RESEARCH INTERESTS

Methods:

• Statistical Learning Methods, e.g. Gradient Boosting for Regression Models


• Maschinelles Lernen


• Regression Models, e.g. Mixed Models, Generalised Additive Models for Location, Scale and Shape (GAMLSS)


• Joint Models for longitudinal and time-to event data


• (Quantitative) Methods of Empirical Social Research

Applications:

• Research on Social Inequality


• Labour Market Sociology

TEACHING

• Praktikum Statistische Modellierung (SoSe 2024, SoSe 2025)


• Generalized Regression (SoSe 23, SoSe 24, SoSe 25)


• Current Topics in Applied Statistics (WiSe 23/24)


• Grundlagen Bayesianische Statistik und statistisches Lernen (WiSe 22/23, WiSe 23/24, WiSe 24/25)

Information on the content of the courses can be found here or in the module descriptions.


Thesis offers can be found here or by personal arrangement. I am happy to support own suggestions for thesis topics related to my reserach interests, both regarding model choice and data set search.

PUBLICATIONS

1. Potts, Sophie, Anja Rappl, Karin Kurz, and Elisabeth Bergherr (Mar. 2026). "Bridging the gap: Introducing joint models for longitudinal and time-to-event data in the social sciences". In: Methodology 22(1), pp. 77–108. DOI: 10.5964/meth.18465.



2. Potts, Sophie, Martin Refisch, and Elisabeth Bergherr (July 2026). "Comparing Information Criteria for Prediction-based Variable Selection". In: Proceedings of the 40th International Workshop on Statistical Modelling (IWSM). Oslo, Norway.



3. Potts, Sophie and Elisabeth Bergherr (July 2025). "Joint Models for Rare Events". In: Proceedings of the 39th International Workshop on Statistical Modelling (IWSM). Limerick, Ireland.



4. Potts, Sophie, Elisabeth Bergherr, Constantin Reinke, and Colin Griesbach (Nov. 2023). "Prediction-based variable selection for component-wise gradient boosting". In: The International Journal of Biostatistics 20(1), pp. 293–314. DOI: 10.1515/ijb-2023-0052.

TALKS AND POSTERS

• IWSM 2025 Joint models for rare events (Poster)


• DagStat 2025: Joint models for rare events (Talk)


• Joint models for longitudinal and time-to-event data in the social sciences (Talk)


• CompStat 2024: Joint models for longitudinal and time-to-event data in the social sciences (Talk)


• IWSM 2024: Joint models for longitudinal and time-to-event data in social science research (Poster)


• IWSM 2023: Prediction-based variable selection for component-wise gradient boosting (Talk)


• DAGStat 2022: Prediction-based variable selection for component-wise gradient boosting (Poster)