Sign in

Robin Blythe

@rbly.bsky.social
687 followers 521 following 2.6K posts

Health economist dabbling in biostats and clinical informatics. orcid.org/0000-0002-3643-4332

PostsRepliesMedia
Robin Blythe @rbly.bsky.social · 14h
Given that coding is basically a mapping from human-to-computer readable language, it really does seem like the natural evolution is to abstract away the human-readable part (ie, code), making it a smooth transition from human lay language straight to binary.
230
Robin Blythe @rbly.bsky.social · 30/09/2026
Actually sort of glad my paper got desk-rejected by Nature Medicine because it means I can keep shit-talking Nature
060
Robin Blythe @rbly.bsky.social · 30/09/2026
How should national expanded carrier screening programs for rare diseases be structured? Evidence from Singapore. First time running a discrete choice experiment and got to do some neat policy analyses. Shoutout to the MVP {logitr} (and, as usual, {ggplot2}) #econsky rdcu.be/6TvA8v58DKpa
020
Robin Blythe @rbly.bsky.social · 30/09/2026
Singapore: Beware of scams! Here's an anti-scam PSA you'll get multiple times a day! Also Singapore: all private companies are allowed to use marketing tactics undifferentiable from how scammers operate. We have no plans to regulate this.
120
Robin Blythe @rbly.bsky.social · 29/09/2026
This was fun. Would have also squeezed Disco Elysium in for a top 10, and this doesn't include what is probably like >8000 hours playing Super Smash Bros and Melee at friends houses
010
Robin Blythe @rbly.bsky.social · 29/09/2026
Applied #statsky and health #econsky friends: any good methods-focused conferences you recommend? I am feeling a little deflated by how little my peers in health economics care about methods. Extra points if there is a focus on Bayes, prediction, and decision theory. I'm in Singapore but can travel!
052
Robin Blythe @rbly.bsky.social · 28/09/2026
Just got a review request from a Frontiers journal that was so intensely personalised to my prior publications (in other journals) that it borders on stalking. A somewhat creepy experience.
000
Robin Blythe @rbly.bsky.social · 27/09/2026
I would like to announce I finally watched the Odyssey and quite enjoyed it, and require no further Discourse at this time. Thank you
070
Robin Blythe @rbly.bsky.social · 26/09/2026
Can we talk about something other than AI for, like, 10 minutes
140
Robin Blythe @rbly.bsky.social · 24/09/2026
This is why it's incredible to me that the US govt is not taking more charge of how these companies interface with their allies. It's using OpenAI and Anthropic in an almost mercantilist way, at least rein in their stupidest behaviours! I guess you need actual experts in charge for that, though.
110
Robin Blythe @rbly.bsky.social · 23/09/2026
New #healtheconomics #preprint in which we propose a proportionality argument for evaluation of clinical AI: that if the model leads to changes in the way patients are treated, then it must meet the evidentiary standards of that treatment.
papers.ssrn.com
<p><span>Towards Rigorous Economic Evidence for Clinical AI</span></p>
<p><span>Clinical artificial intelligence (AI) joins a long list of technologies that promise to improve healthcare, but lack the evidence that it leads to valu
233
Robin Blythe @rbly.bsky.social · 23/09/2026
Any excuse to post my favourite LotR meme
010
Robin Blythe @rbly.bsky.social · 23/09/2026
A lot of bullshit research gets a pass under "exploratory" work. Fishing for p-values is just exploring how to cook the books for your next study. What are some good, actually principled, exploratory stats/Econ methods? #statsky
250
Robin Blythe @rbly.bsky.social · 22/09/2026
It is SO EASY to specify mo() for an ordinal variable in brms that I don't think I can go back to treating them as continuous or dummy vars. Even splines are super easy. Next on my wishlist is that mi() integrates cleanly with mo() and s() 😍 #rstats
081
Robin Blythe @rbly.bsky.social · 20/09/2026
All fields tend to some degree to confuse normative statements for scientific rigour, but I don't think I can tolerate reading any more "Planetary health" papers
140
