A Simulation of Multiple Selection Using Nominal Group Procedures

A common decision problem involves the selection of k alternatives from a larger but finite set. Examples are the selection of personnel for a training program or the funding of research proposals. A related problem is that of screening where a large but finite group of initial candidates is reduced prior to final selection. Choices of this type frequently are made on judgmental grounds by group processes; because of the subjectivity it is difficult to develop objective criteria for choosing one group process over another. In this paper, the comparative effectiveness of three procedures for group opinion summarization: standardized scores, sum of ranks, and zero-one scoring, are studied by means of a simulated nominal group process. The analysis leads to several simple “rules of thumb” which should prove useful to the practicing manager.