The three* most beautiful words in English:

"I don't know"

* four

Can confirm.

"There’s just too many factors impacting on the proportion of p values that will be between 0.01 and 0.05 to assume that variations in it are either an improvement or a worsening in research practices."

Blog post by @freerangestats.info on "fragile" p values.

#MetaSci

<< Overheard in Waitrose >>

Nice title.

✔️

🎻

small error bars (usually a good thing) revealed regression-to-the-mean (usually a bad thing)

in theory I agree with you!

in practice i believe you'll need a lot of highly-motivated scientists prepared to give up science & become journal-creators & political campaigners to turn this juggernaut around.

(this scientist will not join that fight!)

OPEN SCIENCE & ITS ENEMIES, PART III

(with apologies to apostrophe pedant's for the typo in the first post)

theerrorbar.com?e=41

the error bar Podcast · [object Object] · fact-checking brain science news and asking brain scientists about mistakes and controversies

Open Science and It's Enemies.

Part III: The progressives

A-log (podcast) link here:

B-log (text) in next post 👇

open.spotify.com/show/3kQAgTp...

Writing...

Yes, it really does take training to over-ride our incorrect intuitions👍

👇

bsky.app/profile/soap...

but not spelling

Oh wow, thank you! 🙏🙏🙏

(I also teach resesrch methods 🤓)

Part II of my epic criticism of - some parts of - the #openscience / scientific #reform movement. Episode #40

"The populists" (& the podcasters; but not the meek)

Full text transcript, links & references (b-log):
theerrorbar.com?e=40

Podcast audio (a-log):
podcasters.spotify.com/pod/show/the...

New episode coming some week soon...

...has been a long few months 😴

tell me about it

latest episode. Part I of a III-part mini-series on chronic problems in the Open Science movement:

spotifyanchor-web.app.link/e/g4Qa0mBeLJb

THIS 👇

Three latest episodes from the error bar podcast Just die already, 55 plays Your sample is small, 66 plays Don't look back, 88 plays The series suggests a nicely consistent decline in listens since i changed the format, but what will be next month's listenership..?

next month i will have...

A) 44 plays

B) 49.5 plays

C) 10^9 plays coz it will be epic

🧐

🤩

JUST DIE ALREADY

JUST DIE ALREADY

S03E03 #38

in this episode we find out how motivational talk makes you run faster & i ponder on the history of science & the death of [bad] ideas long, long after their time

listen / subscribe: spotifyanchor-web.app.link/e/GtHtUY3dyIb

read: theerrorbar.com?e=38

YOUR SAMPLE IS SMALL

Episode #37 asks whether it's ever really helpful to criticise a study for having a 'small sample'

It also reveals the final truth about men's & women's brains.

Listen podcast

podcasters.spotify.com/pod/show/the...

hello!

THIS is the error bar, a podcast about neuroscience & statistics which focusses on stories that reach the UK news media. the account is owned by @thehandlab.bsky.social

(i've escaped from twitter.)

the latest podcast, Season 03, Episode 01: Don't Look Back is out!

theerrorbar.com?e=36