A Mixed Methods Approach to the Analysis of Bias in Cross-cultural Studies

Abstract

The aim of the study is to illustrate a comprehensive approach to item bias in cross-cultural research by integrating quantitative and qualitative data in a mixed methods design. We studied responses and response processes of Dutch and Spanish participants to quality of life items from five international studies. Item bias was evaluated by analyzing differential item functioning (DIF) with several statistical procedures. Statistical results were combined with qualitative evidence of response processes from cognitive interviewing (CI). These integrated results yielded a coherent picture of bias sources. Nonuniform DIF was linked to differences in the response processes, whereas uniform DIF was associated with differential meaning of the item contents. We conclude that the mixed methods approach holds important promise for understanding bias.