correctly on a test of general knowledge. Two other equally random, but less prototypical ways were used to putatively assign the first position for the other participants. One quarter were told they had gotten the most answers correct on an achievement test, but they had seen that one of the six tests was much easier than the other five. The lucky person would get first place, not the person who knew the most. As a test this was unfair, but as a random device it was fair because tests were assigned randomly (subjects were told). In any case, it was not a prototypical process. Neither was the fourth mechanism, which involved calculating the distance of a participant’s birthday from a randomly selected day of the year. Although participants rated this process as fair, they also rated it as unprototypical.
The study results showed that participants given the two prototypical justifications for their privileged position took nearly 50 percent more of the shared resource than those given the less prototypical justifications. Moreover, the importance of the justification depended on the details of the decision problem. When overuse resulted in zero payoffs for everyone, the effect of the justification was nonexistent; when people were allowed to keep whatever they had taken, the participants with prototypical justifications took nearly twice as much as those with unusual justifications.
Causal attributions are also important with regard to scarcity or abundance of the resource pool. Why there is a lot or a little has been shown to make a difference in how people treat the resource. In a field study of water use during the 1976-77 drought in California, Talarowski (1982) found that people who stayed within their water allocation limits tended to believe the drought was caused by a natural shortage. Those who exceeded their allocation, however, expressed the view that the shortage was people-induced. In this type of study, it is impossible to say whether the beliefs cause the behavior or the behavior causes the beliefs, or whether both are being caused by some other factor.
Rutte et al. (1987) tried to provide an experimental answer to this question. In their study, participants were told that they would be the fifth person of a six-person group to harvest from a shared pool. All subjects saw the harvests of the previous four (bogus) group members. Collectively, these first four members took 20 points (Dutch guilders—the experiment was conducted in the Netherlands). Half of the subjects were told that the pool initially contained 35 points (leaving 15 for the last two members to share) and half were told that it contained 25 (leaving just 5 for the last two members to share). Half of the people in these two conditions were told that all group members knew the size of the pool from the beginning, and the other half were told that the first four were ignorant of the pool size. When everyone knew the pool size, the shortage or abundance would be attributable to the others, whereas it would be attributable to luck when the first four did not know.
When all group members knew the pool size, the behavior of the first four