Roar Stovner

[133] Heterofriendly: The Intuition for Why You Always Need Robust Standard Errors - Data Colada When I taught my first PhD-level methods course, I invited students to submit questions about any topic in statistics or methodology. Six out of 10 students asked about the same topic: robust & cluste...

Answering the most popular question in a PhD methods course
datacolada.org/133

Both crosslink and ripsed look great. Ripsed meant for agents to do search and replace?

Digital products and services are getting worse – but the trend can be reversed A new report from the Norwegian Consumer Council shows how enshittification affects both consumers and society at large. Luckily, it‘s possible to turn the tide.

We just published a new report on enshittification. Pretty proud of this one!

www.forbrukerradet.no/news-in-engl...

Ah. With compaction or with a clean context? I start from a fresh context; my point was that I let Claude write a summary of what it needs after /clear.

Do you use Chainlink? I make Claude write handoff notes to chainlink. "Nearing compaction. Write handoff notes about everything you need to start from a fresh context." Then I tell the new Claude to read the handoff-notes and it will get all the context it needs.

It easily got a star from me. 🌟 Chainlink is good; it's almost hard imagining employees at Anthropic not using it. 🤷‍♂️

Very much worth reading from @jessicahullman.bsky.social. Inspiring and frightening about the possible uses of language models in science.

This sounds like exactly the kind of thing I would like to learn more about, technical but useful knowledge of R internals. Where would you recommend a novice learn about such things?

Just found this old post of @matti.vuorre.com. I have a large library and the Quarto citation picker in visual mode in VS code is so laggy. This is almost instant! Highly recommended if you already use this combo of tools

Min tiltaksliste:
1. bare fagfellevurder for diamond open access eller typiske "university presses"
2. skill tydelig mellom gode tidsskrift og prestisjefulle tidsskrift. Gode tidsskrift er de med best editorial policies, ikke de som publiserer den beste forskningen.

Andre tiltak for enkeltforskere?

At last an AI tool I can get behind

“Upload an architectural render. Get back what it'll actually look like on a random Tuesday in November.”

antirender.com

What would happen with the fertility rates the year everyone decided to wait until they're 31? Such thought experiments might give intuition, and they might illuminate how the formulas work.

Hey, I'm in mathematics education, so I might have some insight although I can't define these fertility rates. One heuristic is that imagining extremes is useful. For example, imagine a society where all women have one child and all give birth when they're 30. 1/

This whole ‘LLMs/GenAI are inherently sycophantic and will only give you what you want to hear and hence it cannot be used for retrieval meme is really oversold and librarians need to be careful not to fall for it (1)

The Iowa Gambling Task is an extreme example of Jingle Fallacy and schmeasurement.

In 100 articles we found 244 different ways of scoring it, 177 were never reused. Correlations between them range -.99 to .99.

At the same time, we show meta-analyses combine these results as if they’re equivalent.

Zotero 8 is out in stable: www.zotero.org/blog/zotero-8/
A bunch of new stuff -- the new citation dialog especially is a huge (and long overdue) improvement; massively speeds up anything I do in Word/GDocs with Zotero

With data from >100k special issues, we set out to tackle this problem.

Guest editors publishing in their own issues is a form of "endogeny". In this study, we define special issues with extreme endogeny as Published In Support of Self (PISS).

So, how much of this literature is PISS?

4/n

An email from Martin Peterson to university administrators.

Martin Peterson's creative response to being banned from teaching Plato (shared with his permission).

How do you provide the pdf and instructions as a sydtem prompt? Does it require API-usage? I can't find a way to change the system prompt with any other interface, and it seems tedious to make an API call every time.

So
#MathsToday
I found out that if you have four adjacent Fibonacci numbers e,f,g,h then eh and 2fg are two elements of a Pythagorean triple.
1,1,2,3 - 3 and 4
1,2,3,5 - 5 and 12
2,3,5,8 - 16 and 30

A new paper by George Borjas—who served this past year in the Trump White House designing some of its anti-immigration policies—claims to display evidence of ideological bias among researchers who study immigration.

doi.org/10.1126/scia...

