Experimental methods in survey research

Abstract

In recent years, interest in experimental approaches has been growing in social science research. In the social sciences that target on society and actors in society, endogeneity always exists between cause and outcome, and methods such as covariate adjustment and natural (quasi-) experiments have been proposed as methods to deal with it. However, the experimental approach is still described as the best method of causal inference, and it has expanded rapidly by establishing an infrastructure that facilitates experimentation. This paper focuses on the theoretical aspects of various methods of survey experiments. Specifically, we explain the outline, implementation, estimation, and note for each of framing experiment, list experiment, and conjoint experiment.