[Longitudinal research in aging populations: participation and the quality of data collected with questionnaires]

In this article methods are discussed which may keep participation in a longitudinal study among elderly persons high, for example adaptation of the interviews and proxy interviews. The LASA sample is described from the start in 1992 until now. The non-response is evaluated and we found that refusals are particularly important in the first part of the longitudinal traject. Also data quality is studied in relation to the aging of the respondents. Although there are theoretical reasons to expect that aging may impair data quality, no support for this hypothesis was found in the present study. Data quality was stable during a period of six years. But data quality seemed poorer for those respondents who dropped out from the study. Item non-response and duration of the interview were higher for drop-outs.