No abstract

This paper describes the cognitive testing of eight self-report questions as part of a broader effort to evaluate and validate the use of these questions in estimating the prevalence of periodontitis in the United States population. This study examined how United States respondents understood and processed the proposed questions in English and Spanish, as well as identifying and correcting for possible response errors. The set of eight questions was selected by the Centers for Disease Control and Prevention Periodontal Disease Surveillance Workgroup for further testing after analytical assessments and field testing of an array of potential questions. Evaluation of these eight oral health questions was based on 40 in-depth, semi-structured cognitive interviews in English and Spanish. Results of this cognitive test study are presented. The recommendations from this cognitive testing evaluation served as the basis to improve the original questions in English and Spanish to be more inclusive and consistent and improve the estimation of periodontal disease in the United States population.

The rise in multilingual surveys within the U.S. federal government and around the world has led to an increased need for assurance of comparable survey questions and resulting data. Cognitive interviewing (CI) is one method that allows us to examine how different language versions are interpreted and thus detects ways in which translated questions are not comparable. However, despite being a well-established question evaluation

This article summarizes the work of the Comparative Cognitive Testing Workgroup, an international coalition of survey methodologists interested in developing an evidence-based methodology for examining the comparability of survey questions within cross-cultural or multinational contexts. To meet this objective, it was necessary to ensure that the cognitive interviewing (CI) method itself did not introduce method bias. Therefore, the workgroup first identified specific characteristics inherent in CI methodology that could undermine the comparability of CI evidence. The group then developed and implemented a protocol addressing those issues. In total, 135 cognitive interviews were conducted by participating countries. Through the process, the group identified various interpretive patterns resulting from sociocultural and language-related differences among countries as well as other patterns of error that would impede comparability of survey data.