| Sumario: | The article attempts to assess response error in vote reports in the 1964 election study carried out by the Survey Research Center (SRC) of the University of Michigan, by comparing estimates of voter turnout and partisan division of the vote with Census Bureau data and official voting tallies. The article is concern with the validity of a survey datum falling under the heading of private business. In the election studies conducted by the University of Michigan SRC, a consistent pattern has emerged in the sample estimates of voter turnout and the partisan division of the vote. Estimates of turnout have consistently exceeded the population figures by about 12 to 13 percent. The consistency, as well as the magnitude, of the overreport suggests that the sample-population differences are not due solely to sampling error. The discrepancy between the official estimate of turnout and the SRC estimate for the 1964 election is not greatly different from that observed in previous studies. In the 1964 study, 77.7 percent of the respondents reported a vote for President whereas 62.8 percent is of the official turnout calculated on the population of civilians of voting age. This is a difference of nearly 15 percent, slightly greater than the 12 to 13 percent observed in previous studies.
|
|---|