Holt, 1993; Roth, 1995). Psychologists became interested in psychological factors such as individual differences (Kelley and Stahelski, 1970; Messick and McClintock, 1968), the effects on behavior of changing the payoffs (Kelley and Grezlak, 1972), and the effects of communication (Dawes, et al., 1977).
More generally, throughout the 1970s and 1980s, psychologists examined factors that influence cooperation across the range of social dilemmas, including commons dilemmas, prisoners’ dilemmas, and public goods tasks (for a broader review of social dilemmas in the social psychological research, see Dawes, 1980; Komorita and Parks, 1994; Messick and Brewer, 1983). Much of the early work on prisoners’ dilemmas was criticized on the grounds that it was atheoretical and that it had little to say about extra-laboratory affairs (Pruitt and Kimmel, 1977).
One interesting theme that has emerged from the more recent research we reviewed is the extent to which people are, or are not, other-regarding (how, if at all, people take others’ welfare into account). The nature in which they are, or they become, other-regarding has become a central research question. Although the hypothesis that people have preferences for the welfare of others is at least as old as Adam Smith’s Theory of Moral Sentiments,1 psychologists have found this question pivotal for understanding choice behavior in interdependent situations. Early efforts in the latter half of the 20th century were made by Sawyer (1966), who tried to measure altruism, by Conrath and Deci (1969), who were estimating a “bivariate” utility function, and by Messick and McClintock (1968), who used a type of random utility model to assess social motives for allocating distributive outcomes in situations of social interdependence. In the Messick and McClintock model, each preference (maximize own outcome in absolute terms, maximize own outcome in relative terms, and maximize joint outcomes of both self and other) had sizable nonzero probabilities. In the 1970s researchers in economics (e.g., Scott, 1972) and in the behavioral sciences (e.g., MacCrimmon and Messick, 1976) began to explore preference structures that could produce behavior that appeared to be altruistic, selfish, and competitive at the same time.
In the 1980s, Messick and Sentis (1985) introduced the concept of a “social utility function” that was later expanded by Lowenstein et al., (1989). A social utility function posits additive preferences for one’s own outcomes and preferences for the difference between one’s outcome and that of others. Both studies found that the latter function takes its maximum when payoffs to self and other are equal, supporting the assumption made by Falk et al. (this volume:Chapter 5). Their economic model further generalizes the social utility component to comparisons with more than one other person.
This chapter focuses on experimental work published in major peer-reviewed journals in psychology. In passing, we note experimental work in economics that bears on variables of interests to psychologists. We included studies that manipu-