Content area
Full text
Previous studies have shown the continuation of a failing project, also known as escalation of commitment, occurs in many aspects of business and government. This study incorporates several established theories to explain the effects of an alternative investment, magnitude of loss and monitoring on the likelihood of continuing a project. The combination of the presence of an alternative investment, "high" magnitude of loss and "low" monitoring was enough to cause decision makers to stop the project suggesting for the first time that decision makers may be willing to stop a project even though it is 90% complete.
(ProQuest: ... denotes formulae omitted.)
INTRODUCTION
Previous studies have shown the continuation of a failing project occurs in many aspects of business and government, and that the commitment to and continuation of a previous decision can even apply to waiting on a bus, attending a play and mountain climbing. The phenomenon of runaway projects is also referred to as overcommitment or escalation of commitment to a failing course of action (Staw, 1976), the sunk cost effect (Northcraft and Wolf, 1984) and entrapment (Brockner, Rubin, and Lang, 1981). Therefore, the reversal of escalating commitments to failing courses of action, either through project termination or redirection, can be called de-escalation of commitment (Keil and Robey, 1999).
Several theories have been suggested to explain the reasons managers continue failing or doubtful projects. Among those theories are Agency Theory (Jensen and Meckling, 1976), Self-Justification Theory (Festinger, 1957), Prospect Theory (Kahneman and Tversky, 1979; Tversky and Kahneman, 1981), Approach Avoidance Theory (Rubin and Brockner, 1975), Self-Efficacy Theory (Bandura, 1977a) and National Culture Theory (Hofstede, 1980, 1983,1984).
This study incorporates Agency Theory, Self-Justification Theory and Approach Avoidance Theory to explain the effects of an alternative investment, magnitude of loss and monitoring on the likelihood of continuing a project. The experimental design of the study was a 2 (presence of an alternative investment: yes or no) x 2 (monitoring: low or high) x 3 (magnitude of loss: low, medium or high) between-subjects factorial design. Likelihood of continuing a project was measured in two ways: first, dichotomously (either "yes" the subjects continued the project, or "no" they did not) and second, on a 0-100 continuous scale. Data were analyzed using binary logistic regression for...