Lightnews — Scholar-powered news

Reposted by Rob Cavanaugh

Maëlle Salmon @masalmon.eu · 15d

"How I, a non-developer, read the tutorial you, a developer, wrote for me, a beginner" by Annie Mueller 😅 😂 😭

anniemueller.com/posts/how-i-...

How I, a non-developer, read the tutorial you, a developer, wrote for me, a beginner - annie's blog

“Hello! I am a developer. Here is my relevant experience: I code in Hoobijag and sometimes jabbernocks and of course ABCDE++++ (but never ABCDE+/^+ are you kidding? ha!)  and I like working with ...

anniemueller.com

15 95 320

Rob Cavanaugh @rbcavanaugh.bsky.social · 21d

So I should just ask students to explain each meme for their stats midterm right?

1 2

Reposted by Rob Cavanaugh

Swathi Kiran@Center for Brain Recovery at BU @swathikiran.bsky.social · Sep 3

Join us Monday, September 15th from 4:00 pm to 5:00 pm ET for a talk by Simona Mancini, Ikerbasque Research Associate Professor / Neurolinguistics and Aphasia group leader at the Basque Center on Cognition Brain and Language.

Register now at https://bit.ly/45LjrF1

3 8

Rob Cavanaugh @rbcavanaugh.bsky.social · Aug 26

fantastic! Straight into the reading list for graduate stats. One thing that might be useful is a conceptual paragraph about how statistical power/sample size estimation changes. I can imagine (enthusiastic) students stuck on how to adjust what they know about study planning.

2

Reposted by Rob Cavanaugh

Julia M. Rohrer @dingdingpeng.the100.ci · Aug 25

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.

49 280 940

Rob Cavanaugh @rbcavanaugh.bsky.social · Aug 23

Oooh and what does the AI boilerplate section look like?

1

Rob Cavanaugh @rbcavanaugh.bsky.social · Aug 19

This is such a thorough study - what a great example of quality science putting an existing paradigm up to a rigorous test. A-priori power analysis. Preregistered. Open data and code. You can tell the authors made sure to anticipate every “but what if…” response to the null evidence.

1 3

Rob Cavanaugh @rbcavanaugh.bsky.social · Aug 18

Sounds a little viffy

1

Rob Cavanaugh @rbcavanaugh.bsky.social · Aug 18

This is awesome! In a weird coincidence I created something similar a few weeks (github.com/rbcavanaugh/..., mostly with Claude because I’m course prepping 2.5 courses and have enough to do already). I’m glad someone has created something with a bit more intention.

GitHub - rbcavanaugh/toddler: Base R Functions to Mess Up Clean Data for Teaching

Base R Functions to Mess Up Clean Data for Teaching - rbcavanaugh/toddler

github.com

1 1 6

Rob Cavanaugh @rbcavanaugh.bsky.social · Aug 18

And where might be put multicolinearity on this figure? 🛋️🍿

2

Rob Cavanaugh @rbcavanaugh.bsky.social · Jul 22

This is fantastic!

There is something funky - I get the error “could not find function is_null()”. Looks like it’s an rlang export?

1 1

Reposted by Rob Cavanaugh

Solomon Kurz @solomonkurz.bsky.social · Jul 21

New #rstats blog up!

solomonkurz.netlify.app/blog/2025-07...

This is the first in a new series discussing causal inference with experimental data using multilevel models. My basic case is g-computation is the way to go.

Within-person factorial experiments, log(normal) reaction-time data | A. Solomon Kurz

Causal inference with the GLMM, Part 1

solomonkurz.netlify.app

7 20 98

Reposted by Rob Cavanaugh

Richard McElreath 🐈‍⬛ @rmcelreath.bsky.social · Jul 9

How can we reform science? I have some ideas. But I am not sure you’ll like them, because they don’t promise much. elevanth.org/blog/2025/07...

Which Kind of Science Reform

What hope is there for science reform, if we can't agree on what to reform? Right now, principles are more important than practices.

elevanth.org

16 130 280

Rob Cavanaugh @rbcavanaugh.bsky.social · Jun 13

So grateful for this amazing work! Very timely presenting on guideline adherence at CAC weeks ago. Here’s guideline adherence to 2 of your guidelines in US community based care (Red=not adherent). Many assumptions but I suspect a reasonable approximation.

2

Rob Cavanaugh @rbcavanaugh.bsky.social · May 12

(for the record, love the latter scenario where if you spend long enough on vacation, and you have kids, you might enter a sort of relaxation enlightenment)

a scatterplot showing that folks with kids who spend enough days on vacation achieve unheard of levels of relaxation.

1 5

Rob Cavanaugh @rbcavanaugh.bsky.social · May 12

gets messy where

relaxation = ifelse(relaxation < median(relaxation), 0, 1) ?

or something like:

exp((1 - kids)*1*vacation) ?

(amazing by the way...I think I might need to go re-run some old models)

1 2

Rob Cavanaugh @rbcavanaugh.bsky.social · May 6

I love this take but I’m skeptical this is a pervasive view (at least in my subfield). curriculum, practice, and reviewers all take a lot of take time to change behavior

2

Rob Cavanaugh @rbcavanaugh.bsky.social · May 6

Yes! Modelbased and parameters are part of my standard workflow. So simple to put the formatted output in a flextable —> word doc or reference output using inline syntax to automate results sections. Saves me loads of time and reduces manual errors.

r.so

3

Rob Cavanaugh @rbcavanaugh.bsky.social · May 6

Wow I had no idea. Thanks!!

2

Rob Cavanaugh @rbcavanaugh.bsky.social · May 6

Ah I remember that post! Yes I think perhaps something like reducing contrasts to treatment and effects coding, noting others exist, and moving on to marginal effects might be a good compromise.

1

Rob Cavanaugh @rbcavanaugh.bsky.social · May 6

Essentially, most of this: debruine.github.io/faux/article...

Contrasts

debruine.github.io

4 1

Rob Cavanaugh @rbcavanaugh.bsky.social · May 6

Sooo I should just swap out my lectures on contrasts for marginal effects right??

1 4

Rob Cavanaugh @rbcavanaugh.bsky.social · Apr 10

I think Ellyn Riley is/was recently looking for a post doc

1

Reposted by Rob Cavanaugh

Brielle Stark, PhD {she} @briellestark.bsky.social · Apr 8

Had a great time sharing about our test-retest, #multimodal #communication #aphasia & peer dataset at the University of Technology Sydney Graduate School of Health seminar series today. Join AphasiaBank (totally free) and get full access to our data! aphasia.talkbank.org/access/Engli...

4 8

Reposted by Rob Cavanaugh

Jeremy Berg @jeremymberg.bsky.social · Mar 29

Funding curves for K and F mechanisms

(Warning: the new F mechanism "curve" is a lot)

a cartoon of homer simpson holding a card

ALT: a cartoon of homer simpson holding a card

media.tenor.com

5 59 110