The Aging in Manitoba (AIM) project includes nearly 9,000 seniors who have been interviewed. Three independent cross-sectional samples (i.e. 1971, 1976, 1983) were subsequently followed (1983–1984, 1990), producing cross-sectional and longitudinal data. A major goal of this paper was to examine non-response over time and representativeness of the panel of AIM survivors. This involved analysis of non-response, calculations of completion/response rates, and comparisons of demographic variables across the AIM sample and the broader Manitoba and Canadian populations. Non-response was very low (4.6% in 1983–1984, 5.0% in 1990) and did not appear to be dependent on such factors as location, gender, education level, or nationality descent. The 95.0 per cent completion rate compares favourably with other longitudinal studies, demonstrating the value of employing rigorous tracking procedures. These findings suggest that attempts to minimize bias and selective attrition in the AIM study have been successful.