An online survey with a large sample-size: A case of the survey for demonstrators after the Great East Japan Earthquake

This research examines the validity of an online survey that focuses on respondents with the rare experience. We use the case of the survey for demonstrators against Japan's nuclear and security policies. An online survey is a simple and inexpensive method to gain access to a large group of potential respondents. However, studies have found that samples can be biased regarding age and education and thus provide inaccurate results. In this research, we illustrate how we address these issues and whether the results of our online survey are different from those of two mail surveys (one using a random sampling method and the other using a probability sampled panel). Results show that while the respondents in the online survey are biased regarding age and education, the same is true for respondents to the mail survey using the random sampling method. The three types of surveys do not largely differ in the distributions of attitudes and their correlations. A larger sample size might weaken its selection bias. These results mitigate negative views of online surveys.