This paper provides a new statistical model for repeated voluntary contribution mechanism games. In a repeated public goods experiment, contributions in the first round are cross-sectionally independent simply because subjects are randomly selected. Meanwhile, contributions to a public account...