Pohontsch N, Meyer T
Institut für Allgemeinmedizin, Universitätsklinikum Hamburg-Eppendorf.
Institut für Epidemiologie, Sozialmedizin und Gesundheitssystemforschung, Medizinische Hochschule Hannover.
Rehabilitation (Stuttg). 2015 Feb;54(1):53-9. doi: 10.1055/s-0034-1394443. Epub 2015 Feb 12.
Questionnaires concerning subjective health status are an important element of rehabilitation research. The appraisal of the quality of these instruments mostly relies on quantitative psychometric analyses. However, these analyses do not explicitly reveal whether or how respondents understand questionnaire content. Over the past few years cognitive interviewing has been increasingly used in questionnaire design and validation. It serves to identify potentially problematic questions, ambiguities and difficulties which could lead to unintended answers. It analyses whether the answers given by respondents represent the intended meaning of the question. Findings derived from cognitive interviewing serve to improve new and further validate well-established questionnaires.The 4-stage model of the survey response process by Tourangeau provides a conceptual basis for cognitive interviewing. The 2 most prominent methods of cognitive interviewing are think aloud and verbal probing. Various authors give recommendations on executing cognitive interviews but almost no recommendations exist on the -indications of the different methods.Potential applications of cognitive interviewing go beyond questionnaire design and improvement. Due to its origin in cognitive science it can also be used to resolve substantive questions, e. g. concerning reasons for discrepancies between the results of 2 different methods of measuring change.
关于主观健康状况的问卷是康复研究的重要组成部分。对这些工具质量的评估大多依赖于定量心理测量分析。然而，这些分析并未明确揭示受访者是否理解或如何理解问卷内容。在过去几年中，认知访谈在问卷设计和验证中越来越多地被使用。它用于识别可能存在问题的问题、歧义以及可能导致意外答案的困难。它分析受访者给出的答案是否代表问题的预期含义。认知访谈得出的结果有助于改进新问卷并进一步验证成熟问卷。图兰乔的调查反应过程四阶段模型为认知访谈提供了概念基础。认知访谈最突出的两种方法是出声思考和言语探查。许多作者给出了进行认知访谈的建议，但对于不同方法的适用情况几乎没有建议。认知访谈的潜在应用超出了问卷设计和改进的范畴。由于其起源于认知科学，它还可用于解决实质性问题，例如关于两种不同测量变化方法的结果存在差异的原因。
