Lightnews — Scholar-powered news

Reposted by Robin Kok

Julia M. Rohrer

@dingdingpeng.the100.ci

Ever stared at a table of regression coefficients & wondered what you're doing with your life?

Very excited to share this gentle introduction to another way of making sense of statistical models (w @vincentab.bsky.social)
Preprint: doi.org/10.31234/osf...
Website: j-rohrer.github.io/marginal-psy...

Models as Prediction Machines: How to Convert Confusing Coefficients into Clear Quantities

Abstract
Psychological researchers usually make sense of regression models by interpreting coefficient estimates directly. This works well enough for simple linear models, but is more challenging for more complex models with, for example, categorical variables, interactions, non-linearities, and hierarchical structures. Here, we introduce an alternative approach to making sense of statistical models. The central idea is to abstract away from the mechanics of estimation, and to treat models as “counterfactual prediction machines,” which are subsequently queried to estimate quantities and conduct tests that matter substantively. This workflow is model-agnostic; it can be applied in a consistent fashion to draw causal or descriptive inference from a wide range of models. We illustrate how to implement this workflow with the marginaleffects package, which supports over 100 different classes of models in R and Python, and present two worked examples. These examples show how the workflow can be applied across designs (e.g., observational study, randomized experiment) to answer different research questions (e.g., associations, causal effects, effect heterogeneity) while facing various challenges (e.g., controlling for confounders in a flexible manner, modelling ordinal outcomes, and interpreting non-linear models).

Figure illustrating model predictions. On the X-axis the predictor, annual gross income in Euro. On the Y-axis the outcome, predicted life satisfaction. A solid line marks the curve of predictions on which individual data points are marked as model-implied outcomes at incomes of interest. Comparing two such predictions gives us a comparison. We can also fit a tangent to the line of predictions, which illustrates the slope at any given point of the curve.

A figure illustrating various ways to include age as a predictor in a model. On the x-axis age (predictor), on the y-axis the outcome (model-implied importance of friends, including confidence intervals).

Illustrated are
1. age as a categorical predictor, resultings in the predictions bouncing around a lot with wide confidence intervals
2. age as a linear predictor, which forces a straight line through the data points that has a very tight confidence band and
3. age splines, which lies somewhere in between as it smoothly follows the data but has more uncertainty than the straight line.

August 25, 2025 at 11:49 AM

Reposted by Robin Kok

Jonathan Peelle

@jpeelle.bsky.social

I was reminded today of the heroic work done by @eikofried.bsky.social and
@robinnkok.bsky.social to see what information Elsevier collects on academics and was re-horrified. 🧵 (1/5)

eiko-fried.com/welcome-to-h...

#academicsky

Welcome to Hotel Elsevier: you can check-out any time you like … not » Eiko Fried

A journey by Robin Kok and Eiko Fried trying to understand what private data Elsevier collects; what private data Elsevier sells; and what to do about it.

eiko-fried.com

July 23, 2025 at 2:57 PM

Robin Kok

@robinnkok.bsky.social

This @xkcd.com hurts a little

July 21, 2025 at 7:24 PM

Robin Kok

@robinnkok.bsky.social

"As fax machines were phased out in the late 2010s, engineers desperately sought new ways to annoy, confuse and enrage users. Microsoft SharePoint turned out to be the ideal replacement" - History of Technology, 3rd ed, 2047.

July 17, 2025 at 2:06 PM

Robin Kok

@robinnkok.bsky.social

For all this talk about U-shaped associations, I have found the optimum amount of yoghurt in your beard is 0.

July 5, 2025 at 11:21 AM

Robin Kok

@robinnkok.bsky.social

LA YARRRRR

simine vazire @simine.com · Jul 2

checks out

Google translate translating “p-hacking” to “p-piratage”

July 3, 2025 at 4:56 AM

Reposted by Robin Kok

Avram Holmes

@avramholmes.bsky.social

The entire field of psychology from 1967 until 2011, and if we are honest with ourselves most of our current research output.

