ALL 200 EVALUATIONS IN THE SAMPLE TO BE ANALYZED BY THE PROJECT HAVE BEEN READ BY MEMBERS OF THE PROJECT TEAM. A LACK OF CRITICAL THINKING ABOUT THE PHYSICAL ELEMENTS OF PROCESS, PARTICULARLY IN THE LOGISTICALLY ORIENTED STUDIES, HAS BEEN GENERALLY FOUND. BECAUSE OF THESE DEFICIENCIES, THE STAFF IS DIRECTING A MAJOR EFFORT TOWARD MODEL-BASED EVALUATION. ONE CLASS OF MODEL USEFUL IN POLICE EXPERIMENTS IS THE HYPERCUBE MODEL, WHICH DEPICTS ANALYTICALLY THE BEHAVIOR OF A SPATIALLY DISTRIBUTED EMERGENCY SYSTEM OPERATING UNDER ALTERNATIVE DEPLOYMENT OPTIONS. AN ABSTRACT OF A RECENT STUDY BY PROJECT STAFF ON THIS MODEL IS INCLUDED. THE PROJECT PRODUCED TWO PUBLICATIONS DURING THIS QUARTER: 'ANALYSIS OF A CONTINGENT EXPERIMENTAL DESIGN, A BEFORE-AND-AFTER EXPERIMENT WITH A BASELINE PERIOD OF RANDOM DURATION' AND 'THE CONGESTED MEDIAN PROBLEM.' THE FIRST STUDY MENTIONED IS A FIRST EXAMPLE OF A MAJOR DEVELOPING THEME OF THE PROJECT--THE NEED FOR FLEXIBILITY, ADAPTIBILITY, AND FEEDBACK IN THE CONDUCT OF EVALUATIONS. A NEED IS SEEN FOR A THEORY OF CONTINGENCY DESIGN OF EXPERIMENTS. SUCH DESIGNS WOULD ALLOW FOR ADAPTIVE UPDATING OF INFORMATION REGARDING PARAMETER ESTIMATES, ADAPTIVE TIMING OF VARIOUS EVALUATION ACTIVITIES, INCLUSION OF CERTAIN UNINTENDED CONSEQUENCES EXPERIENCED IN THE EVALUATION, AND INCORPORATION OF SUBJECTIVE AS WELL AS OBJECTIVE INFORMATION. THE SET OF COMPUTER PROGRAMS FOR USE IN STORING DATA AND DATA ANALYSIS FROM THE EXAMINATION OF THE 200 EVALUATIONS HAS BEEN COMPLETED. A LISTING OF THE 200 EVALUATIONS COMPRISING THE SAMPLE IS GIVEN IN THE APPENDIX, AND THE CODE LIST FOR COMPUTER STORAGE IS PROVIDED. ILLUSTRATIVE RESPONSES OF THE COMPUTER ARE ALSO SHOWN IN THE APPENDIX. FOR REPORTS 1 AND 2, SEE NCJ 52597-52598; SEE 55104 FOR THE FOURTH REPORT. (RCB)