HOW REPRESENTATIVE ARE MATCHED CROSS-SECTIONS - EVIDENCE FROM THE CURRENT POPULATION SURVEY

In this paper we use the Current Population Survey (CPS) to illustrate some problems that arise in analyzing longitudinal data constructed b y matching person records across files of rotating cross-section surve ys. After studying the determinants of attrition from the CPS, we focu s on two issues. The first is the effect of attrition on estimates of several labor market outcomes, such as participation rates, wages, and transitions rates between labor force states. The second is possible uses of the matched CPS to understand the nature of the measurement er ror process.