In spite of the overwhelming amount of literature on the "regression artefact" many analysts of panel data still tend to ignore the consequences of unreliability for the measurement of change. This is illustrated by 2 recent Dutch studies. The regression artefact is explained by confronting the substantive interpretations of the results of these studies with an interpretation in terms of unreliability of measurement. In the first panel study among boys & girls from high schools, questions were asked about their plans for the future. Among other things the figures indicated that those who intended to go to a U had relatively more stable plans than others. Assuming a very slight degree of unreliability & using the fact that the majority planned to go to a U, these results could be explained without assuming that there was any real difference in turnover between the several categories. The other study attempts to determine, among other things, why students fail an exam for the 1st time & pass it a 2nd time. Indications are that among those who failed the 1st time, students with the worst results improved the most. Instead of the substantive explanations previously offered, an interpretation in terms of regression artefacts is made plausible. 6 Tables. AA.
