A mail survey of physical activity habits as related to measured physical fitness

The associations of self-reported measures of physical activity from a mail survey with an objective measure of physical fitness were investigated. Respondents to a health status survey (n = 12,225), conducted in 1982 by the Institute for Aerobics Research in Dallas, TX, formed the population. From this group, males who also had a clinical examination within 60 days of the return of their questionnaire served as subjects. The study subjects (n = 375, mean age = 47.1 years) completed a maximal physical fitness assessment using a modified Balke protocol. The questionnaire included a section of inquiries concerning leisure time physical activity participation in which subjects were asked to quantitatively recall exercise participation for varying periods of time. Reported exercise participation values were converted to estimates of energy expenditure and combined into overall indices of physical activity participation. Multiple regression analyses were used to determine the individual contributions of the physical activity indices in predicting maximal treadmill performance (physical fitness). Significant predictors of physical fitness were age (beta = -0.34), an index of running, walking, and jogging participation (beta = 0.31), and the response to a question on frequency of sweating (beta = 0.35). The multiple correlation coefficient for these variables in predicting physical fitness was 0.65. These results indicate that exercise behavior can be accurately estimated in large populations by using simple questions in a mail survey.