A PERSONALIST APPROACH TO ESTIMATING POPULATIONS OF INSTITU-TIONS

WHEN A BUSINESSMAN OR OTHER DECISION MAKER NEEDS AN ESTIMATE(SAY, OF SALES OF A NEW INDUSTRIAL PRODUCT) AND HE HAS AC- CESS TO RICH AND VARIEDINFORMATION BUT LIMITED RESOURCES HE NORMALLY MAKES HEAVY USE OF JUDGEMENT IN HIS ENQUIRY. PERSONALIST(BAYESIAN) STATISTICS PROVIDES A LOGICAL FRAME- WORK TO GUIDE THE USE OF SUCH JUDGMENT IN CONJUNCTION WITH ANY SAMPLE OR OTHER DATA. IN PARTICULAR, IT PRODUCES ESTIMATES IN THE FORM OF PERSONAL PROBABILITIES CONVENIENT FOR ANALYZING DECISIONS. DECISION SCIENCE CONSORTIUM INC WILL DEVELOP A PERSONALIST APPROACH TO INSTU TIONAL ESTIMATION; PARTICULARLY FOR INDUSTRIAL MARKET RE- SEARCH WHERE JUDGMENT IS CRUCIAL AND CURRENT PRACTICE BASED EITHER ON CLASSICAL STATISTICS OR UNAIDED INTUITION, LEADS TO COSTLY ERRORS. POTENTIAL APPLICATIONS TO ACADEMIC SEARCH AND GOVERNMENT REPORTING ON INSTI TITION WILL ALSO BE CONSIDERED). PHASE I WILL EXPLORE HOW THREE PERSONALIST (BUT UNORTHODOX)CONCEPTS CAN ENHANCE TOH COST-EFFECTIVENESS OF EACH PHASE OFINSTITUTIONAL ESTIMATION, FROM THE PLANNING PHASE (INCLUDINGBUDGET ALLOCATION AND DESIGNING THE MIX OF APPROACHES SUCH SAMPLING, DIRECT ASSESSMENT, AND DESK RESEARCH), TO THE DATACOLLECTION PHASE I INCLUDINIG SAMPLING FRAMES, STRUCTURE SELECTION, AND MEASUREMENT), AND FINALLY TO THE ANALYSI PAHSE (DEVELOPING A FINAL ESTIMATE OR ESTIMATES). DECOMPOSEDERROR ANALYSIS ANALYZES THE ESTIMATION ERROR INTO ITS SEPARBLE SOURCES: RANDOM FLUCTUATIONS, MEAS UREMENT, FRAME BIAS, ETC. (BROWN, 1969). ANALYSIS OF PLURAL RESEARCH DEALS WITH DEVELOPMENT AND INTE-OF SEVERAL ALTERNATIVE ESTIMATES (BROWN & LINDLEY, 1982). PARTIAL CONDITIONING OF CONSEQUENCES ASSESSES THE DECISIONALVALUE OF PROPOSED RESEARCH (BROWN, 1978). PHASE II WILL DEVELOP OPERATIONAL TECHNIQUES AND COMPLIMENTARY DATA RESOURCES, AND TEST THEM ON LIVE MARKET RESEARCH PROBLEMS.