Most disciplines rely on economic games to measure prosocial behavior in controlled experimental settings. However, participants’ comprehension of these games might be lower than desirable, which complicates interpretation of results. We here assess subject comprehension of the payoff structure of five standard economic games commonly used to study prosocial behavior: the Dictator Game, Ultimatum Game, Trust Game, Public Goods Game, and Prisoner’s Dilemma. Participants were recruited from two online platforms: Prolific (n = 528) and CloudResearch (using the CloudResearch MTurk toolkit; n = 540). The Trust Game had the highest level of misunderstanding (70%), followed by the Public Goods Game and the Prisoner’s Dilemma (each at 52%), the Ultimatum Game (27%), and the Dictator Game (24%). Study platform was a significant predictor of misunderstanding in the Dictator Game, Ultimatum Game, and Public Goods Game, with greater misunderstanding on Prolific than CloudResearch. Incentivizing the comprehension questions had no significant impact on misunderstanding in any of the games. The only variable that significantly predicted misunderstanding across all games was numeracy, which was associated with lower misunderstanding. Finally, we found suggestive evidence in exploratory analyses that misunderstanding predicts greater contributions in the Public Goods Game (in line with previous studies) and in the Dictator Game, increased the likelihood to choose the option that maximizes total payoff in the Prisoner’s Dilemma and reduced back transfers in the Trust Game. These findings suggest that misunderstanding may be an important factor in explaining prosocial behavior and that reliance on standard one-shot games may lead researchers to overestimate the importance of social preferences.