Max Melchior Lang

About

I like to think of statistics as the honest interpretation of data — even when the honest interpretation is less catchy than the one we might want to tell. Whether we're fitting models to reality or trying to fit reality to our models is worth thinking about. The latter rather defeats the point, although is often more marketable.

I work on all kinds of (statistical) problems. Currently I am interested in modelling Malaria and Schistosomiasis. This involves spending a lot of time thinking about generalised additive “smooth” models and looking at satellite imagery (GIS) in various geometric configurations (hexagons, rasters, polygons). I'm part of the Schistotrack Group, named after the ongoing cohort study in Uganda. I am supervised by Goylette Chami and Christl Donnelly placed in the Department of Population Health and Statistics at the University of Oxford.

Previously under the supervision of Frauke Kreuter in Munich, I worked on the rather unfashionable problem of asking good questions and collecting decent data to answer them. Understanding that people would rather click boxes than type thoughtful responses matters often more than any sophisticated model, also it's worth remembering that data are only reliable if they're representative of the population you actually want to learn about, not just whoever happened to respond.

This led to various collaborations with 60 Decibels on automated telephone surveys (now featuring LLMs, because everything must), with researchers from LMU Munich on text-based adaptive interviewers (also LLMs, naturally), and with their medical department on conversational bots (which more recently have been rebranded as agents) for consultations and appointment reminders — inevitable perhaps, with Germany aging faster than the healthcare system can manage.

I am always happy to discuss research ideas and look forward to new collaborations. Feel free to reach out via email if you'd like to connect.