Research

The Chair for Psychological Methods, Evaluation, and Statistics conducts methodological research at the intersection of statistics, psychometrics, and machine learning.

We develop methods for detecting parameter differences in psychometric models, as these can affect the fairness of psychological tests. Some methods are based on classical, parametric statistics, while others draw on approaches from machine learning.

In the area of machine learning, we develop methods for assessing stability, as well as for reliably measuring and interpreting the contributions of predictor variables. These topics are also relevant to our work in psychometric computing.

Across all research areas, we contribute to the development of add-on packages for the free, open-source software R. The available R packages are listed below.

For further topics see Publications

SNF project
Development of a Toolbox for Psychological Test Development

SNF project
Detecting Heterogeneity in Complex IRT Models for Measuring Latent Traits

BMBF project
Heterogeneity in IRT Models

DFG project
Methods to Account for Subject-Covariates in IRT-Models

 

BRIDGE Discovery project
Harnessing event and longitudinal data in industry and health sector through privacy preserving technologies

SNF Sinergia project
MULTICAST- A MULTIdisCiplinary Approach to prediction and treatment of Suicidality

DSI project
PREMIA - A Prediction Market with Integrated Algorithms