🧵 Thread—>

International Criminal Court to ditch Microsoft Office for European open source alternative | Euractiv The court will move its internal work environment to Open Desk, a German-developed open source software

love this for you ICC

www.euractiv.com/news/interna...

Enjoyed this presentation of a difficulty of meta-analysing dietary effects.

The same argument applies to my field of teaching: when considering a new teaching strategy, one must consider what it replaces.

There are astonishing things at the British Museum, but it's the 2000-year-old Roman muffin tin that speaks to me. Humans are humans, through time and across continents.

A long time ago, somebody, somewhere undoubtedly got really pissed because she burned the dinner rolls.

Went from "why would I read the memoir of an academic" to "when can I pre-order the memoir" 🤩

"Pleasure and Desire in Mid Life" by Hadas Weiss Pleasure and Desire in MidlifeIn November of last year, I was bubbly with excitement about my life. By December, I was a weepy, barely functional mess. Nothing changed over that month. The excitement ...

more than you ever wanted to know about my midlife crisis 🫣
www.roifaineantpress.com/post/pleasur...

Line chart of Google search interest for “Christmas gift wife” and “Christmas gift husband” from November 18 to December 24. Bold lines show the average pattern across 2020–2024, while lighter shaded lines represent individual years. Shaded areas highlight which term has higher search interest at each point in time. The chart shows that searches for gifts for wives peak later in December than searches for gifts for husbands.

🎄 Hope you’ve got all your presents ready 💝

Google search interest shows a stable pattern:

🎅 “Christmas gift wife” peaks just before Christmas Eve
🎅 “Christmas gift husband” peaks much earlier

#MerryChristmas to all of you! 🎁

📈 Google Trends (Nov 18–Dec 24, 2020–2024)
#dataviz #ggplot2

I drafted a letter to the editor, please help me out by DMing, commenting, emailing feedback if you are an expert on colliderbias, id obviosuly ad you as a author, Ideally we submit within 24-48 hrs, draft: zenodo.org/records/1800... (click download if the pdf doesnt preview on zenodo)

This Dutch infrastructure grant looks like the future! I recommend clicking through to see all the projects awarded grants. The updates to lavaan will benefit thousands of researchers, but so will all the other grants! I wish more grants were of this kind.

Text reads: About synthetic panels Recruiting the right participants for a study can be difficult. You may not get the exact demographics you need, and the shorter the deadline, the less sure you can be that everyone will answer on time. One possible solution can be to use synthetic panels. Synthetic panels are powered by a first party proprietary AI model developed here at Qualtrics. Our synthetic panel is trained on thousands of responses from a variety of demographic backgrounds in order to more accurately predict how certain populations would respond to a survey. Our synthetic panel is based on the United States General Population, and is only available in English. This panel comes with ready-made quotas and target breakouts in order to represent your chosen population and make it easy to launch your survey right away.

Text reads: Question-writing best practices To get the most reliable and actionable results from synthetic audiences, consider these question-writing best practices: Ask forward-looking and attitudinal questions. Synthetic panels perform best with perceptions, preferences, and intent-based questions. For example, “How likely are you to try…?” Synthetic panels are less applicable for studies on past behaviors, detailed recall, brand recall, or awareness questions. For example, “When did you last visit…?”

Text reads: Discussion The current study aimed to conduct a meta-analysis of the TPB when applied to health behaviours which addressed the limitations of previous reviews by including only prospective tests of behaviour, applying RE meta-analytic procedures, correcting correlations for sampling and measurement error, and hierarchically analysing the effect of behaviour type and sample and methodological moderators. Some 237 tests were identified which examined relations amongst model components. Overall the analysis indicated that the TPB could explain 19.3% of the variance in behaviour and 44.3% of the variance in intention across studies. This level of prediction of behaviour is slightly lower than that of previous meta-analytic reviews which have found between 27% (Armitage & Conner, 2001; Hagger et al., 2002) and 36% (Trafimow et al., 2002) of the variance in behaviour to be explained by intention and PBC.

Did you know that from tomorrow, Qualtrics is offering synthetic panels (AI-generated participants)?

Follow me down a rabbit hole I'm calling "doing science is tough and I'm so busy, can't we just make up participants?"