Integrating Rasch and Compositional Modeling for the Analysis of Social Survey Data

Survey data are still ubiquitous in various fields of science, capturing human attitudes and opinions about relevant aspects of daily life. However, they often contain more information than typically con- veyed. Understanding individual response processes can reveal instances of hesitancy and decision uncertainty, which shed light on the unobserved response mechanisms. We present a method to extract as much valuable insight as possible from survey responses using Item Response Theory tree and Compositional Data Analysis. Illustrated with a case study on reactions to the war in Ukraine, our approach provides an alternative framework for analyzing survey data.