"Identifying voter preferences for politicians’ personal attributes: a conjoint experiment in Japan": Replication and Extension

Horiuchi, Smith, and Yamamoto (2020) found through differences between observational data and their conjoint survey results that Japanese voters' voting preferences are dependent on external factors. The conjoint survey results yielded consistent voter preferences across priming on different electoral systems; however, observational data and theoretical explanations show that voters' preferences do vary across electoral systems. I was able to successfuly replicate these results and findings. My extension focuses on examining the difference in marginal means instead of the difference of average marginal component effects, which avoid setting an arbitrary baseline when calculating treatment effects.