Stable distribution of reciprocity motives in a population

Jeroen M Van Baar,Filippo Ricci,Felix H Klaassen,Alan G Sanfey,Luke J Chang

doi:10.1038/s41598-020-74818-y

Abstract

Evolutionary models show that human cooperation can arise through direct reciprocity relationships. However, it remains unclear which psychological mechanisms proximally motivate individuals to reciprocate. Recent evidence suggests that the psychological motives for choosing to reciprocate trust differ between individuals, which raises the question whether these differences have a stable distribution in a population or are rather an artifact of the experimental task. Here, we combine data from three independent trust game studies to find that the relative prevalence of different reciprocity motives is highly stable across participant samples. Furthermore, the distribution of motives is relatively unaffected by changes to the salient features of the experimental paradigm. Finally, the motive classification assigned by our computational modeling analysis corresponds to the participants’ own subjective experience of their psychological decision process, and no existing models of social preference can account for the observed individual differences in reciprocity motives. These findings support the view that reciprocal decision-making is not just regulated by individual differences in 'pro-social’ versus ‘pro-self’ tendencies, but also by trait-like differences across several alternative pro-social motives, whose distribution in a population is stable.

Highlights

Evolutionary models show that human cooperation can arise through direct reciprocity relationships
To formally test whether these differences reflected between-subject differences in reciprocity motives rather than noise around a single decision strategy shared across all participants, we compared the model fit of the three unitary social preference models to that of our Moral Strategy m odel[10]
The improved model fit for the Moral Strategy (MS) model relative to IA was not significant in the × 4/× 6/× 8 context (MS vs. IA: t(93) = −1.31, p = 0.194), mean model performance across the two contexts was significantly better for MS (paired-samples t-test on mean subject Akaike Information Criterion (AIC) across contexts, IA vs. MS: t(93) = 2.07, p = 0.042)

Summary

Introduction

Evolutionary models show that human cooperation can arise through direct reciprocity relationships. Recent evidence shows that different people are motivated by qualitatively different sets of these reciprocity motives, with some individuals exhibiting context-dependent motives such as moral o pportunism[10] This raises the question whether the distribution of reciprocity motives across people in a population is stable and predictable, or whether it may instead change with time and decision task. This is an important question because individual differences in motives for social decision-making—and beliefs about others’ motives—can determine whether an intervention to increase prosocial action backfires or s ucceeds[11,12,13,14,15]. This allows us to test whether individual differences in reciprocity motives are stably distributed in a population or are sensitive to the framing of a task, which speaks to the trait-like nature of reciprocity motives without testing the same individuals multiple times

Methods

Results

Conclusion