Micro panel data on households, individuals, and firms are highly likely to exhibit measurement error. Duncan and Hill (J Labor Econ 3:508–532, 1985) found serious measurement error in average hourly earnings in the Panel Study of Income Dynamics (PSID). This got worse for a 2-year recall as compared to a 1-year recall. Bound et al. (1990) use two validation datasets to study the extent of measurement error in labor market variables. The first dataset is the Panel Study of Income Dynamics Validation Study (PSIDVS), which uses a two-wave panel survey taken in 1983 and 1987 from a single large manufacturing company. The second dataset matches panel data on earnings from the 1977 and 1978 waves of the US Current Population Survey (CPS) to Social Security earnings records for those same individuals.