June 30, 2025 at 3:33 PM

Robin Kok

@robinnkok.bsky.social

Generative AI is basically Orwell's "prolefeed" on steroids

#showerthoughts

June 29, 2025 at 8:22 AM

Reposted by Robin Kok

Mark A. Hanson

@hansonmark.bsky.social

New from The Strain Team:
🎊 Springer Nature Discovers MDPI 🎊

Springer Nature has spawned a copycat journal series called "Discover" mimicking #MDPI journal titles and citation behaviours. We even made a browser game to prove it (see 🔗).

Gross! 😀 1/n

#ResearchIntegrity #SciPub #AcademicSky

Spinger Nature Discovers MDPI – The Strain on Scientific Publishing

Home page for the paper ‘The Strain on Scientific Publishing’ by Mark A Hanson, Dan Brockington, Paolo Crosetto and Pablo Gomez Barreiro

the-strain-on-scientific-publishing.github.io

June 10, 2025 at 5:49 AM

Robin Kok

@robinnkok.bsky.social

Happy weekend!

May 25, 2025 at 7:44 AM

Reposted by Robin Kok

Graham Kendall

@grahamkendall.bsky.social

In a recent paper of mine, I suggested that the article processing charges (APCs) that were paid should be shown on the aricle, and it shoud also be stated which author(s) paid the APC.
🧵

May 23, 2025 at 3:14 AM

Reposted by Robin Kok

Nick Brown

@steamtraen.eu

May 21, 2025 at 8:36 AM

Reposted by Robin Kok

Tim Kaiser

@tkaiser.science

This is also a good time to remind people that, according to the best available evidence, LLM-based chatbot therapy is about as effective as unguided self-help and might not even work for symptoms other than depression.

May 20, 2025 at 3:21 PM

Reposted by Robin Kok

Malte Elson

@malte.the100.ci

May 17, 2025 at 6:01 PM

Reposted by Robin Kok

Gary Collins

@gscollins.bsky.social

NEW PAPER: Updated SPIRIT-2025 for reporting protocols of randomised trials is now available in the @BMJ.com, @JAMA.com, @thelancet.bsky.social, @PLoS.org, and @natmedicine.bsky.social.

—> www.bmj.com/content/389/...

#openscience #transparency #medsky #statssky #episky

April 29, 2025 at 5:59 AM

Robin Kok

@robinnkok.bsky.social

April 21, 2025 at 5:29 PM

Robin Kok

@robinnkok.bsky.social

#academicchatter

April 3, 2025 at 11:16 AM

Robin Kok

@robinnkok.bsky.social

I'm not distracted, I just have a hyperfocus on everything

March 29, 2025 at 10:17 AM

Reposted by Robin Kok

Gordon Hodson

@gordonhodsonphd.bsky.social

#AcademicSky #AcademicPublishing #AcademicChatter

Opening up more journals because we produce too much output that no-one reads or needs

cartoon: driver in 2 lane highway wants road widened, pleased when it's 3 lanes (but it's just as congested)

March 26, 2025 at 5:04 PM

Robin Kok

@robinnkok.bsky.social

Confuse, bewilder and alarm your Gen Z coworkers by citing Sabrina Carpenter lyrics in your meetings.

They: "That meeting was a mess, I have no idea how to follow up"
Me: "You're confused and I'm upset but we never talk about it"

March 19, 2025 at 8:02 AM

Reposted by Robin Kok

Masud Husain

@masudhusain.bsky.social

Why academia is sleepwalking into self-destruction. My editorial @brain1878.bsky.social If you agree with the sentiments please repost. It's important for all our sakes to stop the madness
academic.oup.com/brain/articl...

March 6, 2025 at 7:16 PM

Robin Kok

@robinnkok.bsky.social

The journal that published this AI-generated dross has an impact factor of 6.9 and charges $3450 per paper.

Cheshire @thatsregrettab1.bsky.social · Mar 3

Prov/active paper is retracted by Elsevier’s Process SAFFEETY and Environmental Protection after PPPR caught issues that were not identicated or preventiived during the original editorial assessation and inplematiion. h/t anon Hoya.
pubpeer.com/publications...