Eliciting Expert Opinion in Acquisition Cost and Schedule Estimating

Abstract

Despite the emphasis on data and analytics in acquisition cost and schedule estimating, many estimating situations still require eliciting expert opinion from a subject matter expert. This is problematic, as a 2007 RAND report concludes that there is no standard model for seeking expert input for acquisition estimates. Per the report, the DoD's elicitation methodologies are largely ad hoc, in that they are seldom based on or derived from references to the elicitation literature (Galway, 2007). In this paper, a popular and commonly cited elicitation model "the Stanford Research Institute (SRI) elicitation model" is presented and adapted to the cost and schedule estimating process. It is posited that the consistent application of a formal model would reduce expert biases and improve the acquisition community's risk and uncertainty analyses. This paper also provides the results of an original meta-analysis of published experiments that examine expert elicitation for business and engineering problems. The data reveals that experts are overconfident and struggle to identify the true range of outcomes for both business and engineering problems. However, using a structured elicitation model, training the expert prior to the elicitation, and providing the expert with feedback are shown to decrease expert overconfidence.