Reliability and physiologic correlates of the Harvard alumni activity survey in a general population

The reliability of the Harvard Alumni Activity Survey (HAAS) and its association with physiologic measures was assessed in a large sample of men and women aged 25–65 years residing in the Boston metropolitan area in 1987. Reliability was estimated by comparing HAAS energy expenditure reports (kcal/week) from two separate interviews conducted 7–12 weeks apart. The test-retest reliability coefficient was 0.58 for the entire sample, but was considerably higher ( r = 0.69) for those whose activity patterns had not changed from one interview to the next. HAAS self-reports were compared to two physiologic measures known to be affected by physical activity: high density lipoprotein cholesterol (HDLC) and body mass index (BMI). The natural logarithm of weekly HAAS expenditures was positively correlated with HDLC ( r = 0.14, p < 0.01) and negatively correlated with BMI ( r = 0.13, p < 0.01) for all respondents. These statistically significant associations persisted when adjusted for other covariates influencing physiologic status. The reliability coefficients and physiologic correlations for the HAAS in this sample are comparable to those reported for physical activity instruments requiring more intensive data collection and scoring procedures.