Data were obtained from the New York State Computerized Criminal History/Offender-Based Transaction Statistics data base. Of the 11,098 probation-eligible defendants, 53 percent belonged to a minority group (40 percent black, 13 percent Hispanic). Analysis by region indicates that the disproportionality in incarceration rates was low in New York City and high in upstate and suburban regions. In the latter, white defendants predominated and were less likely than minorities to be incarcerated. A binary logit analysis technique was used to control for the effects of prior record, type of crime (personal, property, drug), seriousness (Class A, B, C), and race. Results of these analyses indicate that different variables have different weightings in determining whether or not a defendant is incarcerated in different regions of the State. While prior record was the most influential variable in all three regions, its impact was greatest in New York City. The magnitude of the race effect, adjusting statistically for the other factors, was negligible in New York City but appreciable in upstate. While race had less weight than prior record, it was more influential than offender age. The statistical model developed by the analyses can be used to determine the probabilities of incarceration in the three regions for hypothetical cases varying in the pertinent variables. A simulation indicates that similarly situated hypothetical defendants face systematically different risks of incarceration depending on where the case is decided and whether the defendant is black of white. Seven notes, tabular data, and 16 references are included.