THE CRIMINAL JUSTICE SYSTEM IS COMPRISED OF A SERIES OF INTERRELATED DECISIONS, STARTING WITH THE DECISION OF A VICTIM OR WITNESS TO REPORT AN OFFENSE THROUGH THE DECISION OF A PAROLE BOARD TO RELEASE A CONVICTED OFFENDER FROM PRISON. DISCRETION AVAILABLE TO CRIMINAL JUSTICE DECISIONMAKERS IS CONSIDERABLE, AND THERE IS ALWAYS A RISK THAT TWO IDENTICAL INCIDENTS WILL BE TREATED DIFFERENTLY. DECISION MODELS HAVE BEEN DEVISED TO FACILITATE THE PROCESSING OF LARGE CASE VOLUMES AT A VARIETY OF DECISION POINTS. THESE MODELS HAVE THE CAPABILITY, IF PROPERLY CONSTRUCTED AND USED, TO ALLOCATE RESOURCES MORE EQUITABLY AND MAKE THE CRIMINAL JUSTICE SYSTEM MORE EFFICIENT. INFORMATION ELEMENTS IN THE SRI BURGLARY DECISION MODEL INCLUDE THE ESTIMATED TIME OF OCCURRENCE OF OFFENSE, WITNESS REPORTS OF OFFENSE, ON-VIEW REPORTS OF OFFENSE, USABLE FINGERPRINTS, SUSPECT INFORMATION DEVELOPED (DESCRIPTION OR NAME), AND VEHICLE DESCRIPTION. INFORMATION ELEMENTS ARE WEIGHTED TO REPRESENT THE VALUE OF AN ELEMENT IN MAKING AN ARREST FOR BURGLARY. IN A FIELD TEST OF THE SRI MODEL, A MANUAL GIVING INSTRUCTIONS FOR CONDUCTING THE REPLICATION WAS WRITTEN IN ORDER TO REDUCE PROCEDURAL VARIANCE. THE MANUAL EXPLAINED HOW TO SAMPLE BURGLARY CASES FROM POLICE DEPARTMENT FILES, PREPARE CASES PRIOR TO DATA COLLECTION, CODE RELEVANT DATA ON FORMS, AND TABULATE RESULTS. IN ADDITION, THE MANUAL DISTINGUISHED BETWEEN DEPARTMENTS THAT FORMALLY SCREENED BURGLARY CASES AND THOSE WITH NO FORMAL SCREENING POLICIES. THE SELECTION OF SITE POLICE DEPARTMENTS WAS LIMITED TO DEPARTMENTS WITH CHIEF EXECUTIVES WHO WERE MEMBERS OF THE POLICE EXECUTIVE RESEARCH FORUM. A TARGET SAMPLE SIZE OF 500 BURGLARY CASES WAS ESTABLISHED FOR EVERY PARTICIPATING LAW ENFORCEMENT AGENCY. FIFTEEN POLICE DEPARTMENTS THAT PARTICIPATED IN THE FIELD REPLICATION DID NOT USE A FORMAL SCREENING PROCEDURE DURING 1977. THE REMAINING 12 POLICE DEPARTMENTS HAD SOME TYPE OF FORMAL SCREENING. ANALYSIS SHOWED THAT THE SRI MODEL, ON THE AVERAGE, WAS RELATIVELY ACCURATE IN PREDICTING THE OUTCOME OF BURGLARY CASE INVESTIGATIONS BUT DID NOT DEMONSTRATE WHETHER THE MODEL WAS BETTER THAN NONSTATISTICAL CASE SCREENING PROCEDURES. IN GENERAL, STATISTICALLY WEIGHTED FELONY CASE SCREENING MODELS ARE USEFUL FOR SCREENING CASES PRIOR TO THEIR ASSIGNMENT TO DETECTIVES. IT IS RECOMMENDED, HOWEVER, THAT THE SRI MODEL NOT BE USED BY LAW ENFORCEMENT AGENCIES EXACTLY AS CONSTRUCTED WITHOUT FIRST TESTING IT. NOTES ARE INCLUDED. (DEP)