ABSTRACT
Since the end of the 1960s, researchers on both sides of the Atlantic have reported a deterioration in response rates on face-to-face interview surveys of general populations. To acknowledge that increasing survey nonresponse has sensitized methodologists to the strength of possibly already existing privacy values is distinct, however, from a demonstration that a growth in these values helps account for the response trend. The comparison between predicted and actual nonresponse in effect estimates the shortcomings of a model that includes only researcher input, neglecting historical context. True, the computations cannot deny the possibility that supposed historical effect really involves inputs unspecified in the model. The trend is, starting with American interview surveys, a pronounced increase in nonresponse, particularly between the 1960s and the 1970s. The American commitment to mass education, in which markedly higher proportions than in Canada or Great Britain experience post-secondary education, predicts higher response on American surveys.
