THE ROLE OF PROJECT COMPLETION INFORMATION IN RESOURCE ALLOCATION DECISIONS.

Prior studies have suggested that costs, the amount of money already invested in a project, influence resource allocation decisions but have often confounded sunk costs with the degree to which a project is completed. To address this issue, we varied information about both sunk costs and project completion in two experiments. Our results suggest that degree of project completion may dominate any sunk cost effects that are present in resource allocation decisions. An expanding literature from a variety of theoretical paradigms suggests that, contrary to normative models of decision making, consideration of sunk costs-the amount of a budget invested in a project-may enter into resource allocation decisions. In the reported study, we focused on situations in which a decision maker faces a choice between continued investment in a project or withdrawal with certain loss of some or all prior investments. According to traditional economic theory, a rational decision maker in the situation just described would consider only incremental costs and benefits. However, a growing body of research suggests that sunk costs may influence the decision to continue investment in an ongoing project. For example, Arkes and Blumer (1985) found that subjects faced with a decision to invest more funds in an ongoing research and development project were more willing to make the investment than those asked if they would invest the same amount of funds to start up the same R&D project. In a replication study using five levels of prior investment, Garland (1990) found a linear relationship between the amount of sunk costs and willingness to continue investment in the project. Faced with mounting evidence for the existence of irrational sunk cost effects on resource allocation decisions, we examined the studies in which such effects have been observed. Such an examination revealed that at least two variables are in need of closer scrutiny. First, a number of studies have used decision scenarios in which sunk costs have been confounded with the degree of project completion. Second, most studies have included negative information about project success as a constant condition.