Reported increases in nonresponse rates to sample surveys have not been systematically documented to date. Data from the National Election Studies and the Surveys of Consumer Attitudes, two well-known continuing studies conducted by the Survey Research Center at the University of Michigan, permit the assessment of long-term trends for the two major components of nonresponse, refusals and other noninterviews, by urban subgroups using time series regression techniques. The analyses clearly demonstrate that there have been substantial increases in total nonresponse due primarily to increases in the percentages of respondents who refuse to be interviewed, and that these trends are related to the level of urbanization.