Statistics in Modern Japan

摘要

Statistics inevitably involve errors. It is widely known that sampling errors occur in sampling surveys. And whether a statistical survey is of the sampling type or a complete enumeration, there will always be non-sampling errors caused by, among other things, mistakes in pollees' reports, oversights in the survey, refusal to participate in the survey, and mistakes in tabulation. Government statistics in Japan are said to be relatively free of non-sampling errors and therefore quite reliable. I believe there are three reasons for this.