Payoff-based learning best explains the rate of decline in cooperation across 237 public-goods games

Abstract

What motivates human behaviour in social dilemmas? The results of public goods games are commonly interpreted as showing that humans are altruistically motivated to benefit others. However, there is a competing 'confused learners' hypothesis: that individuals start the game either uncertain or mistaken (confused), and then learn from experience how to improve their payoff (payoff-based learning). We: (1) show that these competing hypotheses can be differentiated by how they predict contributions should decline over time; and (2) use meta-data from 237 published public-goods games to test between these competing hypotheses. We find, as predicted by the confused learners hypothesis, that contributions declined faster when individuals have more influence over their own payoffs. This prediction arises because more influence leads to a greater correlation between contributions and payoffs, facilitating learning. Our results suggest that humans, in general, are not altruistically motivated to benefit others, but instead learn to help themselves. Humans often face opportunities to improve group welfare but at an individual cost ('social dilemmas') 1,2 . For example, an individual may act to pay more taxes, practice social distancing during a pandemic, and/or reduce their carbon footprint 3,4 . Human behavior in such situations is often studied experimentally with the public-goods game 5-7 . In the linear public-goods game, individuals can contribute financially to a group fund, which multiplies all contributions by M. The total product is then shared out equally between the N group members, providing each individual a return of M/N per unit contributed, termed the marginal per capita return (MPCR). Consequently, whenever the multiplier is smaller than the group size (M<N), the group does best if everyone contributes fully (cooperates), but individuals maximize their financial gain by not contributing. The results from hundreds of linear public goods experiments have consistently shown two patterns: (1) behaviour varies, but initial contributions average just below half (40-50%); and(2) average contributions decline as individuals repeat the game 8 . One potential explanation for these patterns is that most individuals are altruistically motivated to contribute, but are also averse to unfair outcomes ('inequity aversion') 9-15 . This preference for fair outcomes causes them to limit their altruism to match, or just undercut, what they expect others will contribute, leading to intermediate contributions ('conditional cooperation') [16][17][18] . Then, as these conditional cooperators encounter non-cooperators, who do not contribute as much, they resent the greater payoff those non-cooperators receive 10,12,14 . This causes the conditional cooperators to revise their overly optimistic expectations about their group mates, and reduce their level of contribution. The inequity aversion interpretation is based on the assumption that we can "safely assume that the players understood the game" 9,17 . However, an alternative ex...

Social Learning Strategies and Cooperative Behaviour: Evidence of Payoff Bias, but Not Prestige or Conformity, in a Social Dilemma Game

“…1a. Which social learning strategies, if any, do participants use?-We predict that payoff biased learning will have the strongest influence on cooperative behaviour [45] followed by prestige and then conformity [38,42,43]. 2a.…”

“…Further experimental evidence suggests that participants are more likely to exhibit a payoff bias than conformity in a cooperation game and reduce their contributions [42,43] and also decrease their contributions when reminded how their behaviour was benefiting others [44]. A recent analysis of 237 PGGs also showed that declines in contributions were most consistent with improving personal payoffs [45]. Furthermore, cooperation was also higher when participants had no information on the behaviour of their group mates [46].…”

“…What effect does access to social information have on cooperative behaviour? We predict lower cooperation with access to social information than when individuals make decisions asocially because we expect a payoff bias to decrease cooperation [45,46].…”

A preference to learn from successful rather than common behaviours in human social dilemmas

“…We wanted multiple rounds of 'model' data so that we could observe more learning and short-term cultural evolution. However, contributions normally decline in public goods games [32]. If individuals observe declining levels of cooperation, it is hard to test if individuals are learning from success or trying to match a declining group average [36,50].…”

“…In the public goods game, individuals can make monetary contributions to a group fund. Financially speaking, the group does best if everyone contributes fully, but individuals do best by contributing nothing, hence the social dilemma [28][29][30][31][32]. Many studies have shown that individuals tend to contribute partially and that most will condition their cooperation to local levels (either perfectly or imperfectly) [2,[33][34][35][36][37][38][39][40][41][42].…”