UPON REQUEST FROM THE COURTS, CALIFORNIA DEPARTMENT OF CORRECTIONS (CDC) EVALUATES CONVICTED FELONS AND RECOMMENDS SENTENCES. THE RESEARCH INVESTIGATED SENTENCING EVALUATIONS CONDUCTED AT CALIFORNIA INSTITUTION FOR WOMEN (CIW) BETWEEN JUNE 1972 AND MAY 1975. RANDOM SAMPLES (200 CASES EACH) WERE SELECTED FROM THE FOLLOWING COMMITMENT TYPES: (1) PRESENTENCE EVALUATIONS CASES; (2) PROVISIONAL COMMITMENT EVALUATION CASES; AND (3) CASES SENTENCED DIRECTLY TO CIW WITHOUT EVALUATION (STRAIGHT COMMITMENTS). PRESENTENCE EVALUATION CASES ARE SENT TO CDC FOR A CERTAIN PERIOD AND ARE THEN RETURNED TO COURT FOR FINAL SENTENCING. PROVISIONAL COMMITMENT CASES ARE SENTENCED TO STATE PRISON, BUT THE JUDGE REQUESTS A SENTENCING RECOMMENDATION FROM CDC AND RESERVES AUTHORITY TO WITHDRAW THE PRISON COMMITMENT AND TO RESENTENCE THE OFFENDER. CASES WERE CODED ON 20 PRECOMMITMENT VARIABLES, SUCH AS SEVERITY OF OFFENSE, WEAPON INVOLVED, PRIOR ARRESTS, AGE, OCCUPATION, AND DEGREE OF COUNTY URBANIZATION. IT WAS FOUND THAT HIGHLY URBANIZED COUNTIES WERE MORE LIKELY TO USE PRESENTENCE EVALUATIONS; SEMIURBAN COUNTIES, PROVISIONAL COMMITMENT; AND RURAL COUNTIES, STRAIGHT COMMITMENTS. ANALYSIS SHOWED THAT THE EVALUATION PROGRAM ITSELF HAD A GREATER IMPACT ON THE SEVERITY OF CIW'S RECOMMENDATIONS THAN ANY OTHER VARIABLE. WOMEN GIVEN PROVISIONAL COMMITMENT EVALUATIONS WERE MORE THAN TWICE AS LIKELY AS WOMEN IN THE PRESENTENCE EVALUAION PROGRAM TO BE RECOMMENDED FOR PRISON SENTENCES. SENTENCES RECOMMENDED FOR WHITE WOMEN WERE SIGNIFICANTLY MORE SEVERE THAN THOSE RECOMMENDED FOR BLACK OR MEXICAN-AMERICAN WOMEN. ANALYSIS OF SENTENCE IMPOSED AFTER EVALUATION SHOWED THAT RECOMMENDATIONS FROM INDIVIDUAL EVALUATION STAFF MEMBERS AND THE COLLECTIVE CIW STAFF RECOMMENDATIONS EACH HAD AN INFLUENCE ON FINAL SENTENCE, BEYOND THAT ACCOUNTED FOR BY OFFENSE, PRIOR RECORD, AND BIOGRAPHICAL VARIABLES. GENERALIZATIONS REGARDING POSTRELEASE CRIMINAL INVOLVEMENT WERE TENTATIVE. FIVE PRECOMMITMENT VARIABLES WERE FOUND TO BE SUFFICIENT TO PREDICT POSTRELEASE OUTCOME. THESE INLCUDED OFFENSE SEVERITY AND FOUR MEASURES OF PRIOR RECORD. NONE OF THE CIW RECOMMENDATIONS HAD ANY INFLUENCE UPON POSTRELEASE OUTCOME. APPENDIXES DETAIL THE VARIABLES USED AND DESCRIBED RECODING, TRANSFORMATION, AND COMPUTATION OF THE VARIABLES. REFERENCES ARE PROVIDED. (AUTHOR ABSTRACT MODIFIED--VDA)