A comment on ‘growth and inequality in public good provision’: Testing the robustness and generalizability of dynamic public good games

I revisit the dynamic public goods game by (Gächter et al.,2017) that captures temporal interdependencies in cooperation. I first reproduce parts of the original analysis and find only minor and inconsequential discrepancies. Using both student and a more representative sample, I then replicate findings about growth and inequality in public good provision. Finally, I examine whether the robustness of results also translates into generalizability. Specifically, I test whether behavior in this experimentally-induced social dilemma predicts real climate action through voluntary carbon offsetting. Despite the game’s enhanced ecological validity through temporal interdependencies, I find no correlation between game behavior and climate action in either sample. This suggestive evidence indicates that laboratory paradigms, even when incorporating key real-world features, may better serve to isolate specific behavioral mechanisms than predict field behavior. • Reproduces dynamic public goods experiment. • Compares different samples. • Finds similar behavior across samples that are also similar to behavior in standard public goods experiments. • Shows that contribution behavior in this abstract game does not generalize to voluntary climate action.