A welcome and much-needed addition to the literature on survey data quality in social research, McNabb's book examines the most common sources of nonsampling error: frame error; measurement error; response error, nonresponse error, and interviewer error. Offering the only comprehensive and non-technical treatment available, the book's focus on controlling error shows readers how to eliminate the opportunity for error to occur, and features revealing examples of past and current efforts to control the incidence and effects of nonsampling error. Most importantly, it gives readers the tools they need to understand, identify, address, and prevent the most prevalent and difficult-to-control types of survey errors.
The inferential paradigm of probability sampling demands 100 percent cooperation[and response honesty] to guarantee the unbiasedness of a survey estimate. Current best practices argue that researchers should attempt to maximize response rates and to minimize risk of nonresponse errors.
In practice, this chapter-introductory statement by Olson may be summarized thus: The best guarantee of unbiased survey results is for all members of a sample to participate in the survey when asked and to respond to all questions on the questionnaire or panel record. But, that is getting harder and harder to make happen; nonresponse rates continue to climb.
Nonresponse errors result from a failure of the individuals or agency designing a statistical survey to collect complete information from all cases in ...