Personal Follow-Up in a Mail Survey: Its Contribution and Its Cost

This article addresses itself to the following questions: Who are the non-respondents to a mail survey, and what effect does their absence have on the achieved sample? What types of people are brought in through personal contact? Does their inclusion improve the sample estimates appreciably over the mail response alone, and at what cost? Can extrapolation of the data be used to improve the estimates? On data for which parameters are not known, to what extent does the personal follow-up change the estimates derived from mail questionnaires? Because the population from which their respondents are drawn had previously been completely enumerated, the authors can give rather precise answers to all these questions, of concern to everyone doing survey research.