Speaker：Tetsuro KOBAYASHI [Department of Media and Communication, City University of Hong Kong]

Date：October. 19, 2021／15:00‐16:40 (JST)

Location：Zoom

Target: Open to the public

Abstract：Experiments have come to be a well-established tool for causal inference in the political sciences but, in practice, there is a significant bias towards surveys and the application of other experimental designs is relatively rare. The situation is the same for political communication research, which typically focuses on the media effects on political attitudes and behavior. In most laboratory and survey experiments, the treatment is administered once, with only a short period of time before outcomes are measured. As such, the persistence and long-term impacts of treatment effects are often unclear, so there is the lingering criticism that only so-called "knee-jerk reactions" can be measured. The experimental conditions in laboratory and survey experiments can also be unrealistic, and the highly hypothetical nature of stimuli in scenario-based experiments raises questions about the ecological validity of findings. Field experiments, in which treatments are embedded in the daily lives of participants and administered over a prolonged period, are one way to deal with such limitations. Field experiments generally feature significant levels of noise due to the experiment environment and are costly to implement, but the spread of digital media is making it possible to design experiments more flexibly. In this presentation, I would like to introduce field experiments that I have conducted in the past, and discuss the methodological and ethical hurdles we encountered during those experiments. Specifically, I will talk about field experiments using mobile apps, using social media to investigate an election campaign, and using a mock portal site to investigate political learning.