2.2) increased the incidence of ED. Current smokers free of comorbidity were also at higher risk of ED (RR=1.3, 95% CI: 0.8-2.1), but no effect was observed among past smokers. Our results indicate that sociodemographic and life-style factors, except age and obesity, have little influence on ED.