Factors affecting escalation in auction games
Auction games have been relatively neglected by empirical researchers in spite of their apparent relevance to a wide range of social problems. Following a discussion of social dilemmas in general, and an exhaustive review of research into psychological traps and auction games in particular, six new experiments on factors affecting escalation in auction games are reported. In the first experiment, male and female subjects participated in three successive auction games with variable prize values. A serial decrease in escalation with experience was observed among males only; no significant effect of prize value was found. Experiment two compared subjects' responses to programmed strategies in two-person and three-person groups, either face-to-face or via a computer simulation designed to provide greater control and to facilitate economical data-collection. A significant interaction emerged: escalation was significantly greater in the computer simulation than the face-to-face condition in the two-person groups only. The successful computer simulation was retained for all subsequent experiments, and experiment three examined the relationship between escalation and personality. A significant positive relationship was found between radicalism and escalation. In the fourth experiment no significant differences were found between individual and team bidding (in teams of two or three), but in the fifth experiment a non-significant trend towards greater escalation in individual than in three-person team bidding was observed. The sex difference, and the interaction reported by previous researchers between sex and individual versus team bidding, were non-significant, although there was a trend towards greater escalation by individuals than teams. Experiment six compared individual bidding in two-person, three-person, and five-person groups. Escalation was found to be significantly greater in the three-person and five-person groups than in the two-person groups. Methodological difficulties and problems of interpretation in all six experiments are discussed and areas in need of further investigation are suggested.