The study was undertaken as part of Sarah Fisher's PhD research. It investigated whether a classic risky-choice framing effect depends on the semantic inequivalence of the options presented in alternative framing conditions. The dataset comprises response data from two experiments, both of which were run online using Qualtrics/Amazon Mechanical Turk. The experiments involved Mechanical Turk Workers completing a choice task, based on a short scenario concerning the outbreak of an unusual Asian disease. In the first experiment (n=305), each participant received one of four versions of the scenario. In the second experiment (n=308), each participant received one of three versions of the scenario. After completing the choice task, participants answered a small number of demographic/screening questions.