DATA FROM THE STATE'S MAJOR COURTS AND FROM THE CONNECTICUT BOARD OF PAROLE AND THE CONNECTICUT DEPARTMENT OF ADULT PROBATION FOR FEBRUARY AND MARCH OF 1975 FORM THE BASIS OF THIS ANALYSIS. THE PAROLE ANALYSIS USED 21 VARIABLES RELATED TO THE OFFENDER'S PERSONAL CHARACTERISTICS, CRIMINAL HISTORY, AND PRISON BEHAVIOR. THE JUDICIAL ANALYSIS USED 14 PERSONAL AND CASE RELATED VARIABLES. TABLES PRESENT THE STUDY DATA. IT WAS FOUND THAT THE PAROLE BOARD ADHERES TO ITS STATED GOALS WHEN CONSIDERING A CASE. WORK AND EDUCATIONAL RELEASE PARTICIPATION, POSITIVE JOB PLANS, ACCEPTANCE OF DRUG TREATMENT, AND PARTICIPATION IN JOB TRAINING ALL POSITIVELY AFFECT PAROLE DECISIONS. PREVIOUS PAROLE REVOCATIONS STRONGLY AFFECT DENIAL. VARIABLES RELATED TO THE OFFENSE AND PAST CRIMINAL HISTORY SEEM UNIMPORTANT. THE JUDICIAL DECISIONS, ON THE OTHER HAND, SHOW WIDE VARIATIONS, SOME OF WHICH ARE DIFFICULT TO EXPLAIN. UNDERSTANDABLY, THOSE DETAINED PRIOR TO SENTENCING ARE MORE LIKELY TO RECEIVE PRISON SENTENCES THAN THOSE WHO ARE RELEASED ON BAIL. PROBATION OFFICER RECOMMENDATION FOR A SUSPENDED SENTENCE HAS CONSIDERABLE IMPACT. REPRESENTATION BY A PUBLIC DEFENDER OR A PRIVATE ATTORNEY IS NOT SIGNIFICANT. DEFENDANTS WHO REFUSE TO PLEA BARGAIN GENERALLY RECEIVE MORE SEVERE SENTENCES. STRANGELY, DEFENDANTS WITH THE GREATEST HISTORY OF VIOLENCE RECEIVE THE LIGHTER SENTENCES. IT IS CONCLUDED THAT THIS STUDY POINTS UP THE NEED FOR SENTENCING GUIDELINES TO CONTROL JUDICIAL DISPARITY. IT SEEMS TO SUPPORT THOSE WHO MAINTAIN THAT EXTRANEOUS FACTORS CARRY UNDUE WEIGHT IN SENTENCING DECISIONS. FOOTNOTES CONTAIN EXTENSIVE ADDITIONAL DATA. (GLR)