Nonresponse on Surveys: A Canada-United States Comparison

Abstract

This paper compares survey response rates in Canada and the United States. Initial unsystematic evidence suggests a large margin in favour of the United States on national inter- view surveys. Three hypotheses are entertained: that the difference in response rates is attribut- able to fieldwork procedures employed in each country; that the difference reflects socioeconomic differences between Canada and the United States; and that a cultural resistance to surveying depresses Canadian rates of response below the level expected on the basis of the socioeconomic structure of the society. Using response rate per survey as the unit of analysis, Canadian and American samples of response (N = 443), with codes for details on fieldwork procedures, are used to test the hypotheses. National interview surveys in Canada achieve some 10-15 percentage points lower response than comparable surveys in the United States. We can establish about 7 of these points as attrib- utable to factors other than simply the number of call-backs. The remaining margin does seem due to the lower frequency of call-backs in Canadian surveys. We conclude that it would be diffi- cult to greatly increase call-backs in national-scale surveys, given the Canadian realities of geog- raphy, and economy. Alternative methods of data collection, or surveys on the regional or local scale, may hold a comparative advantage in the Canadian context.