Ensuring valid comparisons of self-reports in heterogeneous populations and marginalised groups (ENSURE)

PH-LENS研究单位的许多子项目都依赖于通过调查收集数据。在对难民和其他边缘人口进行调查时，应调整数据收集工具以供这些群体使用。根据德国数据论坛（RatSWD）的质量标准，在开发新的或改编现有测量仪器时应确保高可靠性和有效性。在比较调查中，测量不变性——任何组间比较的先决条件——应被额外考虑，这意味着组成员不应使比较结果产生偏差。在跨文化调查中，建立调查数据的测量不变性一直是一个方法上的挑战。不同的事前方法和事后数据分析方法被用来达到和评估可比性。一种方法是认知预测试，它在问题引导过程中被用来建立更高的测量质量和跨文化可比性。另一种方法是纳入锚定的小片段，允许在考虑到答案选项的不同用法的情况下对数据进行工作后调整。修正响应集是锚定小片段的潜在替代方案。然而，这些方法是否能有效地建立测量不变性和提高测量质量，还缺乏经验证据。第一个目标是确保与PH-LENS相关的仪器适配的高质量测量（可靠性和有效性）和测量不变性。在与子项目NEXUS、LARGE、OTHER和TREAT的密切双向合作中，测量仪器将进行调整、认知测试和改进。将评估适应仪器的测量质量和测量不变性。第二个目标是评估开发高信度和效度的测量不变仪器的程序。本部分的重点是比较跨文化认知预测试作为事前方法与事后数据纠正时使用锚定小片段和纠正反应集。为了比较这两种方法，将对一个大州（巴登-符腾堡州）的难民进行一项实验研究，与子项目NEXUS和TREAT共享资源。第三个目标是评估测量不变性的事后分析统计方法（贝叶斯扫描电镜，对齐方法，精确测量不变性）在检测认知前测试已知的可比性问题方面的敏感性。PH-LENS的子项目受益于测量仪器的质量保证，以及用于异质人群和边缘群体的经过调整、评估和记录的高质量测量仪器。实现测量不变性的高质量测量工具和方法也将有利于健康、行为和社会科学研究的进一步领域。

Many sub-projects in the research unit PH-LENS rely on data collection via surveys. When refugees and other marginalised populations are surveyed, data collection instruments should be adapted for use in these groups. According to the quality standards of the German Data Forum (RatSWD), high reliability and validity should be ensured when developing new or adapting existing measurement instruments. In comparative surveys, measurement invariance – a prerequisite for any between-group comparisons – should be additionally taken into account, which means that group membership should not bias the results of comparisons. Establishing measurement invariance in survey data has been a methodological challenge in cross-cultural surveys. Different ex-ante approaches and ex-post data analysis methods have been used to reach and evaluate comparability. One method is cognitive pretesting, which has been used to establish higher measurement quality and cross-cultural comparability during the piloting of questions. Another method is incorporating anchored vignettes that allow an ex-post adjustment of the data taking different usage of answer options into account. Correcting for response sets is a potential alternative to anchored vignettes. However, empirical evidence is lacking on whether these methods are effective in establishing measurement invariance and can improve measurement quality. The first objective is to ensure a high quality of measurement (reliability and validity) and measurement invariance of adaptations of instruments relevant to the PH-LENS. In close bi-directional work with the sub-projects NEXUS, LARGE, OTHER and TREAT, measurement instruments will be adapted, cognitively tested, and improved. Measurement quality and measurement invariance of adapted instruments will be assessed.The second objective is to evaluate procedures for developing measurement-invariant instruments of high reliability and validity. This part focuses on a comparison of cross-cultural cognitive pretests as an ex-ante method with ex-post data corrections when using anchored vignettes and correcting for response sets. To compare the methods, an experimental study on refugees in a large state (Baden-Wuerttemberg) will be conducted, sharing resources with the sub-projects NEXUS and TREAT.The third objective is to evaluate statistical methods of ex-post-factum analysis of measurement invariance (Bayesian SEM, alignment method, exact measurement invariance) with respect to their sensitivity in detection of problems of comparability known from cognitive pretests.The sub-projects of the PH-LENS benefit from the quality assurance of the measurement instruments as well as from the adapted, evaluated and documented high-quality measurement instruments for use in heterogeneous populations and marginalised groups. High-quality measurement instruments and methods to reach measurement invariance will also benefit further areas of health, behavioural and social science research.