Lightnews — Scholar-powered news

Mattan S. Ben-Shachar

@mattansb.msbstats.info

3.1K followers 330 following 1.2K posts

Statistics lecturer | Freelance statistical consultant & research analyst | #rstats dev @easystats.github.io home.msbstats.info (He/Him)

Posts Media Videos Starter Packs

Pinned

Mattan S. Ben-Shachar @mattansb.msbstats.info · Oct 26

1k followers! Calls for a re-introduction:

Statistics lecturer, freelance stats consultant, & #rstats dev @easystats.bsky.social 📊

I try to help social scientists make better inferences from their data & communicate their findings 👨‍🏫

Hope to bring the #stats twitter/R community vibes over to bsky!

3 2 31

Mattan S. Ben-Shachar @mattansb.msbstats.info · 12h

Regarding your point on search engines: using an old fashioned Search Engine requires levels of abstraction that using LLMs does not - more abstraction => more learning on the user side (this is true in general). Research supports this idea.

Mattan S. Ben-Shachar @mattansb.msbstats.info · 12h

I can't speak about experts using LLMs - I'm an expert in what I do, and I haven't found LLMs to be transformational in my work. Little helpers at best. But I keep hearing others saying it's very helpful and useful, and I see no reason not to believe them (and you).

Mattan S. Ben-Shachar @mattansb.msbstats.info · 12h

I should have said: Interacting with LLMs as a novice is not like interacting with an expert as a novice.

As an educator, I'm interested in how non experts use tools to become experts, and I feel (and research seems to back me up) LLMs are not good tools for this.

Reposted by Mattan S. Ben-Shachar

Paul Hünermund @p-hunermund.com · 16h

AI is a heavy blow to the "writing is thinking" crowd

8 1 10

Mattan S. Ben-Shachar @mattansb.msbstats.info · 13h

#statsmeme

Khoa @khoavuumn.bsky.social · 14h

1 7

Mattan S. Ben-Shachar @mattansb.msbstats.info · 13h

I'm here replying to you, so weird of you to assume I'm not open to evidence that would counter what I'm saying. Suit yourself.

Mattan S. Ben-Shachar @mattansb.msbstats.info · 13h

Chat says yes!!!!

Mattan S. Ben-Shachar @mattansb.msbstats.info · 13h

The value isn't in the anxiety - it's in getting real meanigful feedback about my question from an expert. Something I haven't seen an LLM do well *at all*.

Also anxiety isn't removed (which was my point) - it is alleviated by avoidance, which only leads to more anxiety later.

Mattan S. Ben-Shachar @mattansb.msbstats.info · 14h

The interaction with the LLM hasn't taught me anything about how to ask good questions.

2 1

Mattan S. Ben-Shachar @mattansb.msbstats.info · 14h

I have a question -> I've never asked an expert -> I'll ask the LLM -> LLM always says "great question" which I can tell is BS ->

and the cycle repeats (this is classic "feeding the anxiety" - a great opportunity for research), and I never learn how to ask a great question.

1 2

Mattan S. Ben-Shachar @mattansb.msbstats.info · 14h

Yes, but you're asking a non-human (non-)expert - my guess is that this probably leads to some form of avoidance learning, making one less likely to ask a human expert any questions of fear of them perceived as dumb.

1 1

Reposted by Mattan S. Ben-Shachar

Matti Vuorre @matti.vuorre.com · 19h

Against Publishing: universonline.nl/nieuws/2025/...

Preprints are read, shared, and cited, yet still dismissed as incomplete until blessed by a publisher. I argue that the true measure of scholarship lies in open exchange, not in the industry’s gatekeeping of what counts as published.

6 35 89

Mattan S. Ben-Shachar @mattansb.msbstats.info · 14h

E.g. of an opportunity robbed by LLM use

Tim Kellogg @timkellogg.me · 14h

by not being afraid of asking dumb questions

1 1

Mattan S. Ben-Shachar @mattansb.msbstats.info · 17h

These survivors also don't realize they are living survivor bias, and will often down play the role of luck or how much rejections sting.

(Or my favorite: claim the peer review always makes the ms better... this is total BS)

1 1 2

Mattan S. Ben-Shachar @mattansb.msbstats.info · 22h

The jokes write themselves (but not literally)

Mattan S. Ben-Shachar @mattansb.msbstats.info · 3d

"Productivity is through the roof!"

Mattan S. Ben-Shachar @mattansb.msbstats.info · 22h

Finally, a science reform we can all endorse!

Nick Brown @steamtraen.eu · 1d

Next time an institution tells you how seriously it takes research misconduct, ask them if it's *this* seriously. www.bmj.com/content/297/...

In 1916 the BMJ published an article about the work done by James Shearer, an American physician working in the British Army as a sergeant (because he had no British qualification). He had described a
"delineator" which was better than x rays for portraying gunshot wounds. This caused a sensation and a lot of interest — but on investigation the work was found to have been invented. The BMJ published a retraction, but Shearer was tried by court martial and sentenced to death by firing squad.

2 3

Reposted by Mattan S. Ben-Shachar

Nick Brown @steamtraen.eu · 1d

Next time an institution tells you how seriously it takes research misconduct, ask them if it's *this* seriously. www.bmj.com/content/297/...

15 210 570

Mattan S. Ben-Shachar @mattansb.msbstats.info · 22h

Ah yes, I'm familiar with the sub-class of methods people that only believe results from simple analysis (what are we hiding in our multiple regressions???)

Mattan S. Ben-Shachar @mattansb.msbstats.info · 1d

I don't know who Malte is, but if anyone is attributing forethought to what psychologists measure... 😶‍🌫️

1 1

Mattan S. Ben-Shachar @mattansb.msbstats.info · 1d

Oh, I thought it was okay for any even number of measures!

Mattan S. Ben-Shachar @mattansb.msbstats.info · 1d

Wow super useful!

Mattan S. Ben-Shachar @mattansb.msbstats.info · 1d

But surely if I have *3* measures, you wouldn't...

1 1

Mattan S. Ben-Shachar @mattansb.msbstats.info · 1d

#statsmeme

1 3 23

Reposted by Mattan S. Ben-Shachar

Julia M. Rohrer @dingdingpeng.the100.ci · 1d

New blog post!

Let's say you have two measures meant to capture the same confounder. They're highly correlated. Can you still proceed with your regression analysis?

(I admit, the title is a bit of a spoiler)

www.the100.ci/2025/10/13/i...

If you have two measures of the same confounder, you can just include both of them in your regression model

Sometimes, researchers worry about multicollinearity in situations where it’s actually a non-issue. Here’s one such scenario. Imagine a situation where you are interested in the effect of X on Y (X...

www.the100.ci

9 38 110

Reposted by Mattan S. Ben-Shachar

Joe Bak-Coleman @jbakcoleman.bsky.social · 1d

Standardized effect sizes can mean very different things, depending on what they mean for society more broadly. We probably shouldn't use them.

1 6