Measurement Error in the SIPP: Evidence from Matched Administrative Records: Working Paper 2007-03

Working Paper
January 30, 2007

Julian Cristia and Jonathan A. Schwabish

Validation studies that compare survey-reported earnings to administrative-recorded earnings are useful to assess the extent and implications of measurement error in labor market data. While previous work typically used small restrictive samples with topcoded earnings, this paper uses data from the 1996 Survey of Income and Program Participation (SIPP) Panel matched to Social Security administrative records. This large representative sample contains uncapped administrative earnings and allows us to provide more definitive evidence on measurement error. Results show that SIPP respondents, on average, underreport earnings by a significant amount ($2,800). Consistent with previous studies we find that measurement error is negatively correlated with true earnings. Finally, unlike previous work, we find a consistent pattern between measurement error and covariates: factors positively associated with earnings are negatively correlated with measurement error.