Abstract.According to Champ et al. (1997), we have to differentiate between willingness to pay (WTP) and willingness to donate (WTD) in eliciting individual preferences for public goods. Because strong incentives exist to understate the willingness to spend on the public good and to freeride on other people's contributions, WTP forms the upper limit for WTD. This paper presents an experimental study on the size of the gap between WTP and WTD in the context of green electricity promotion, where making electricity markets ''greener'' can be interpreted contributing to the public good "environmental quality". Our theoretical model is based on the impure public good model Sandler, 1984, 1994). It enables us to directly investigate the extent of free-riding in the sample. In the individual-choice scenario, subjects are induced to act as market participants and thereby reveal their WTP, while the public-choice scenario is designed as a majority vote on the preferred level of green electricity. Since free-riding is impossible, the latter treatment reveals subjects' WTD. The data analysis reveals a strong treatment effect concerning the decision mode: under the public choice treatments, participants were willing to pay about three times higher contributions to green electricity than under the individual choice treatments. This observation highlights the large extent of free-riding in the private provision of the public good "environmental quality".