De kwaliteit van schriftelijke vragen over alcoholconsumptie: vergelijking van een bureauonderzoek met een veldonderzoek

Samenvatting

Data quality (validity, accuracy, completeness, and reliability) is a generally acknowledged problem in survey research. One main source of error is the wording and presentation of questions. Here, the results of two approaches, ie, an expert review and a field study, applied to the same set of six questions on alcohol consumption contained in a Dutch Survey. are compared. The expert review involved close reading, comparison of questions with other survey questionnaires, and interviewing authors of the questionnaire. The field study was a three-step procedure consisting of think-aloud protocol analysis, cognitive interviewing, and intensive substantive interviewing. In comparing the results of these two studies, it appeared that most problems identified in the expert review were confirmed in the field study. But the field study identified many problems that were not discovered in the expert review. These respondent problems stemmed mostly from irregular drinking patterns and local normative meanings. Findings indicate that the three-step test interview is an approach in its own right that identifies problems that cannot be identified through other methods. It should, therefore, be used more widely to pretest and test survey questions. 1 Table, 2 Appendixes, 43 References. Adapted from the source document.