Abstract
This paper aims to classify changes in sports and recreational activities from 1982 to 2001 and to investigate future trends in the numbers of participants of various types of activities based on age-period-cohort analyses. The effects of age, period, and cohort on participation rates are estimated from data classified by age group and survey period using the Bayesian logit cohort model. A total of 50 activities for both sexes are classified into four clusters forage effects, five for period effects, and six for cohort effects using cluster analysis. The mainresults are : (1) From the clusters for age effects, people increase or keep their participationin some activities, while they disengage themselves from participation in many activities asthey age. (2) From the clusters for period effects, participation in some activities such as golfare affected by the time trends related to household income, economic cycle, and fashion ofcomics. (3) From the clusters for cohort effects, each cohort tends to continue participation inparticular activities depending on the historical circumstances where its members grew up.