ACCURACY OF THE COLLEGE ALUMNUS PHYSICAL-ACTIVITY QUESTIONNAIRE

The validity and reproducibility of the Physical Activity Index from t he College Alumnus Questionnaire (PAI-CAQ) were determined in 78 men a nd women (21-59 yr) with a broad range of physical activity habits. Th e PAI-CAQ was computed as the sum of energy expended in stair climbing , walking, and sports and recreational physical activity recalled from the past week. Data were validated against measures of cardiorespirat ory fitness, body fatness, motion detection, and physical activity rec ords. All physical activity was recorded by participants for six 48-ho ur periods. Each day of the week, including weekends, were represented at least twice in the physical activity records. Age-adjusted correla tion coefficients between like activities on the College Alumnus Quest ionnaire and physical activity records ranged from 0.25 to 0.65 in men and 0.28 to 0.86 in women. Correlations between the PAI-CAQ and valid ation criteria that reflected total and heavy-intensity physical activ ities were higher (r = 0.34-0.69, p < 0.05) than for lighter-intensity physical activities and motion detection expressed in MET-min.d-1 and kcal.d-1 (r < 0.35, p > 0.05). Test-retest reproducibility was higher over 1 month (r = 0.72) than over 8 and 9 months (r = 0.34 and 0.43) (p < 0.05). Energy expended in walking and stair climbing was underest imated on the College Alumnus Questionnaire, resulting in lower PAI-CA Q scores, as compared to the physical activity records.