Pinned Loading

1. dices-safety-audit dices-safety-audit Public

   A psychometric audit of the DICES-350 conversational-AI safety taxonomy, testing whether its categories are reliable, dimensionally coherent, and empirically distinct

   Python

2. typology-paper typology-paper Public

   A two-dimensional typology of 29 OECD countries classifying institutional exposure to AI-driven labor market change. Replication code, data, and working paper (Mazurek, 2026).

   Python

3. eval-rubric-design eval-rubric-design Public

   How I design AI harm-evaluation rubrics: a query safety taxonomy for a general-purpose conversational assistant, and the reasoning behind each decision. Designs, not run evaluations.

4. ess-redistribution-analysis ess-redistribution-analysis Public

   Multilevel analysis of redistribution preferences across 28 European countries using ESS Round 9 data, with agent-based simulation of inequality shocks

   HTML

5. ThesisPipelineRework ThesisPipelineRework Public

   Pipeline rebuild of MSc thesis on interest group prominence in U.S. congressional speech. Improved classifier (F1=0.91, kappa=0.82), multilevel models, interactive dashboard.

   Jupyter Notebook

6. public-attitudes-research public-attitudes-research Public

   Multilevel analysis of neighbourhood income composition and support for redistribution (SCoRE Netherlands 2017 + CBS)

   R