Mixed strategies for improving data quality: the contribution of qualitative procedures to survey research

This article aims to illustrate how qualitative research strategies contribute to the implementation of the standard survey model by monitoring and improving the quality of collected data and thus the quality of survey findings as well. The methodological cycle for survey data-quality construction using qualitative strategies comprises three sequential stages that alternate with standard procedures: the pilot study, pretesting, and deviant case analysis. If the first two of these stages have enjoyed a certain recognition in the literature, the same cannot be said for deviant case analysis. Deviant case analysis is a research strategy originally proposed by Lazarsfeld and promoted within the Colombia School from the 1940s to the 1960s, unfortunately without subsequently receiving the attention it deserved within the area of survey research in general. It refers to searching outside the collected empirical base for clues concerning anomalous responses revealed by statistical analysis tha

This article aims to illustrate how qualitative research strategies contribute to the implementation of the standard survey model by monitoring and improving the quality of collected data and thus the quality of survey findings as well. The methodological cycle for survey data-quality construction using qualitative strategies comprises three sequential stages that alternate with standard procedures: the pilot study, pretesting, and deviant case analysis. If the first two of these stages have enjoyed a certain recognition in the literature, the same cannot be said for deviant case analysis. Deviant case analysis is a research strategy originally proposed by Lazarsfeld and promoted within the Colombia School from the 1940s to the 1960s, unfortunately without subsequently receiving the attention it deserved within the area of survey research in general. It refers to searching outside the collected empirical base for clues concerning anomalous responses revealed by statistical analysis that either deviate from research hypotheses or produce contradictory classificatory results. © 2013 Springer Science+Business Media Dordrecht.