Methodeology Research
Our research focuses on the development and application of quantitative methods for questions related to effectiveness research, the measurement of psychological traits such as competencies, interests, and attitudes, and the replication of research findings.
The goal is to increase the validity of empirical findings by fully utilizing data and interpreting results reliably. The research addresses, how to better control for confounding factors in causal inference, how to measure psychological traits fairly, and how to integrate new data sources, such as process and product data from digital surveys. These methodological developments are also applied to meta scientific questions regarding replication and systematic research synthesis.
This research focus provides important input for empirical psychological research, with the findings highlighting new possibilities for statistical modeling and making methodological innovations applicable in practice.
Causal Inference
Causal inferences regarding the evaluation of treatment effects and the explanation of human behavior and experience are fundamental to psychological research. We develop methods that simultaneously account for various methodological challenges (such as selection effects, measurement error, non-adherence to treatment instructions, and clustered data) and enable the investigation of the differential efficacy of treatments (such as group- and individual-specific effects). A current project in collaboration with the CREATE Cluster of Excellence at the University of Oslo investigates the integration of response times into efficacy analyses using the example of app-based learning interventions.
Psychometrics
The fair (longitudinal) measurement of constructs that cannot be directly observed, such as competencies, interests, and attitudes, serves as the starting point for many further research questions. In this context, we develop psychometric approaches to address methodological challenges (such as the multidimensionality of measurements and violations of measurement invariance) and examine the validity of measurements. These methods are applied, for example, in large-scale educational studies in collaboration with the Leibniz Institute for Educational Trajectories (LIfBi). Currently, we are using the diversity of data in this context to apply and further develop various machine learning methods for investigating test fairness.
Meta Science
Research synthesis and the replication of scientific findings are central to a thorough understanding of results, as well as to theory formation and testing. With regard to this research focus, we evaluate the causes of effect heterogeneity, which is regularly found in replication studies and meta analyses. The fundamental approach to explaining the variation in scientific findings involves applying methods from causal theory and measurement theory to meta-scientific applications. Existing collaborations within the DFG Program Meta-Rep and with the Collaboratory Replication Lab provide interdisciplinary and international insights into current replication research.