The Cornerstones of Survey Research

Abstract

The goal of this chapter is to introduce the readers to the central issues in survey quality, to discuss the decisions that must be made when designing and implementing a survey, and to review the current methodological and statistical knowledge as presented in the Handbook. Principles important for designing good quality surveys are discussed in terms of the four cornerstones of survey research: coverage, sampling, response, and measurement. These four cornerstones are connected via the specification of the research problem. Specification is concerned with conceptualisation and operationalization, where the research question is translated into survey questions for specific populations and ultimately into observed variables. Coverage and sampling refer to statistical issues that arise because we survey a sample instead of the entire population. Coverage errors occur when the sampling frame does not exactly matches the intended population. Sampling errors occur because we observe only a sample and not the population. If the sample is a probability sample, statistical theory can estimate the amount of sampling error. Response errors, usually designated as nonresponse errors, occur when some of the intended respondents cannot be reached or refuse to respond. If the nonrespondents are not a random sample from all intended respondents, nonresponse can result in biased estimates. Finally, measurement errors refer to discrepancies between the observed measurement and the true value due to the measurement instrument. Measurement errors can originate from many sources, including the interviewer, the respondents, and the specific survey questions.