Together with my co-host @smreeteemehta we've started a new podcast: Nullius in Verba. A podcast about science - what it is, and what it could be. The introduction episode is up now! .Apple: .Spotify:..RSS:.

Here's Selvin and Stuart, in 1966, discussing 'data dredging' in surveys, including 'fishing' and 'hunting' for significance. We've known about alpha inflation for so long! Glad preregistration is finally being more widely adopted!

One thing I like about this paper is that we worked together with some of the people we were criticizing, and who had promoted Type S and M errors. They are friends, which helps, but it was not given we could co-author. The paper is much better and nuanced for it in the end.

Here's Selvin and Stuart, in 1966, discussing 'data dredging' in surveys, including 'fishing' and 'hunting' for significance. We've known about alpha inflation for so long! Glad preregistration is finally being more widely adopted!

New prologue episode of our podcast Nullius in Verba! Smriti reads the first chapter of Bunge's (1959) book Metascientific Queries, which asks 'Why Metascience?

RT @lakens: Prologus 64: Why Metascience? (M. Bunge)

Prologus 64: Why Metascience? (M. Bunge)

J. D. Bernal in 1939 discussing how graduate students (only a thing from 1917 in the UK) are evaluated based on number and bulk, the publication pressure stifling originality, and filling the literature with useless papers.

RT @giladfeldman: Been doing workshops about Peer Community In @PeerCommunityIn / PCI Registered Reports @PCI_RegReports / PCI Psychology &….

Saddened to hear about the death of Prof. Andrea Evers. Such an impressively talented scholar. A great loss for Dutch academia.

We found that the results of 42% of registered studies on the OSF are never publicly shared Most common reasons are null results and bad logistics. Plan your studies carefully (equivalence testing and good time management) to prevent research waste!

In our next episode of Nullius In Verba we will dive into the history of metascience. The first mention of 'metascience' in the literature we could find is Morris, 1938. Do you know of an earlier mention of the term 'metascience'?

RT @mzloteanu: @lakens @dylanarmbruste3 Read the paper, some v convincing points. I saw you mention Kaizer's work (gamma error) and using t….

RT @lakens: New Preprint: Rethinking Type S and M Errors. We argue Type S and M errors are not useful when designing or evaluating studies,….

RT @PadovanRiccardo: Just completed Improving Your Statistical Inferences 🎓.Thanks to @lakens for making such a high‑quality, free course o….

I always found it surprising so many people discussed Type S and M errors as if they are useful, and I hope our paper will make the scientific community rethink the usefulness of these novel statistical concepts, and appreciate the alternative approaches available a bit more!.

We admit Type S and M errors might be referenced in the discussion of badly designed and/or p-hacked studies, and they can be used in education to teach about one-sided tests and selection bias (although critical effect sizes might be more intuitive).

(We also show the BUCSS approach by Anderson et al., 2017, to compute bias adjusted sample sizes, and P-uniform, are basically identical approaches to adjust for bias, something I think no one noticed until now).

We look in the literature, and see people massively misuse Type 2 errors - they try to use it to adjust individual effect estimate. It can't, and was never intended to. Use p-uniform, and you can compute bias adjusted effect sizes.

We analyze the argument that the null is never true, and argue it is incoherent to remove only the value 0 from the space of possible hypotheses. After all, no effect is exactly 0.3 either. If you think the null is not plausible, don't test against it - do a minimum effect test.

What I like about this paper is that the last 2 co-authors - Enrico Toffalini and Gianmarco Altoè - have published papers promoting Type S And M errors, and they made the PRDA R package - they were our internal critics, making sure our points were fair and balanced.