Gerd Gigerenzer: Mindless Statistics

At his talk at the Broken Science Epistemology Camp in March 2024, psychologist Gerd Gigerenzer addresses what he calls ‘mindless statistics.’ Rather than think critically about the implications, impact, and meaning of studies, researchers are stuck in a system which rewards publication and statistical significance. The pressure to publish has encouraged them to p-hack, slice data, and essentially cheat their way into showing their research is important.
Professor Gigerenzer dives into the history of statistical analysis in his field of psychology, which has led to a paradigm of rituals, in place of meaningful thought.
Professor Gigerenzer is director emeritus of the Center for Adaptive Behavior and Cognition at the Max Planck Institute for Human Development in Berlin.
In an effort to maintain the integrity of open discussion, we are asking all viewers to direct their thoughts and comments to the forums and posts on our website: brokenscience....

[@andrewkatz5946]

Happy to see talks like this posted online for others to hear

[@JConabike]

What a nice and clear talk about how Science has been Broken.

[@BCarli1395]

This is a very helpful perspective on the teaching and use of statistics. Prof also gives a clever sociological analysis of statistics (in practice) rather than the usual statistical analysis of a sociological practice, describing p-hacking as a “ritual “. He even brings a little Freudian analysis to the discussion! Thank you, Prof Gigerenzer.

[@letMeSayThatInIrish]

It bothers me no end that an important video like this gets so few views. But they say the youtube suggestion algorithm favours videos with comments, so maybe if I add this comment?

[@harrisoncurtis5016]

This man wrote the paper that this talk is largely based on in 2004!!!

[@tobykelsey4459]

Apart from Gigerenzer's 2004 paper with the same title, what are the best resources or publications to learn more about this topic?

[@gnoelalexmay]

I'm afraid I don't have a good answer, (I'm only 25 minutes in, so may be misunderstanding your request), but a place to start might be the work of John Ionnidis. Also, Kary Mullis had some very interesting views on the way modern science was done and how scientific consensus was being protected. I hope this helps 🙏 (also to boost the video in the algorithm!)

[@ckashino]

This is made exponentially worse by “biostatisticians” who train outside of any contextual field who mindlessly report to the untrained clinicians what they “found”.

[@ungarlinski7965]

Gosh, this is just second-semester statistics class at the undergrad level. And I mean for non-stats majors. Really easy stuff. Goes to show how academic papers are not really read. Bizarre that this misuse goes on. Otoh, in the 90s there are stats that say over 90% of college students cheat at least once. With the internet now it's worse. So perhaps not too surprising after all. It's just the effect of high stakes stuff. And then instructors have no power to discipline students ... and the cycle continues.