Research

I am primarily interested in robust statistics, which is concerned with situations where statistical modeling assumptions fail to hold true. In particular, I study robustness issues in the analysis of categorical data, with a special interest in questionnaire responses. Conventional methods for the analysis of such data can easily yield unreliable results if a small number of data points are noisy, such as study subjects not responding accurately or truthfully. With my coauthors, I develop methods that are designed to be robust against such data contamination. In addition, I am working on machine learning theory and applications.

All of my developed methods are readily available as open-source software packages; see the software page for an overview.

Besides my methodological work in statistics, I enjoy working on applied problems in the health and social sciences.

Working papers

  • Robust estimation of polychoric correlation
    Welz, M., Mair, P., and Alfons, A.
    Under review. [arXiv]

  • Robust estimation and inference for categorical data
    Welz, M..
    In preparation for submission. [arXiv]

  • When respondents don’t care anymore: Identifying the onset of careless responding
    Welz, M. and Alfons, A.
    Under review. [arXiv]

  • How much carelessness is too much? quantifying the impact of careless responding
    Welz, M., Archimbaud, A., and Alfons, A.
    In preparation for submission. [PsyArXiv]

  • Matched samples in strategic management research: Review and recommendations
    Bergh, D., Gallegos Quezada, J., Moelijker, R., Tang, Y., and Welz, M.
    Under revision.

Work in progress

  • Robust estimation of structural equation models, with Patrick Mair and Andreas Alfons..

  • Robust estimation of polyserial correlation

Publications in peer-reviewed journals

  1. A comparative analysis of heterogeneity in lung cancer screening effectiveness in two randomised controlled trials
    Welz, M., van der Aalst, C., Alfons, A., Naghi, A., Heuvelmans, M., Groen, H. J., de Jong, P. A., Aerts, J., Oudkerk, M., de Koning, H., and ten Haaf, K.
    Nature Communications, 16, 8060, 2025.
    [open access]

  2. Open science perspectives on machine learning for the identification of careless responding: A new hope or phantom menace?
    Alfons, A. and Welz, M.
    Social and Personality Psychology Compass, 18(2), e12941, 2024.
    [open access] [code]

  3. Forecasting Real GDP Growth for Africa
    Franses, P.H. and Welz, M.
    Econometrics, 10(1):3, 2022.
    [open access] [code]

  4. Evaluating Heterogeneous Forecasts for Vintages of Macroeconomic Variables
    Franses, P.H. and Welz, M.
    Journal of Forecasting, 2022.
    [open access] [code]

PhD thesis

  • Robust categorical data analysis
    Welz, M., 2025.
    [Doctoral Thesis, Erasmus University Rotterdam]
    [open access]
    I successfully defended my thesis on June 6, 2025. You can find some pictures of the defense here.