Lightnews — Scholar-powered news

Jan Vanhove

@janhove.bsky.social

56 followers 54 following 37 posts

Senior lecturer at the Department of Multilingualism at https://unifr.ch/. Master's student statistics at https://unibe.ch/. Saxophonist with https://beatmoustache.ch/. Aficionado of Danish ska. Blog & website: https://janhove.github.io

unifr.ch

Posts Media Videos Starter Packs

Pinned

Jan Vanhove @janhove.bsky.social · Sep 12

Revised and translated to English: The course materials for 'Introduction to quantitative data analysis'. Available from github.com/janhove/Anal....

Jan Vanhove @janhove.bsky.social · 6d

New blog post: Clarifying research questions by sketching possible outcomes

janhove.github.io/posts/2025-1...

1 2

Jan Vanhove @janhove.bsky.social · 6d

High-level whining: Replace your osf.io view-only links once your paper's accepted.

Jan Vanhove @janhove.bsky.social · 8d

For the record, I've got dibs on "The Reverend Bass and His Flat Priors".

Jason Moore @moorejh.bsky.social · 8d

Jan Vanhove @janhove.bsky.social · 9d

Otto in het ootje.

Jan Vanhove @janhove.bsky.social · 9d

Schwestersprachstreich.

1 1

Reposted by Jan Vanhove

Jo Wolff @jowolff.bsky.social · 11d

I wouldn’t normally endorse AI prompts but these are indeed essential for all academics.

profserious @profserious.bsky.social · 12d

Here are your 10 -essential- AI prompts for academics ... make your life easy with help from @profserious.bsky.social profserious.substack.com/p/10-ai-prom...

10 AI Prompts for Academics

making those hard jobs a little easier ...

profserious.substack.com

6 37 100

Jan Vanhove @janhove.bsky.social · 15d

Whether I did a good job is not for me to say. But chapters 9 and 10 in github.com/janhove/Anal... The article by Schad et al. is my go-to reference.

github.com

Jan Vanhove @janhove.bsky.social · 16d

that's nice to read, thanks :)

Jan Vanhove @janhove.bsky.social · 27d

Kötbullar, Hoegaarden

Jan Vanhove @janhove.bsky.social · Sep 13

"It is well known that horses don't require dentistry (Rohlfing 2025)."

Ingo Rohlfing @ingorohlfing.bsky.social · Sep 13

Don't look a gift horse in the mouth.

Paul Hünermund @p-hunermund.com · Sep 13

Every now and then, I make the mistake of checking what people are citing my papers for. And without fail, it's a fast track to disappointment. 😭

2 3

Jan Vanhove @janhove.bsky.social · Sep 12

Revised and translated to English: The course materials for 'Introduction to quantitative data analysis'. Available from github.com/janhove/Anal....

Jan Vanhove @janhove.bsky.social · Sep 12

I've updated the lecture notes for the class on Quantitative Methodology - available from github.com/janhove/Quan....

Jan Vanhove @janhove.bsky.social · Sep 7

That doesn't have to be due to AI, though. I've seen published studies from years ago that do this. (It's true that for *constant* total n, balanced groups have somewhat more power. But throwing away data to achieve equal groups is of course silly.)

Jan Vanhove @janhove.bsky.social · Feb 23

The NZ badminton team used to be called... well, you'll look it up.

1 1

Jan Vanhove @janhove.bsky.social · Feb 16

"The (perhaps surprising) answer is that under any null hypothesis, the p-values are uniformly distributed: all p-values between 0 and 1 are equally likely." This is incorrect It's a good enough approximation for reasonably continuous data and the standard tests, but it's not true in general.

1 1

Reposted by Jan Vanhove

JAMES HEATHERS @jamesheathers.bsky.social · Feb 14

Friends, I have written you a book on forensic metascience.

It is free. You can have it. Happy St. Valentine's Day.

If you wish to give me a gift back, you can use it to cause trouble - the greatest gift of all.

open.substack.com/pub/jamescla...

I Have Written You A Book On Forensic Metascience

Use it to cause trouble

open.substack.com

12 150 360

Jan Vanhove @janhove.bsky.social · Feb 14

thanks, James :)

Jan Vanhove @janhove.bsky.social · Feb 14

"Aspects of the Theory of Forensic Metascience" would be the linguist's way out ;)

1 3

Jan Vanhove @janhove.bsky.social · Feb 6

Sure, but they involve arbitrary choices (bandwidth, kernel used) and the assumption that a density actually exists :)

Jan Vanhove @janhove.bsky.social · Feb 6

In fairness, the most common alternatives are worse. But boxplots and/or violin plots with the actual data points on top are pretty nifty, imo.

1 3

Jan Vanhove @janhove.bsky.social · Feb 6

They are somewhat counterintuitive (boxplots too, btw), but they don't involve any tuning parameters, don't assume that the distribution has a probability density, and it's very easy to construct confidence bands for them (unlike for kernel density estimates). But this just underscores your point ;)

1 2

Jan Vanhove @janhove.bsky.social · Nov 26

(Of course, the best Dutch word is the Belgian Dutch word for turquise.)

Jan Vanhove @janhove.bsky.social · Nov 26

Wait till you ask a Dutch person (not a Fleming!) what they call a pair of jeans.

Reposted by Jan Vanhove

Nick Brown @steamtraen.eu · Nov 14

French wordplay jokes are the best...

1 4

Jan Vanhove @janhove.bsky.social · Nov 4

The approximation becomes better as n -> \infty.