Explaining differences in labour market transitions between panel and life-course data in West-Germany

摘要: Panel and life-course data are ideally suited to unravelling labour market dynamics, but their designs differ, with potential consequences for the estimated relationships. To gauge extent which these two produce dissimilar transition rates causation thereof, we use German Life History Study Socio-Economic Panel. Life-course in particular suffer from recall effects due memory bias causing understated probabilities. seam spurious transitions between statuses recalled activity calendars that generate heaps at time points cause overstated We combine datasets estimate multilevel (multistate) discrete-time models event history model states taking factors into account. Though find much lower study, confirming results of Solga (Qual Quant 35:291–309, 2001) this Journal East-Germany, part difference can be explained by short spells bias. The on exit, re-entry job mobility combined show indeed a negative retrospective design effect. Another specification includes length period shows no significant decrease probabilities increasing length, suggesting effect is other differences.