Attributions of trust and the calculus of reciprocity

This research investigated trust and reciprocity in two experiments using the Trust Game. In the Trust Game, Player 1 can ‘‘trust’’ an unknown Player 2 by sending some portion of a monetary endowment. The amount sent triples on its way to Player 2, who can then ‘‘reciprocate’’ by returning as much as he or she wishes to Player 1. Initial endowments were either $10 or $20 and were known to recipients; amounts sent were experimentally manipulated and varied from $2 to the entire endowment. Although many trusted parties returned enough money to equalize outcomes, trustors only benefited, on average, when they sent all or almost all of their endowments. Results suggested that recipients viewed sending less than everything as a lack of trust and that felt obligations mediated choices to reciprocate. These and other results contrast markedly with traditional, incremental models of the trust process, which suggest that initial trustors should take small risks and build trust gradually. ! 2003 Elsevier Science (USA). All rights reserved. Trust and reciprocity are intimately related: reciprocation of an initially trusting act can instigate a beneficial cycle of increasing trust and reciprocation. Unreciprocated trust, however, can seriously damage a relationship and weaken its prospects. Traditional, incremental models of trust development (e.g., Rempel, Holmes, & Zanna, 1985) suggest that trust initiators should be careful, because trust involves risk. Thus, it is wise for trustors to take relatively small risks initially, increasing their risks as a relationship develops. Trusted parties, however, may view caution on the part of trustors skeptically, wondering why they have not trusted more (Weber, Malhotra, & Murnighan, 2002). Trusted parties may view small initial acts of trust negatively, or may not recognize them as ‘‘trusting acts’’ at all. Either reaction reduces the likelihood of reciprocity. If the likelihood of reciprocity depends on the clarity of a trusting act, then trust initiators have a difficult dilemma—how much must they risk to engender reciprocity? A !generalized norm of reciprocity" (Gouldner, 1960, p. 170) suggests that initial trusting acts will be reciprocated regardless of their size, because people understand that they should help (and not injure) those who have helped them. Yet egocentric biases (Ross & Sicoly, 1982), such as the tendency to devalue others" contributions (Messick & Sentis, 1979), will reduce the likelihood of reciprocity, particularly if cautious trustors take only small risks. In the experiments presented here, we investigated trusted parties" decisions to reciprocate by manipulating the amount of trust shown (and risks taken) by initial trustors in the Trust Game, first investigated by (Berg, Dickhaut, & McCabe, 1995).