Share this page!

Enter URL or ID to Unroll

You can paste full URL like: https://x.com/threadreaderapp/status/1644127596119195649
or just the ID like: 1644127596119195649

How to get URL link on X (Twitter) App

  1. On the Twitter thread, click on or icon on the bottom
  2. Click again on or Share Via icon
  3. Click on Copy Link to Tweet
  4. Paste it above and click "Unroll Thread"!
  5. More info at Twitter Help
Nïck Brown🌻 Profile picture
Nïck Brown🌻
@sTeamTraen
🏴󠁧󠁢󠁥󠁮󠁧󠁿🇮🇪🇪🇺 in 🇪🇸. PhD in psychology & self-appointed data police cadet. Interested in the lower tail of many distributions. Not yet disabled.
3 subscribers
Jan 12, 2024 • 11 tweets • 6 min read
Have we had this one yet?
Left: Oxman (2011), Virtual & Physical Prototyping 6(1), 3–31.
Right: Suresh (2001), Science 292(5526), 2447–2451. doi.org/10.1080/174527…
doi.org/10.1126/scienc…
Image Amusingly, Oxman's 2011 article (left) was then apparently plagiarised in this 2013 thesis proposal (right). researchgate.net/publication/23…

Image
Image
Sep 13, 2023 • 7 tweets • 2 min read
You know how when there's a big mystery, people speculate earnestly for years, and it turns out to have been something banal?

It's going to be fun* when we realise that the reproducibility crisis is in large part down to fraud and we look back at the convoluted explanations. /1 We have known since Wolins (1962) that less than half of psychologists share their data when asked. This is a robustly reproducible result (ho ho).

Ask your non-scientist friends why they think that might be. See how many don't say that the authors might be hiding something. /2
Image
Image
Nov 29, 2022 • 9 tweets • 3 min read
Seven years after @JamesHeathers and I first started to write about it, this article by Nicolas Guéguen has been retracted today. 🥂🍾🥳🎉

/1 I hope that the copyright people at Taylor and Francis will not mind if I post some of the "highlights" from this article here.

/2
Sep 7, 2022 • 16 tweets • 3 min read
In case anyone is interested, there is an absolutely wild story of alleged cheating at the very top of the chess world right now. /1 theguardian.com/sport/2022/sep… On Sunday 4 September, world champion Magnus Carlsen lost to 19yo US grandmaster Hans Moke Niemann in a tournament. /2
Sep 7, 2022 • 6 tweets • 2 min read
The hydroxychloroquine article by Gautret et al. sciencedirect.com/science/articl… has 18 authors. For the better part of a year every one of them, even if their names were only on the paper as a courtesy, has been fully aware of the detailed allegations of faking. /1 The French investigative site @Mediapart basically described exactly what the latest French government report has confirmed back in November 2021. /2 mediapart.fr/en/journal/fra…
Sep 6, 2022 • 4 tweets • 2 min read
French government inquiry confirms (as reported earlier by investigative media) that the initial IHU-Marseille hydroxychloroquine study was faked, with different criteria for +ve/-ve PCR tests being applied to the two groups. This is straight-up scientific fraud. I would like to post a translation via DeepL, but I can't copy the text from the PDF due to some or other document security feature. So maybe you could ask a French-speaking friend if necessary.
Feb 15, 2022 • 10 tweets • 4 min read
Another comment strikes a blow to the French "belief drives #LongCovid" study in JAMA Internal Medicine. The authors knew, but somehow failed to mention, that participants with positive serology tests had been told that their results were unreliable. /1 jamanetwork.com/journals/jamai… Image The authors—and, even more so, the media outlets that covered the story—used the fact that people who "believed" that they had had Covid were more likely to report LC symptoms than those with positive serology tests, to insinuate that people with LC never actually had Covid. /2
Dec 28, 2021 • 46 tweets • 9 min read
With @JulieOudet’s permission, I have tried to translate her astonishing thread. All errors are mine, including the mix of UK and US slang & concepts. I start each translated tweet with quote marks (“) to emphasise that the words are not mine. My comments are in [brackets]. /0 “The last few moments of 2 days off. I have a knot in my stomach at the thought of going back to work. So here’s a random thread that nobody is forcing you to read: /1
Dec 6, 2021 • 22 tweets • 6 min read
I can't get very excited about the results from Merck's "Protocol 002" study of Molnupiravir. (thread)

