抄録
In this paper, I examined the possible use of random digit dialing sampling (RDD), focusing on the method improved by Waksberg which is often used by the public health professions. Since the telephone possession rate is 90% or more, the probability of selection bias is very low. There is the possibility that individuals who live out of the target area may inadvertently be included in the survey since the object area of the telephone may not coincide with the area of an administrative district. However, this situation can be avoided by confirming the living area at the time of enforcement or by expanding the area of the survey. The numbers called and the times of the survey calls also have an effect on the proportion of survey participants. The importance of making a sufficient number of calls and at the most opportune survey time is recommended for emphasis at the survey design stage.
