| Popis: |
The main advantage of panel analysis is that changes can be investigated at the individual level. However, the observed individual changes result not only from underlying, true changes but also from random fluctuations. The problem is that these random fluctuations may lead to systematically looking patterns of change that seemingly beg for substantive, theoretical explanations. Notorious in this respect is the regression to the mean phenomenon that has misled investigators everywhere and in all times. This phenomenon is more aptly called regression to the mode and in this latter form also affects the observed changes in categorical panel data. Latent class models are excellently suited to find out about the regression to the mode and prevent misleading explanations of panel data. |