It's the second document here: fda.gov/advisory-commi… /1 Here's the main result. Fewer all-cause hospitalizations or deaths in the treatment condition than with placebo, p=0.0218. /2
Nov 21, 2021 • 31 tweets • 16 min read
I love international borders; the less noticeable, the better. So here is a thread of some remarkably unremarkable frontier crossings from Google Earth. /1 After all, a nation state is merely, in Max Weber's words, "a human community that (successfully) claims the monopoly of the legitimate use of physical force within a given territory", which is silly and boring. /2
Aug 21, 2021 • 4 tweets • 1 min read
Is there a political party, anywhere in the EU, whose programme includes a call to abolish existing government-issued ID cards? /1 I know the various Pirate parties are skeptical of government use of information about citizens, but I don't think even they are calling for the abolition of ID cards. /2
Jun 18, 2021 • 14 tweets • 3 min read
It occurs to me that as written, Baron & Kenny's IV-M-DV mediation model, combined with a sample size of 10,000 (available in any decent public dataset) enables you to show that literally anything "partially mediates" the relation between literally any two other variables. /1 Step 1: Does IV "significantly" predict DV? Almost certainly, because IV and DV will be correlated at least .02 by chance, and r(9,998)=.02 has p=.046. /2
Jun 17, 2021 • 7 tweets • 2 min read
I'm testing a new home-made captcha system. It doesn't have to be especially robust because it will only be used on one site. On the next 3 tweets are examples of the modes that it works in (the user sees one chosen at random). There is a poll at the end. /1 Version 1

/2 Image
Mar 16, 2021 • 6 tweets • 1 min read
If it turns out that the AZ vaccine has a real deleterious effect, we're probably finally going to have to face up to the rather dishonest way in which we have elided "minimal risk" and "zero risk" for many years in vaccine discussions. /1 It is not unreasonable to expect that a complex product like a vaccine might interact with certain factors in a small number of people to produce undesirable effects, including death. /2
Mar 11, 2021 • 12 tweets • 4 min read
Sometimes an advertising slogan goes way beyond its original purpose and ends up in the language for years or decades, either as-is or as a source of riffing. Wendy's "Where's the beef?" dominated a US election. /1 "Does exactly what is says on the tin" is a standard form of approval in the UK, and in France, a whole generation of younger people doesn't know that the adverb "vachement" ("extremely") comes from "vachement bon" in ads for La Vache Qui Rit processed cheese. /2
Mar 10, 2021 • 9 tweets • 3 min read
Do you ever mark up PDF files in Adobe Reader? Did you know that whether you write a text comment, add suggested copy-edits, or simply highlight a word, your Adobe username is attached to every annotation? Might that not be what you want in every case? Read on... /1 Let's imagine that I'm reading a PDF that has been e-mailed to me at some point, and a couple of my esteemed colleagues read it at some earlier point. I might find comments like this. /2
Mar 9, 2021 • 5 tweets • 2 min read
Everyday stuff I wish psychologists could explain, part 94: Why people believe in "secret burglar marks on the street outside your house". /1 Image If I've understood correctly, a burglar uses these signs to communicates to all the other burglars in the neighbourhood that he's done all the reconnaissance work, but doesn't mind if a different burglar then waltzes in and steals all the stuff. /2
Mar 5, 2021 • 4 tweets • 1 min read
Why I posted this poll:

I just installed R-4.0.4. On 4.0.3 I had 783 packages. After upgrading them all I had over 800 (some have spawned new dependencies).

I didn't recognise 95% of the names, so I decided to see how many I use. /1 I searched every R source file that I've run in the past 4 years for library() calls. There were 36 distinct packages. So I reset my library to zero and told R to install those 36. Including dependencies, I ended up with 610. /2
Feb 27, 2021 • 13 tweets • 6 min read
A thread about the "Proxalutamide prevents hospitalisation and death in male COVID-19 patients" study that has been attracting some attention. /1 The study protocol and analysis plan is here [PDF]: clinicaltrials.gov/ProvidedDocs/2…
You can also read about the results at the trial registration page, clinicaltrials.gov/ct2/show/study…, in the "Study Results" tab. /2
Dec 25, 2020 • 14 tweets • 5 min read
A thread on the UK's new "points-based immigration system".

Spoiler: It's not very points-based. /1 This page gov.uk/government/pub… is an "accessible version" of the description of the scheme. Presumably an incomprehensible version is available for masochists. /2
Dec 18, 2020 • 19 tweets • 6 min read
A recent Brexiteer theme is how the UK can manage just fiene without food products from the EU. Just search for mentions of "Somerset Brie" and you'll find dozens of pompous claims from true patriots who eat no other cheese. But what do we know about Somerset Brie?

A thread. /1 A quick Google reveals that Somerset Brie is made by a company called Lubborn Cheese Limited. Their address is 1 Manor Farm Cottages, Cricket St Thomas, Chard. Could they be more English? Let's check their web site, shall we? /2