Voluntary cooperation in public goods problems crucially affects the functioning and long-term fate of economic and political systems. Previous research emphasizes that cooperation in public goods games correlates with expectations about cooperation by others among students and other selected demographic subgroups. However, determining if this reciprocity effect is causal and a general feature of individual behavior requires the use of randomized experiments in combination with large-scale samples that are representative of the population. We fi elded large-scale representative surveys (N=8,500) in France, Germany, the United Kingdom, and the United States that included a public goods game in combination with a novel randomized experiment and a survey instrument eliciting individual's conditional contribution schedules. We find a positive causal effect of higher expected cooperation on individual contributions that is most pronounced among positive reciprocity types which account for about 50% of all individuals. We also show that positive reciprocity is unevenly distributed: It is more widespread among richer, younger and more educated respondents. Therefore, socio-demographic characteristics matter for understanding behavior in social dilemmas because of their association with conditionally cooperative strategies.