The determination of risk assessment is an important technique of the "new penology" in its search for more efficient and effective ways to handle offenders. The current study examines the validity of a risk assessment instrument and its ability to predict recidivism among a sample of 133 male and female juvenile offenders in one of Iowa's eight judicial districts. The study instrument was a risk assessment tool utilized by juvenile court officers at the dispositional stage of juvenile court case processing. The results indicate that the study instrument is statistically valid and provides a 34.2 percent improvement over chance in the prediction of recidivism. This finding contradicts previous research that cautioned against using an instrument that was validated at one stage of juvenile court case processing at another stage. In addition, it was found that leaving the variable "sex" in the risk assessment model and making adjustments in the risk categories for females, resulted in higher mean cost rating scores than when treating females the same as males or when scoring males as if they were females;Results from a path analysis indicated that the number of re-referrals to juvenile court was predicted directly by the completion of sanctions and the length of non-recidivism, but there was no direct association found between the risk assessment score and the number of re-referrals to juvenile court. There was also no relationship found between the juvenile court ordered disposition and the number of re-referrals. The reduced models for males and females showed that there were both similarities and differences by sex in the prediction of recidivism. The primary difference by sex that was found was that the juvenile court officer's recommendations and the juvenile court ordered disposition had no association with the number of re-referrals for females, while an indirect association was found to exist for males. In other words, the intrusiveness of the disposition had some impact on male offenders in terms of re-referral, while other factors such as the completion of sanctions and the length of non-recidivism predicted the number of re-referrals for females.