It has become common knowledge that employment during adolescence is positively associated with a variety of delinquent and problem behaviors. What is not so common knowledge is the reason for this positive association. One possibility is that youth employment causes problem behavior, while a second (but not necessarily mutually exclusive) possibility is that the positive association is a spurious consequence of self-selection. Prior research in this tradition has not adequately dealt with the issue of selection effects, and thus runs the risk of mistaking a selection effect for a causal effect. This paper uses three waves of data from the National Longitudinal Survey of Youth 1997 to examine in greater detail the association between adolescent employment and problem behavior. Our analysis deals directly with self-selection by introducing controls for observed and unobserved heterogeneity through the use of a random effect probit model. This paper builds on prior research by estimating all models separately by race and sex, modeling two distinct types of employment (freelance and employer work), and accounting for seasonal variation in employment (school-year vs. summer work). Preliminary results indicate that school-year employment (particularly intensive school-year employment) is unrelated to problem behavior net of controls for observed and unobserved heterogeneity, contrary to findings from prior research. Additionally, both summer work and freelance work are associated with increases in problem behavior, but with differing sex-race interactions.
(Return to Program Resources)