The Spiral of Silence Revisited Examining through a CASI (Computer Assisted Self-administered Interview) experimental survey

This paper presents the results from nationwide experiments conducted with the CASI (Computer Assisted Self-administered Interview). Experiments in political science have been either limited to natural experiments or to non-randomly selected samples ranging from laboratory-based experiments to web survey experiments. It has been de facto impossible to conduct experiments with randomly selected samples as it is difficult to randomly split samples in the paper-and-pencil based interviews and to assign treatments based on responses obtained from the same survey. In light of these limitations, we have developed a CASI system in which randomly selected respondents can respond directly to an automated questionnaire in the CASI program. The CASI system is innovative as the use of computers facilitates complicated tasks of experiments and at the same time holds external validity based on random samples selected nationwide, which is absent in laboratory-based experiments and web survey experiments. We as a team of researchers headed by Aiji Tanaka have already conducted CASI surveys seven times since 2007 and have kept on refining the program each time. The latest CASI survey conducted in October 2011 was above all devoted to a series of experimental questions after conducting some pilot studies through web surveys. We hope to contribute to the panel by presenting our results from the above survey experiments and adding survey experiments to the list of methodological varieties.