Using cognitive interviewing and behavioral coding to determine measurement equivalence across linguistic and cultural groups: An example from the international tobacco control policy evaluation project

This study examined and compared results from two questionnaire pretesting methods (i.e., behavioral coding and cognitive interviewing [CI]) to assess systematic measurement bias in survey questions for adult smokers across six countries (United States, Australia, Uruguay, Mexico, Malaysia, and Thailand). Protocol development and translation involved multiple bilingual partners in each linguistic/cultural group. The study was conducted with convenience samples of 20 adult smokers in each country. Behavioral coding and CI methods produced similar conclusions regarding measurement bias for some questions; however, CI was more likely to identify potential response errors than behavioral coding. Coordinated qualitative pretesting of survey questions (or postsurvey evaluation) is feasible across cultural groups and can provide important information on comprehension and comparability. The CI appears to be a more robust technique than behavioral coding, although combinations of the two might be even better.