Project Project
-A +A

The Role of Group Competition and Cultural Group Composition in the Provision of Public Goods: An Evolutionary Approach

Research Fields
Migration, Integration, and Intercultural Conflicts
Project Management
Jeroen van den Bergh, PhD (Universidad de Barcelona)
Susanne Rebers, PhD (Vrije Universiteit Amsterdam); Dr. Julian Garcia (Universiteit van Amsterdam)
Dutch Science Foundation

Theoretical background and objectives

The explanation of the human capacity to cooperate in larger groups constitutes one of the major puzzles that preoccupy evolutionary biologists and anthropologists. A currently popular explana­tion, multilevel selection theory (e.g., Wilson 2002), proposes that the human capacity for collec­ti­ve action has evolved by way of a process of intergroup competition in which cooperative beha­viour is selected at the group level because more cooperative groups are more successful (in terms of population growth, outcomes of intergroup conflicts, and cultural imitation of successful groups by less successful ones) than less cooperative groups. Many theorists of cultural group selection argue that this evolutionary process should have produced a behavioural pattern of “paro­chialism” (Bowles and Gintis 2004), which combines ingroup favouritism with outgroup hosti­li­ty. The question of the evolutionary basis of human collective action is relevant for the study of immigration and integration because if correct, the theory of cultural group selection has several important implications for interethnic relations, for instance that intragroup cooperation will be more difficult in culturally heterogeneous communities (see project 6.2) or that intergroup conflict is more likely to occur when cultural differences between groups are stark (see projects in cluster 3), especially if groups live highly segregated lives (see projects in cluster 4). This pro­ject inquires into what biologists (Tinbergen 1963) refer to as 'ultimate', i.e. evolutio­na­­­ry, causes of intra- and between-group behaviour. It does so along two empirical paths of in­quiry. The first, implemented by collaborators Garcia and van den Bergh, mathematically models group selection processes and conducts numerical simulations under varying parameters. The second component of the project, in which the research unit is involved, tests the behavioural implications of cultural group selection theories in a series of public goods experiments.

Research design, data and methodology

A series of pilot experiments was conducted to establish whether different results were obtained in online experiments using a web-based tool, or classical laboratory experiments. Because these re­vealed no significant behavioural differences and online experiments were easier and less cost­ly to implement, we stuck to online experiments except in the second experiment, where groups repeatedly interacted, which was too impractical to implement online. In a first experiment we tested whether more homogeneous ingroups were more successful in overcoming public good provision problems, by giving subjects information on several cultural characteristics of their fellow group members (religion and political affiliation), as well as on a trivial trait as a control (birth month). The experiment included a further control condition in which all traits were ran­dom­ly assigned. Second, we conducted single-group repeated public goods experiments to in­vestigate whether more homogeneous groups were more successful in resisting the trend of de­creasing cooperation levels that is routinely observed in repeated games. In a third set of expe­ri­ments, we tested whether between-group competition for a valued good increased within-group cooperation, and additionally investigated whether this was more strongly the case if groups were culturally different from one another (using political affiliation and German vs. Dutch natio­nality). Finally, we introduced a punishment option in the between-group competitive game, and tested whether between-group competition led to increased punishment of free riders.


The first experiment revealed significant ingroup favouritism, but we found no evidence of out­group hostility as culturally contrasting fellow group members were not treated more unfa­vou­rab­ly than randomly assigned group members. The second experiment showed higher levels of cooperation in homogeneous groups, but group homogeneity did not moderate the decline in coope­ration levels across the rounds of the game. Contrary to the parochialism thesis, but in line with recent work in sociology and experimental economics on the effects of diversity on coope­ration (see project 6.2), we find that group homogeneity raises contribution levels of individuals re­gard­less of whether they are in a minority position in a group dominated by cultu­ral­ly different others. This suggests that the positive effect of group homogeneity on coope­ration is not prima­rily a result of parochial ingroup/outgroup biases, but due to other advantages of homo­geneous groups. The third experiment showed that in line with group selection theories, intergroup com­pe­ti­tion raises within-group cooperation levels: Subjects com­pet­ing in a cultural in­group against an outgroup cooperated significantly more than those who were placed in an out­group compet­ing against a group of cultural ingroup members. However, the size of this paro­chialism effect was small compared to that of intergroup compe­ti­tion. In line with the expecta­tions, the final experiment showed that subjects were more likely to punish free riders in inter­group compete­tion settings. Taken together, the results provide strong evidence that group cooperation increa­ses with group homogeneity and intergroup competition. Com­pared to the importance of these factors, support for cultural ingroup/out­group biases as predic­ted by the parochialism thesis is weaker and less consistent. Translated to interethnic relations in the con­text of immigration, the findings support the idea that cultural diversity may undermine com­mu­nities' capacity for collec­tive action and raise the potential for intergroup con­flict. How­ever, the fact that effects of shared fate (i.e. membership of the same payoff group) are much stronger than those of shared cultural group membership suggests that the extent to which such negative outcomes will occur depends importantly on how ethnicity interacts with socio-economic status and material interests.