Revisiting the Effect of Criminal Justice Involvement on Employment using the NLSY

• In Event: Stigma and Labor Market Disconnection: Criminal-Record-Based Risk Control, Its Effects on Employment Opportunity, and Efforts to Overcome Related Workforce Disconnection

Abstract

The negative relationship between criminality and involvement with the criminal justice system and employment and wages has been long-studied in particular in the 1979 and 1997 National Longitudinal Surveys of Youth. While many of the extant studies using the NLSY are observational, at least one uses quasi-experimental methods as an attempt to estimate a causal relationship of criminality and employment and/or wages. However, though prior research has recognized that there is attrition bias in the NLSY, methods used to deal with it have unnecessarily removed large numbers of observations. Alternative methods to attempt to correct for this attrition bias are necessary, especially as the NLSY cohorts mature. Furthermore, even in the quasi-experimental designs, little attention has been paid to the sample design or the sampling weights. Using the NLSY79 and 97, this paper will present estimates of the causal relationship of criminality and criminal justice system involvement using quasi-experimental econometric techniques, using the complex sampling design and sampling weights and exploring statistical adjustments to correct for attrition bias.

Author

• Ryan Kling, Abt Associates Inc.