The probability of technical success and R&D appraisal

Abstract

Two models that assess the probability of technical success are presented. The first model describes projects that are either complete successes or complete failures. The second model describes projects that may have partial success. The first model uses an expert evaluator's assessments of the probability of success of selected projects to parametrize a logit probability distribution. The results give implicit weights which can be employed in future analyses to calculate the expected level of success of new projects. The second model allows projects to be partially successful. Several methods of using an expert evaluator's opinion to determine the parameters of a beta distribution are developed.