Correction of influence using the different reaction labels of rating scales: A case of trust in public institutions measured by social survey

Abstract

This research introduces a method to compare the responses to rating scales with different rating labels by examining the effects of the 2011 Great East Japan Earthquake on trust in public institutions. Although both the World Values Survey and the Asian Barometer Survey, conducted before and afer the earthquake, included questions about trust in the Self-Defense Force, the police, the courts, the media, political parties, and parliament, their rating labels were different for each survey. Therefore, this research used Bayesian statistical modeling to remove the influence of the label differences and examined the changes in trust before and afer the earthquake. The results demonstrate that while trust in the Self-Defense Force increased, trust in the courts, the media, political parties, and parliament decreased. The correction of social survey data using statistical modeling, as demonstrated in this research, is effective in examining the influence of earthquake disasters and for verifying the short-term or long-term social changes.