Year: 1995 Source: American Journal of Epidemiology, v.141, no.9, (1995), p.885 SIEC No: 19950175

This Letter to the Editor notes in perfecting the authors’ methodology for weighted analysis of longitudinal data, they realized that the weighting used in earlier papers (ref: SIEC AN 910900) may be less accurate & appropriate than their current approach. Although when used in prevalence estimates, the earlier weights are entirely appropriate & correct, those same weights when used in logistic regression analysis may not precisely define the reference population. Details are provided.