Data were gathered for 1981-84 in the juvenile courts in Boston, Detroit, Newark, and Phoenix for a sample of 201 youths against whom prosecutors filed petitions for transfer. The analysis compared the characteristics of the 38 percent of the youths who were transferred to the criminal court with those of the 62 percent who were retained by the juvenile court. Findings showed that the best predictors of transfer for the total sample were older age at the time of the offense and the number of the victims in the offense. However, the four sites varied greatly in the determinants of the transfer decision. Results suggested the need for formal standards to promote equity and fairness in transfer decisions. Figures, tables, and 45 references.