Teaser image to Multiverse Analysis: On the Robustness of Functional Form and Data Pre-Processing Decisions

Colloquium

Multiverse Analysis: On the Robustness of Functional Form and Data Pre-Processing Decisions

Cristobal Young, Cornell University

   05.06.2024

   4:15 pm - 5:45 pm

   LMU Munich, Department of Sociology, IfS 309 and online via zoom

Functional form assumptions are central ingredients of a model specification.

Just as there are many possible control variables, there is also an abundance of estimation commands and strategies one could invoke, including ordinary least squares (OLS), logit, matching, and many more. How much do empirical results depend on the choice of functional form? In this talk we demonstrate the functional form multiverse with two empirical applications: how job loss affects wellbeing in panel data and the effect of education on voting for Trump. We find in our cases that OLS and logit produce very similar results, but that matching estimators can be surprisingly unstable. We also reconsider a key many-analysts study and find that human researchers produce a much wider range of results than does the multiverse algorithm.


Related

Link to From Concepts to Certificates: Towards Controllable Generative Models

Lecture  •  20.11.2026  •  Oettingenstraße 67, Room 151, Munich

From Concepts to Certificates: Towards Controllable Generative Models

Francesco Leofante from Imperial College London presents an approach to controlling generative models through concept unlearning.

Read more
Link to Who’s Listening? Levels of Automation for AI That Hears Students’ Math Talk

Guest Talk  •  14.12.2026  •  Marsstraße 20, Room 605, 80335 Munich

Who’s Listening? Levels of Automation for AI That Hears Students’ Math Talk

Dora Demszky from Stanford explores how much work AI should take on when listening to students’ small-group math discussions.

Read more
Back to Top