Typically, laboratory experiments suffer from homogeneous subject pools and self-selection 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...