The non-respondent problem in questionnaire research

Abstract

It is suggested that the real solution to the problem of bias detection, in the event that a numerically large non-respondent group remains after 2 or 3 mail questionnaire follow-ups, is to make up a random sample of this hard core; then to make repeated and persistent attempts at contact. The aim here is a 100% response. An illustration of the application of this technique is given in connection with a large-scale follow-up of 1300 veterans where, after 2 mailings, there still remained some 400 non-respondents. An incidental check, of the test-retest type, is pointed out. (PsycInfo Database Record (c) 2025 APA, all rights reserved)