Jack Wilkinson

Is it fine if we call it ‘INSPECT-NRS’ (Non-Randomised Studies) even though it doesn’t really make sense (CT = clinical or controlled trials)

One of the most important rules of stats social media is never post about odds ratios. I apologise for my behaviour and will use this as an opportunity for leaning

Yeah, but am reviewing your draft and that isn’t what you wrote 😜

Whatever your thoughts on relative risks vs odds ratios, I hate the argument that ORs can’t be interpreted as RRs and so shouldn’t be used. Sausages are not doughnuts, and trees are not fish, what’s your point?

Introduction to INSPECT-SR Training Workshop March (Europe) An introductory 2-hour online workshop will introduce participants to the INSPECT-SR tool for assessing trustworthiness of randomised controlled...

And here is a March 6th event, timed for Europe: www.trybooking.com/uk/FZUN

INSPECT-SR Training Workshop April (N America) An introductory 2-hour online workshop will introduce participants to the INSPECT-SR tool for assessing trustworthiness of randomised controlled...

Intro to INSPECT-SR online training workshop - this one has been timed for North America region. April 30th - I *think* this works out as 10am Pacific...but check that. www.trybooking.com/uk/GATM

New paper, on a worrying trend in meta-science: the practice of anonymising datasets on, e.g., published articles. We argue that this is at odds with norms established in research synthesis, explore arguments for anonymisation, provide counterpoints, and demonstrate implications and epistemic costs.

Introduction to INSPECT-SR Training Workshop March (Europe) An introductory 2-hour online workshop will introduce participants to the INSPECT-SR tool for assessing trustworthiness of randomised controlled...

Next free online INSPECT-SR training workshop on March 6th. Register here: www.trybooking.com/uk/FZUN Timed for Europe. Will try to add something for North America in the next few days...

Yeah it’s a tool for assessing trials, not the same as the original

INSPECT-SR: a tool for assessing trustworthiness of randomised controlled trials The integrity of evidence synthesis is threatened by problematic randomised controlled trials (RCTs). These are RCTs where there are serious concerns about the trustworthiness of the data or findings....

This is what we ended up with: www.medrxiv.org/content/10.1...

If you're interested in data sleuthing but aren't sure where to start,

or if you're conducting a systematic review/meta-analysis and want to ensure you're not including junk studies,

check out this Cochrane training session on Trustworthiness Assessment by @jdwilko.bsky.social

Recording now available: www.youtube.com/watch?v=u8vQ...

How serious are the problems? I would now send to the publisher’s research integrity team

Join this grateful bunch, find out how you could use INSPECT-SR in our March session with Jack!

www.cochrane.org/events/inspe...

Pleased to share that we were awarded a 2026 #NHMRC Ideas grant via @sydney.edu.au to study/ raise awareness of published human #CellLines that may not exist. It was a tough year, but this outcome shows that NHMRC can fund topics in #metascience, as well as laboratory & clinical research
#PaperMills

Nice Jamie! Congrats. Really great work, this

RegCheck is one of our group's most exciting projects IMO, and I'm particularly happy to see the codebase now be open scourced!

It’s rewarding to hear from systematic reviewers who have used INSPECT-SR to identify problematic trials (when they have used it correctly!)

I don’t think many of the suggested tests are useful to be honest…in some cases I would question their validity. There is a difference between thoughtful recommendation of reasonable supplementary analysis and a list of every method that has ever been proposed, no?

Was referring to the peer reviewers here.

Trying to work out why all peer reviewers of systematic reviews now suggest including every sensitivity analysis ever devised (do a leave-one-out, do a sequential analysis, do p-curve, how about some outlier detection?) Is the answer ‘they are all now written by LLMs?’

How to read ‘evidence pyramids’ To get past the pointless bit of the arguments

Everyone please red this and then vigorously attack @timpmorris.bsky.social 😜

tpmorris.substack.com/p/how-to-rea...

Running a free two-hour INSPECT-SR workshop tonight 10pm UK time (Thur Jan 15th). Timed for the Australians but may work for some other folks - send me a message in the next couple of hours if you’d like me to add you. INSPECT-SR is a tool to identify problematic (inc. fake) RCTs.

Answer was ‘yes’. Mega. Going to try to make all my lectures as good as Soulwax from now on

I’m off to see Soulwax. Will that be good in 2026?

Some common, fatal flaws in systematic reviews of observational studies - PubMed When evidence from randomized controlled trials about the effectiveness and safety of an intervention is unclear, researchers may choose to review the nonrandomized evidence. All systematic reviews pose considerable challenges, and the level of methodological expertise required to undertake a useful …

I agree - they should have cited mine instead pubmed.ncbi.nlm.nih.gov/38309515/

They didn’t even make the post-hoc reg match what was reported in the review.

This is good. Authors ask an LLM to write a systematic review. The LLM states ‘prospectively registered’ and includes a fake registration number. When queried, authors send a corrected reg number…which has been created after the query. You love to see it. Happy new year!

www.cochrane.org/events/inspe...

One of my best Xmas presents was getting it converted to play Japanese games and a copy of this: