Common ratio and common consequence effects arise from true preferences

Recent contributions suggest that the empirical evidence for the common ratio effect could be explained as noise instead of underlying preferences under "common assumptions." We revisit this argument using a more general method which allows to unambiguously distinguish noise from preferences nonparametrically and with less stringent assumptions. The results are independent of the assumed behavioral model or how noise affects choices. Applying this method to new experimental data we show that there is a systematic preference for the common ratio and the common consequence effects which cannot be explained by noise.