Objectives: This article examines factors associated with starting or sustaining physical activity during leisure time.
Data source: The data are from the longitudinal household components of the National Population Health Survey. The sample consisted of 11,026 respondents who were aged 20 and older in 1994/95.
Analytical techniques: Multiple logistic regression was used to identify variables that independently predicted the adoption or maintenance of leisure-time physical activity.
Main results: Among people who had been inactive in 1994/95, the two-year incidence of starting at least moderate physical activity by 1996/97 or 1998/99 was 24 cases per 100 person-years. For people who had been at least moderately active in 1994/95, the two-year incidence of ceasing to maintain that level of activity was 32 cases per 100 person-years. Many predictors of starting or sustaining activity were the same: sex, age, educational attainment, smoking, and sense of mastery. However, some factors were significant for one sex only. For instance, overweight and the presence of children were deterrents for women, but not for men. Social involvement and smoking status were significant for men, but not for women.