Socioeconomic Status and Familial Variables in Mail Questionnaire Responses

The study of bias introduced by nonreturns in mail questionnaire research has survived the either-or and all-or-none foci of the early questionnaire-versus-interview polemics in which the finding of non-returnee bias was frequently a sufficient rationale for throwing out the method with the data.2 In fact, it could be argued that a typology is now needed to differentiate among the varieties of methods, substantive areas, and research designs associated with findings on respondent-non-respondent differences, particularly when such findings are regarded as potential means for maximizing the number of returns3 and eventually may make it possible to correct for predictable amounts and kinds of bias when reporting mail-questionnaire data. The present paper is intended to further the efforts to ascertain and correct for predictable directions and amounts of non-response biases. This intent rests on the assumption that such biases will become measurable conditions of questionnaire research when sufficient descriptive data become available from studies in which the substantive area, the sponsor, the subjects' involvement, and the research methods are comDarable and sDecified. The data reported here have relevance for mail-questionnaire research in general and implications for mail-questionnaire research on the family in particular. In fact, mail-questionnaire research would appear, from the present study, to elicit a disproportionately high response from subjects whose self-reported backgrounds in "normal" lower-middle-class families4 skew familial data in the direction of what has become the predominant textbook model of the "nice," "happy," stable, middle-class family.