This study addressed the methodological quality of longitudinal research examining R. Karasek and T. Theorell's (1990) demand-control-(support) model and reviewed the results of the best of this research. Five criteria for evaluating methodological quality were used: type of design, length of time lags, quality of measures, method of analysis, and nonresponse analysis. These criteria were applied to 45 longitudinal studies, of which 19 (42%) obtained acceptable scores on all criteria. These high-quality studies provided only modest support for the hypothesis that especially the combination of high demands and low control results in high job strain. However, good evidence was found for lagged causal effects of work characteristics, especially for self-reported health or well-being outcomes.