THIS IS THE FIRST OF A SERIES OF REPORTS ASSESSING CRIMINAL SENTENCING IN ORDER TO ISOLATE THOSE FACETS OF THE PROCESS THAT ARE IN NEED OF REFORM. SEVERAL LIMITATIONS WERE FOUND IN EARLIER STUDIES. FIRST, MANY FOCUSED ON A SINGLE OFFENSE OR ON SIMILAR OFFENSES. SECOND, A LARGE NUMBER FAILED TO EXPLORE SYSTEMATICALLY BEYOND THE ZERO- OR FIRST-ORDER CORRELATION THE SIGNIFICANCE OF THE RELATIONSHIP BETWEEN A PARTICULAR OFFENDER CHARACTERISTIC (E.G., SEX, RACE, PRIOR RECORD) AND SENTENCE. A THIRD CONSTRAINT WAS THE RELATIVELY SMALL SCOPE OF THE DATA BASES USED BY EARLIER STUDIES. THE FOURTH LIMITATION WAS A FAILURE TO DIFFERENTIATE BETWEEN THE TYPE OF SENTENCE (I.E., INCARCERATION OR SOME OTHER SENTENCE) AND THE LENGTH OF INCARCERATION. A RESEARCH DESIGN IS PROPOSED, INCORPORATING THE FOLLOWING FEATURES IN ORDER TO OVERCOME THESE LIMITATIONS: (1) THE STUDY WILL EXAMINE SEVERAL CRIMES SIGNIFICANTLY DIFFERENT IN TYPE, INCLUDING A VIOLENT OFFENSE, PROPERTY OFFENSES, WHITE COLLAR OFFENSES, DRUG OFFENSES, AND SELECTIVE SERVICE VIOLATIONS; (2) IT WILL INCORPORATE A NUMBER OF CONTROLS SUFFICIENT FOR AN EXAMINATION OF CORRELATIONS BEYOND THE ZERO- OR FIRST-ORDER LEVEL; (3) ITS DATA BASE, A FEDERAL INFORMATION SOURCE, INCLUDES A LARGE NUMBER OF CASES, CAN ACCOMODATE REGIONAL ANALYSIS, AND INCLUDES MORE THAN 1 FISCAL YEAR; AND (4) WHILE MOST STUDIES HAVE VIEWED SENTENCING AS INVOLVING A SINGLE DECISION, THIS STUDY WILL DISTINGUISH BETWEEN THE JUDGE'S DECISION ABOUT THE LENGTH OF INCARCERATION AND THE VITAL DETERMINATION OF WHETHER TO INCARCERATE THE OFFENDER AT ALL. DETAILS ARE PROVIDED ON STUDIES OF SENTENCING TO DATE, A PROPOSAL FOR A NEW RESEARCH PERSPECTIVE, AND THE RESEARCH AND ANALYTICAL DESIGN. APPENDED MATERIAL INCLUDE DEFINITIONS OF INDEPENDENT VARIABLES AND FEDERAL CRIMINAL SENTENCING PROVISIONS. TABULAR AND GRAPHIC DATA ARE PROVIDED. SEE ALSO NCJ 33684-33686. (AUTHOR ABSTRACT MODIFIED--KBL)