iiinova — where your digital self keeps evolving

Reposted by Rich Morris

Hane @hane.bsky.social · 23/04/2026

I'm a mid-career researcher insofar as my research career is mid

Rich Morris @sometimesdata.bsky.social · 18/03/2026

I'm trying to find examples of replicable mistakes by Chat GPT. I seem to recall there was a thread on here about common causal inference mistakes made by AI. Does anyone have a tip? #causalsky

140

Rich Morris @sometimesdata.bsky.social · 19/02/2026

Some of the creators of Claude or other LLMs assign a greater than 0% chance the LLM is conscious. Something as high as 20 to 80%. This would be ethically problematic but I don't think this is right. Claude is non-committal on the topic but it also says it doesn't experience the passage of time.

120

Rich Morris @sometimesdata.bsky.social · 14/02/2026

“Every student learns that correlation does not imply causation. Few learn the converse: absence of correlation does not imply absence of causation. Both mistakes are common in applied work, and understanding why reveals something fundamental about what causal inference requires.”

140

Rich Morris @sometimesdata.bsky.social · 22/01/2026

Mental health interventions are now digital but over 50% of the variance in effectiveness for stress is due to differences in design and features. Expert design is needed and throwing everything at the wall is not helpful: www.jmir.org/2025/1/e71253/

jmir.org

Optimizing Workplace Digital Mental Health Interventions: Systematic Review and Meta-Analysis

Background: Digital mental health interventions (DMHIs) are widely used in workplaces to address common mental health conditions such as depression, anxiety, and stress. Their content and design vary ...

021

Reposted by Rich Morris

The R Foundation @r-foundation.bsky.social · 22/01/2026

We are responding to this call for evidence: ec.europa.eu/info/law/bet... Do you have concrete examples of the added value of #Rstats in the public or private sectors? Include the most important factors (risk, lock-in, security, innovation...) to assess the added value. Thanks!

ec.europa.eu

European Commission - Have your say

73537

Rich Morris @sometimesdata.bsky.social · 27/12/2025

Is there something in the water of young men today?

310

Rich Morris @sometimesdata.bsky.social · 10/12/2025

Does anyone else just surf LinkedIn hoping to find a job using R, Quarto, or Shiny? #rstats

190

Reposted by Rich Morris

Danger, Bees! @foolishhoomans.bsky.social · 12/10/2025

America is great when kids can't play soccer because the secret police may steal their parents

0234

Rich Morris @sometimesdata.bsky.social · 01/08/2025

I feel like we are in a golden age of AI right now, like when google first appeared and you could get your search results without adverts. How long before AI results start incorporating ads?

210

Rich Morris @sometimesdata.bsky.social · 28/07/2025

"Either politics throws open the doors, or it won’t be long before they’re kicked in by angry citizens as they smash every last stick of furniture and walk out with the chandelier of power" -this was released two years before Jan 6th open.spotify.com/track/4Sw9Xa...

open.spotify.com

Steve And Fatima

Deadbeat · Wax Poetic For This Our Great Resolve · Song · 2018

020

Reposted by Rich Morris

Coach Finstock @coachfinstock.bsky.social · 03/07/2025

Yeah so there's no way anyone is coming here for World Cup or Olympics. if you can bag up the son of one of the most famous boxers to ever live, you can do that to anyone.

221050368

Rich Morris @sometimesdata.bsky.social · 03/07/2025

protip for Gen Z working with older colleagues: if discussing data or studies which occurred entirely after the year 2000, you can refer to them as *21st Century*. It will sound futuristic and sci-fi to anyone who grew up on Buck Rogers, Flash Gordon (i.e., Gen X, Boomers) - very impressive!

020

Rich Morris @sometimesdata.bsky.social · 03/07/2025

This is a neat summary of the history of applied statistics I’ve read

110

Reposted by Rich Morris

Darren Dahly @statsepi.bsky.social · 03/07/2025

We constantly empower people that have no interest in the truth whatsoever and I struggle to understand what can be done about that.

1173

Reposted by Rich Morris

Bruno Rodrigues @brodriguesco.bsky.social · 03/07/2025

New blog post: You can outsource the grunt work to an LLM, not expertise brodrigues.co/posts/2025-0...

