Anonymity vs. Identification in Studies of Public Opinion

The extremely rapid growth of attitude and opinion research as a field of endeavor has not, in the eyes of some critics, produced the methodological advance which one might expect during the first twenty or twenty-five years of what its practitioners continually refer to as an infant science. This alleged methodological lag may, if it exists, be attributable to the very rapidity of growth, since there have been, as Stouffer has pointed out, so many substantive problems to be met, such a demand for the services of research technicians in problem solving, so many "fires to be put out," that little time and energy have been available for concentration on problems of method. On the other hand, it may be that the popular success of such enterprises as the Gallup Poll have been considered some sort of empirical validation of the methods used by the best-known pollsters.Whether or not methodological progress in attitude measurement has been satisfactory, the major emphasis of what work has been done has been upon the two most basic problems - sampling and questionnaire construction. Relatively little attention has been paid to what might be considered less basic or less important problems like the one to be discussed here; the effect of anonymity as against respondent identification in determining response.In attitude or opinion studies, is there a difference between responses given under anonymous conditions and responses given under conditions calling for identification of the respondent?