Aggregating qualitative district-level campaign assessments to forecast election results: Evidence from Japan
针对日本媒体以定性评估而非民调数字报道竞选的情况，本研究开发了一种聚合定性选区竞选报道的方法，并应用于预测2017年日本众议院选举结果，验证了其准确性。
The poll aggregation is conducted in the USA and European democracies for electoral forecasting. However, this has not been the case in Japan because the news media report on electoral campaigns with qualitative assessments rather than poll numbers, although these assessments are based on extensive polling. Our study developed an approach to aggregate qualitative district-level election campaign coverage in Japan, applied the method to forecast the outcomes of the 2017 general election for Japan’s Lower House of the National Diet, and assessed the accuracy of the forecast against the actual results. We integrated the qualitative assessments by using the item response theory, which effectively predicted the electoral results. The method proposed in this paper can be applied to aggregate qualitative assessments by experts in other countries, such as the Cook Political Report in the USA. This would improve the accuracy of election forecasting when combined with existing approaches.