NCJ Number
67082
Date Published
1979
Length
103 pages
Annotation
THIS IS A REPORT OF THE POLICE EXECUTIVE RESEARCH FORUM'S REPLICATION IN 26 MEMBER DEPARTMENTS OF A MODEL FOR SCREENING BURGLARY CASES BEFORE ASSIGNING THEM FOR INVESTIGATION IN ORDER TO SELECT THE MOST SOLVABLE.
Abstract
FORUM DEPARTMENTS AND RESEARCHERS SOUGHT TO DETERMINE THE PERFORMANCE OF THE STANFORD RESEARCH INSTITUTE'S (SRI) INFORMATION MODEL BY TESTING THE MODEL'S PREDICTED CASE OUTCOMES WITH ACTUAL BURGLARY CASE INVESTIGATION RESULTS. IN DEVELOPING THE SRI MODEL, WHICH IS CRIME-SPECIFIC AND STATISTICALLY DERIVED, RESEARCHERS ISOLATED 25 INFORMATION ELEMENTS BY USING DISCRIMINANT ANALYSIS AND FINDING 6 ELEMENTS TO BE SIGNIFICANTLY RELATED TO THE ARREST OF A SUBJECT. THE STUDY FOUND THAT THE SRI DECISION MODEL PREDICTED THE RESULTS OF BURGLARY INVESTIGATIONS ACCURATELY IN 85 PERCENT OF THE 12,001 CASES EXAMINED, AND THAT IT WAS POSSIBLE TO REPLICATE RESEARCH IN MANY POLICE DEPARTMENTS WITH THE ACTIVE PARTICIPATION OF DEPARTMENT STAFFS RATHER THAN OUTSIDE EVALUATORS. IN ADDITION, DATA FROM NINE AGENCIES WITH FORMAL SCREENING PROCEDURES REVEALED THAT A STATISTICALLY DERIVED MODEL IS MORE ACCURATE THAN A FORMAL CASE SCREENING PROCESS BASED ON INTUITION AND EXPERIENCE, SHOWS CONSISTENT RESULTS, AND ASSIGNS MANY FEWER CASES THAN NONSTATISTICAL SCREENING PROCEDURES AND SAVED MORE INVESTIGATIVE RESOURCES. CHAPTERS OF THE REPORT DESCRIBE DATA COLLECTION PROCEDURES FOR THE STUDY, ANALYZE TWO EVALUATIONS, EXPLAIN HOW AGENCIES CAN DEVELOP AND USE THE SRI MODEL, AND DISCUSS FUTURE RESEARCH IMPLICATIONS. APPENDIXES INCLUDE FREQUENCY OF INFORMATION ELEMENTS AND CALCULATING THE LOST CLEARANCE RATE IN SCREENING AGENCIES. TABULAR DATA ARE GIVEN. (AUTHOR ABSTRACT MODIFIED--AOP)