Discusses problems of bias in longitudinal studies resulting from low recruitment and high attrition rates. Participation rates in longitudinal studies where the dependent variable is antisocial child behavior or related areas are reviewed. Participation and recruiting experiences are described from the 1st 3 yrs of a 5-yr study of delinquency in 4th-grade males, using a risk-sampling procedure. The strategies used to attain high recruitment and retention rates were high pay for the Ss, trained home visitors paid on an incentive system, and written materials targeted to the sample. These more than doubled the recruiting rate obtained from a pilot study. No families have dropped out; the sample is not biased toward less aggressive Ss; and the Ss seem to be at risk for psychopathological behaviors.