Correlates of Respondent Accuracy in the Denver Validity Survey

Although social research has seen substantial advances in methodology in recent years, the “Denver Validity Survey” (as it is called by those familiar with its findings), which was conducted almost twenty years ago, still remains the only probability sample survey designed specifically for the purpose of measuring against official records the relative accuracy of responses on a range of questions of fact. Previous articles have discussed the project design and the sample plan, the relationship of interviewer characteristics to responses received, and the aggregate findings regarding respondent accuracy on factual items.1 The primary focus of the present article will be upon the relative accuracy of respondents of different characteristics—such as age, sex, and socioeconomic status—on the various types of factual questions.