Recent Methods of the Questionnaire Research with the Large Sample Size (2): Issues of Telephone and RDD Surveys

How to Deal with Coverage and Nonresponse Errors

Abstract

Early telephone surveys were conducted by selecting samples from telephone directories and public registers. Due to errors in survey results caused by under-coverage of individuals not listed in telephone directories, telephone survey methodology changed from the original list-assisted model to Random Digit Dialing (RDD). Surveys for samples added cell phone numbers are able to cover cell-phone-only (CPO) individuals, which now includes many young and middle-aged persons. The downsides to mobile surveys are that they have a very low response rate, and that in Japan it is difficult to obtain proper regional information from a telephone number. This article offers information to aid in the process of deciding whether or not to use telephone surveys.