0122

Rich Morris @sometimesdata.bsky.social · 03/07/2025

Add this to my reading list: “For “random intercept” models, adding the group-level averages of all included variables as regressors (or a variety of equivalent procedures such as group-wise centering) relieves the bias otherwise induced by this regularization”

010

Rich Morris @sometimesdata.bsky.social · 23/06/2025

Spoke to Jen a few weeks ago because I am also trying to pivot out of academia. Can confirm Jen is a great person to help you focus on what’s important (to anyone else who is in a similar situation). She would be a great asset to any community building enterprise

040

Reposted by Rich Morris

Thom Volker @thomvolker.bsky.social · 18/06/2025

After two weeks, I'm finally done! In this post, I explain different approaches for solving linear regression in R: directly, using QR, singular value and Cholesky decompositions, and do some benchmarking for comparison with in-built approaches. thomvolker.github.io/blog/2506_re...

611933

Reposted by Rich Morris

Chris Kenny, PhD @chriskenny.bsky.social · 17/06/2025

A lawsuit in Rockland County, NY, alleges election irregularities in the 2024 Presidential and Senate races. The plaintiffs claim statistical anomalies and uncounted votes. But do these claims hold up? I analyzed the data and I don't think they do: christophertkenny.com/posts/2025-0...

christophertkenny.com

683

Rich Morris @sometimesdata.bsky.social · 12/06/2025

Lord save me! Where do people get the idea the y-axis must always include zero, and every panel in a collection of panels must have the same y-axis scale? These a fine “rules of thumb” but they are not stone commandments handed down by the #dataviz gods!!! #rstats

4280

Reposted by Rich Morris

Matti Vuorre @matti.vuorre.com · 04/06/2025

The quarto-preprint extension v0.16 (github.com/mvuorre/quar...) now enables full-width content in two-column #quartopub #typst documents (and more!; via github.com/christopherk...). Hope this makes Quarto a bit more useful for writing well-typeset reproducible #rstats manuscripts 💙

A screenshot of a two-column PDF produced with Quarto and Typst illustrating a full-width figure that spans both columns.

1267

Reposted by Rich Morris

Calvin and Hobbes @calvin-and-hobbes.bsky.social · 03/06/2025

When your past is questionable—revise it! #CalvinandHobbes #FakeHistory

06313

Rich Morris @sometimesdata.bsky.social · 27/05/2025

One of my pet peeves with dplyr verbs in R is filter(). It should remove the rows which meet the condition(s). What is the most readable way to remove rows which meet a joint condition? e.g., remove rows where col_a == "x" & col_b > y? #rstats

541

Rich Morris @sometimesdata.bsky.social · 26/05/2025

Bluesky snark at its finest. Just learnt so much from the linked responses. The debate is alive! :-)

020

Reposted by Rich Morris

Julia M. Rohrer @dingdingpeng.the100.ci · 18/05/2025

Trying to keep an open mind only to end up confirming my prejudices

051

Rich Morris @sometimesdata.bsky.social · 15/05/2025

Just had a terrible experience at BMC Public Health. According to @springernature.com Australian universities can publish in their OA journals for free, so you might *think* an OA journal such as BMC Public Health would be covered... you would be wrong. Over $5000 wrong!

120

Rich Morris @sometimesdata.bsky.social · 15/05/2025

The UCLA stats group do fantastic work explaining things. This tutorial on SEM in R is gold: stats.oarc.ucla.edu/r/seminars/r...

stats.oarc.ucla.edu

Introduction to Structural Equation Modeling (SEM) in R with lavaan

100

Rich Morris @sometimesdata.bsky.social · 13/05/2025

I should really read the fcuking paper before butting in on expert discussion here!

100

Rich Morris @sometimesdata.bsky.social · 12/05/2025

TIL the difference between interaction and effect modification: "[An] interaction is the effect of 2 interventions whereas effect modification is the effect of one intervention varying across strata of a second variable." I think this means things like sex cannot interact with other variables...

