Abstract

Decision Models For The Selection Of Research Projects * t t Robert H. Cramer and Barnard E. Smith INTRODUCTION Research management is by nature a process of risk taking. Each project offers an unknown and highly variable payoff for an often un­ certain expenditure. This is in contrast to plant investment where expenditures may be accurately estimated and payoffs, while not cer­ tain, are at least determined within a reasonably narrow range. Due to this difference, the techniques for decision making which have been developed for plant expenditures offer little attraction to the research manager. Such techniques usually assume that revenues and expenses are precisely determined. If an estimate must be viewed as having a distribution, it is assumed adequate to use the expected value of this distribution as the cost certain. While this assumption seems harmless enough, it ignores a significant aspect of the decision pro­ cess. Implicit in this assumption is that the decision maker values every dollar as being equally important. A research manager cannot accept this assumption. Today's re­ search is characterized by massive expenditures with the hope of massive payoffs. Only at the lowest level of research activity are the expenses A more detailed