Typically, laboratory experiments suffer from homogeneous subject pools and selfselection biases. The usefulness of survey data is limited by measurement error and by the questionability of their behavioral relevance. Here we present a method integrating interactive experiments and...