In this paper, we test for scope effects in Contingent Valuation applying different distributional assumptions for WTP, a non-parametric estimation and an estimation based on an open-ended format. Mean WTP is sensitive to the distributional assumption, but so is the scope effect. The nonparametric model, without conditions on the distribution, is the best able to identify scope effects. More sophisticated models, such as the spike model, and open-ended follow-up question give more information about individual WTP and are thus more powerful in revealing scope effects. For small sample size, a non-parametric analysis, a spike model or an open-ended format may therefore be better than the classical parametric dichotomous choice analysis to identify scope effects.