230

Reposted by Rich Morris

Collin Berke @collinberke.bsky.social · 11/05/2025

Spent more time exploring {scales} for #ggplot2. I found the `label_dictionary()` function to be useful for relabeling the values of a discrete scale using a lookup table: scales.r-lib.org/reference/la... Here's what I came up with, using some NCAA basketball data from the #rstats {hoopR} package👇

4114

Rich Morris @sometimesdata.bsky.social · 09/05/2025

Reporting observational research: www.auajournals.org/doi/full/10.... Some simple guidelines and good recommendations for language in intro, method and discussion.

auajournals.org

Guidelines for Reporting Observational Research in Urology: The Importance of Clear Reference to Causality | Journal of Urology

000

Rich Morris @sometimesdata.bsky.social · 07/05/2025

stringr::str_wrap(width = ...) is fantastic! Works for facet headings as well: facet_wrap(~str_wrap(myfacet, width = 10)

030

Reposted by Rich Morris

Aleksander Molak @alxndrmlk.bsky.social · 05/05/2025

The founder of scikit-learn: "Without causal thinking, machine learning is not enough for optimal clinical decisions." 🧵 1/ #CausalSky

791

Rich Morris @sometimesdata.bsky.social · 30/04/2025

In psychology, an operational definition specifies how a variable or concept is to be measured and defined within a particular study. How does this compare to a statistical estimand, which also defines what a study aims to estimate? I know they aren't the same but they seem quite similar #StatsSky

000

Reposted by Rich Morris

JD Long @jdlong.cerebralmastication.com · 27/04/2025

Well I feel like a fool. I’ve been saying NFTs and crypto serve no fundamental purpose. And I totally didn’t entertain the use case of bribing the US President. So embarrassing

4333

Reposted by Rich Morris

Erik Reinbergs @ereinbergs.bsky.social · 20/04/2025

I knew about with() but how did I not know about within()!? #rstats

0102

Rich Morris @sometimesdata.bsky.social · 11/04/2025

Starting a PhD makes you sick. After entering a PhD program, Swedish students seek more psychotropics: swopec.hhs.se/lunewp/abs/l...

130

Rich Morris @sometimesdata.bsky.social · 09/04/2025

Why not subset, transform and aggregate? (filter and mutate were terrible verb choices anyway - we filter _out_, nobody filters _in_)

010

Rich Morris @sometimesdata.bsky.social · 31/03/2025

How do you write your probabilistic models?

error term convention (with stochastic error terms)

100

Rich Morris @sometimesdata.bsky.social · 31/03/2025

Every few years I have to remind myself of this simple and clear post describing how modes, medians and means arise naturally from zero-one loss (accuracy), absolute difference (MAE), and squared error (MSE), respectively: www.johnmyleswhite.com/notebook/201...

johnmyleswhite.com

Modes, Medians and Means: A Unifying Perspective

Introduction / Warning Any traditional introductory statistics course will teach students the definitions of modes, medians and means. But, because introductory courses can’t assume that students have...

020

Rich Morris @sometimesdata.bsky.social · 29/03/2025

for the last few weeks I've been getting more interaction on bsky than ex-twitter - more interesting people here, and more interesting content. I still visit ex out of habit, but bsky is where it's at, and I'm here for it

020

Rich Morris @sometimesdata.bsky.social · 29/03/2025

This is the first nuanced view of the role of replication in science I've read in a while. Like a breath of fresh air after sitting through the replication crisis in [insert favourite science here].

261

Rich Morris @sometimesdata.bsky.social · 18/03/2025

Learning maths again with my 11yo. Why does multiplying two negative numbers produce a positive? Consider: (-1)*[1+(-1)] = 0 since [1+(-1)] is zero By distribution: (-1)(1) + (-1)(-1) = 0 (-1) + (-1)(-1) = 0, so (-1)(-1) must be positive one

100

Reposted by Rich Morris

Adam Kucharski @adamjkucharski.bsky.social · 15/03/2025

While I was reading, it occurred to me that this was a version of a mathematical puzzle I was already familiar with. It’s commonly known as the ‘German tank problem’, and it shows how we can sometimes extract a lot of valuable information from relatively little data. So, how does it work? 5/

1111

Rich Morris @sometimesdata.bsky.social · 12/03/2025

Just updated to macOS Sequoia 15.3.2. Apple finally allowed me to set the scroll direction on my mouse and trackpad independently! Such a clever little company!!!

100