Lightnews — Scholar-powered news

Kevin M. King

@kevinmking.bsky.social

6.1K followers 800 following 1.5K posts

He/him. Something something Self control and quant methods. Professor of Psychology at University of Washington. All opinions are my own and correct. Co-host of https://thatimplementationsciencepodcast.podbean.com/ https://faculty.washington.edu/kingkm

thatimplementationsciencepodcast.podbean.com

Posts Media Videos Starter Packs

Kevin M. King @kevinmking.bsky.social · 4h

But for real, what did people do before that?

What do they do in fields where they don't care about rigor, like Social Psychology?

1 1

Kevin M. King @kevinmking.bsky.social · 4h

1 1

Reposted by Kevin M. King

Katie Wilson for Seattle @wilsonforseattle.bsky.social · 13h

We are facing an affordability crisis in our city that is worse than any that we have ever seen before. More people have died outside in the last four years than in any previous term. It doesn’t have to be this way.

2 18 88

Kevin M. King @kevinmking.bsky.social · 12h

I'm curious about data harmonization methods. What do folks use besides MNLFA? Z scores? I use POMP but it's not super popular. Any other methods?

Kevin M. King @kevinmking.bsky.social · 13h

Nice work on the bagels, those look solid!

Reposted by Kevin M. King

Jean-François Bonnefon @jfbonnefon.bsky.social · 18h

Data available upon request

Marjorie James Keenan @unenthusiast.com · 1d

In honour of spooky month, share a 4 word horror story that only someone in your profession would understand.

rm -rf ~/

hammancheez @hammancheez.bsky.social · 1d

"The chancellor approved it"

1 4 28

Reposted by Kevin M. King

Ed Zitron @edzitron.com · 2d

Premium: The AI Bubble's promises are impossible. NVIDIA's customers are running out of money, GPUs die in 3-5 years, most 1GW data centers will never get built, and OpenAI's Abilene data center doesn't won't have the power it needs before 2028 - if it ever does.
www.wheresyoured.at/the-ai-bubbl...

The AI Bubble's Impossible Promises

Readers: I’ve done a very generous “free” portion of this newsletter, but I do recommend paying for premium to get the in-depth analysis underpinning the intro. That being said, I want as many people ...

www.wheresyoured.at

43 340 1.4K

Kevin M. King @kevinmking.bsky.social · 1d

Give America Ebola again.

Joshua Weitz @joshuasweitz.bsky.social · 1d

"Roughly 70 Epidemic Intelligence Service officers — the so-called “disease detectives” who respond to outbreaks around the globe — received layoff notices"

There is no 'make healthy' objective here, but there will be costs and consequences, to readiness, safety, and lives.

Sheryl Gay Stolberg @sherylnyt.bsky.social · 2d

FULL STORY HERE: Emails began flooding CDC inboxes late Friday night announcing dozens of layoffs. HR people brought back from furlough to oversee the RIFs. Disease detectives, entire Washington office, global health, injury prevention, MMWR staff among those hit.
www.nytimes.com/2025/10/11/u...

Reposted by Kevin M. King

Catherine Rampell @crampell.bsky.social · 1d

Per CDC source: At internal leadership meeting this afternoon, it was shared that 1,257 people were RIF'd at CDC. Number may not be final. Apparently, chief of staff was unaware that RIFs were going to occur so not clear who is making decisions. (Short 🧵)

91 1.6K 3.4K

Reposted by Kevin M. King

Jed Brown @jedbrown.org · Aug 18

Needless to say, an LLM cannot reliably determine whether a source supports a given claim, just as it cannot determine whether a claim is true. This will automate generation of documents that make false claims, with citations that exist and may, on superficial reading, support those false claims.

1 7 78

Reposted by Kevin M. King

Virginia Gewin @virginiagewin.bsky.social · 2d

Journalist here

I’m interested in talking to a federal agency scientist who was fired, then rehired. I can keep you anonymous. I’m on Signal ginnyg.04

Reposts are appreciated!

4 310 310

Reposted by Kevin M. King

Colette Delawalla @cdelawalla.bsky.social · 2d

RIF UPDATE:

We @standupforscience.bsky.social are in communication with our network. So far, we have not heard of any actual RIFs, only Russ’ tweet.

To press—we will send a release and facilitate interviews if we do hear abt them taking place.

1 27 90

Kevin M. King @kevinmking.bsky.social · 2d

Those fuckers.

Sheryl Gay Stolberg @sherylnyt.bsky.social · 2d

BREAKING: Friday night massacre underway at CDC. Doznes of "disease detectives," high-level scientists, entire Washington staff and editors of the MMWR (Morbidity and Mortality Weekly Report) have all been RIFed and received the following notice:

1 3

Kevin M. King @kevinmking.bsky.social · 2d

Totally. Seattle is spoiled by the constant winning ways of the Sounders. They're just not used to losing.

Kevin M. King @kevinmking.bsky.social · 2d

Your hands probably died waiting for the Bayesian model to converge.

1 3

Kevin M. King @kevinmking.bsky.social · 2d

It turns out the imputation does matter, BUT the method (Bayesian versus frequentist) doesn't. Unfortunately for me and the months I spent running the brms code!

1 2

Kevin M. King @kevinmking.bsky.social · 2d

@charlymarie.bsky.social , you were right. I'd messed up the plot code.

1 4

Kevin M. King @kevinmking.bsky.social · 2d

Reposting to show how I'm a moron, but differently than I thought.

Imputation does change things (estimates, but interestingly not really CIs). And the choice of glmm versus Bayesian doesn't matter tremendously (except that the latter takes orders of magnitude longer to run).

2 1 8

Kevin M. King @kevinmking.bsky.social · 3d

But currently triple checking this because maybe something is wrong.

1 2

Kevin M. King @kevinmking.bsky.social · 3d

Yeah I noticed that and checked.

It's a big GLMM with something like 9K observations from ~430 people when not accounting for missing, and 15K observations from ~500 in the imputed data. So it's possible the differences in CIs are just quite small.

1 1

Kevin M. King @kevinmking.bsky.social · 4d

Solomon we're in full on nihilism mode tonight. Won't you join us???

a cartoon of homer simpson and bart simpson sitting on a couch with the words join us above them .

ALT: a cartoon of homer simpson and bart simpson sitting on a couch with the words join us above them .

media.tenor.com

Kevin M. King @kevinmking.bsky.social · 4d

Especially when the effort does not produce a noticeable improvement.

It would be like if someone showed that MLR handled zero-inflated count data as well as a proper model. I'd just quit and go farm sheep.

Kevin M. King @kevinmking.bsky.social · 4d

I think this all just shows us that quant psych folks lie.

Kevin M. King @kevinmking.bsky.social · 4d

Months, Andrea. I mean, much of that and as avoidance but still

1 1

Kevin M. King @kevinmking.bsky.social · 4d

(but thank you for the ggtip() )