Check Me, Replicate Me

Everybody makes mistakes. Sometimes, this happens unnoticed. And sometimes, it doesn’t even matter. However, there are some areas where it becomes particularly troublesome when we do so. And science, I think, is one such area. Luckily for us, science (and philosophy, I might add) is a collaborative enterprise. And uncovering mistakes is an integral part of it. In this spirit, I invite you to check and replicate my work. To help you with this, I maintain repositories containing essential files (e.g., data and analysis files) for research projects where I am the first or corresponding author (see publications).

Check My Work

For a long time, inspired by Gilad Feldman,¹ I offered to pay a bounty for finding errors in my preprints and published work. With the rise of LLMs, Gilad Feldman has discontinued his bounty program in September 2026, writing:

With the introduction of powerful frontier LLMs it’s now easier than ever to check manuscripts. In the last week I received several emails sending me reports of automated checks flagging potential minor errors in my published manuscripts. This is wonderful, truly, and I love it that people are now leveraging LLM models to check previous literature. But it also means that the rewards set in place before are no longer needed, what is needed is for every researcher, me included, to check their prior work with LLMs, and correct the public record.

And so, I am following suit. Of course, reports of errors in my work remain very welcome. I collect all errors found in my preprints and published work in a separate section below.

Replicate My Work

Should you be interested in replicating my work, there are a few things I offer to make this as easy as possible for you:

For projects where I am the first or corresponding author, I maintain repositories containing all the essential files.
If anything is missing from a repository, please don’t hesitate to contact me, and I’ll try to add the missing files or information if possible.
If you are working on replicating my work, I will be happy to give you feedback and answer any questions you have.

Published Errata

Bauer, Alexander Max, Adele Diederich, Stefan Traub, and Arne Robert Weiss (2025): »Correction to: Thinking About Need«, The Journal of Economic Inequality. Link

Further Errors

In Austin in the Lab, we reported our sample’s mean age as 48.45; it should be 48.50.
In Poisoned Babies, Shot Fathers, and Ruined Experiments, seven numerical errors were found.
- GFCI replication, statement (2): the reported 95% CI belongs to a different statement; it should be [3.434, 4.799].
- Poisoned Cup, IOC exclusion (2): the running text gives N = 81; it should be N = 89.
- Poisoned Cup replication: the analysis file divides by √50 rather than √71; reported r = 1.166 and r = –0.842 should be r = 0.978 and r = –0.706.
- GFCI CRC exclusion: the analysis file divides by √50 rather than √67; reported r = 0.917 and r = 0.254 should be r = 0.792 and r = –0.219.
- Revolver, simultaneous exclusion: the sample should be 31 males, not 30.
- GFCI CRC exclusion: the mean age is 46.269, not 46.267.

Gilad Feldman got the idea from Stuart Ritchie, who himself adapted it from Ruben Arslan, José Luis Ricón, and the Knuth Reward Check. ↩