Robin Blythe @rbly.bsky.social · 17/09/2026
Just as the AI hype types are (poorly) reinventing statistics, the anti-AI crew are (poorly) reinventing decision theory and economics. Please just open a book.
1111
Robin Blythe @rbly.bsky.social · 16/09/2026
We must pace the frontier of logistic regression. It is too powerful, and the risk to humanity of incorrect interpretations of odds ratios is an existential threat #statsky #ai
3274
Robin Blythe @rbly.bsky.social · 14/09/2026
Been teaching grad students a bit lately, and hoo boy, nothing makes you question education quality as much as your complete lack of confidence in your own knowledge.
150
Reposted by Robin Blythe
Martin Modrák @modrakm.bsky.social · 14/09/2026
I will be looking for a PhD student in Bayesian statistics (in collaboration with Paul Bürkner), located in Prague. Forward to students if you have somebody who could be interested. www.martinmodrak.cz/grammo-call/
martinmodrak.cz
Grammo - Let's Research Bayes!
03224
Reposted by Robin Blythe
Journal of Clinical Epidemiology @jclinepi.bsky.social · 10/09/2026
Randomized, controlled, and real - Journal of Clinical Epidemiology www.jclinepi.com/article/S089...
jclinepi.com
Randomized, controlled, and real
The growing use of the term “real-world evidence” (RWE) has introduced a classification of clinical research based on its perceived proximity to routine practice rather than its methodological charact...
055
Robin Blythe @rbly.bsky.social · 10/09/2026
Many of the managers working at these companies are not smart. You can do maths good but if you're consistently handing society reasons to regulate or even destroy you because your actions are irresponsible or delulu, that is idiot behaviour. Ok, now back to statsposting
131
Robin Blythe @rbly.bsky.social · 09/09/2026
#medsky friends: I'm interested in identifying appropriate outcomes for deterioration prediction. It's often death or ICU admit which are poor proxies of a deteriorated state. Also useful: outcomes to identify pts who should be considered palliative/end-of-life, so curative Tx is non-beneficial. TIA
211
Robin Blythe @rbly.bsky.social · 09/09/2026
Me: here's the model object, the paper describing the model, and annotated code so you can use it to estimate the outcome with uncertainty bands for any hypothetical scenario you like. Stakeholder: Yes, but can it deliver ✨insights✨???
151
Robin Blythe @rbly.bsky.social · 07/09/2026
Friendly reminder: Just because they publish in Q1 journals doesn't mean a research group isn't a paper mill! There's plenty of room to call out shitty citation-stuffing associational research for everyone!
042
Robin Blythe @rbly.bsky.social · 03/09/2026
Increasingly finding the 'predictive vs causal' internalised modelling dichotomy a yoke around my neck when submitting papers. Building mechanistic models for simulation is equally valid but fits neither label, and reviewers seem to hate what haven't been exposed to before. #statsky #episky
020
Robin Blythe @rbly.bsky.social · 02/09/2026
Posting about using AI to share a cool project ❌ Posting about using AI to filter out the abusive-Luddite faction ✅
020
Robin Blythe @rbly.bsky.social · 01/09/2026
It's a bit alarming when you put up data and code for a paper under review to facilitate reviewers, then you check the zenodo repo and there are *way* too many downloads for you to not get scared you're about to get scooped. Big barrier to open science IMO
060
Robin Blythe @rbly.bsky.social · 29/08/2026
The worst part of this is that their recommended fix is to collapse continuous vars into 2-3 categories which I know would irritate @f2harrell.bsky.social (it sure bothers me)
470
Robin Blythe @rbly.bsky.social · 29/08/2026
Often having collaborators claim you can't use continuous predictors where there are small sample sizes for individual values of that predictor. Where does this come from? It's incomprehensible to me. That's the entire point of a parametric model! I'm so confused #statsky #rstats
2140
Robin Blythe @rbly.bsky.social · 28/08/2026
For the non health-economists out there, this is like building a linear regression for hypertension with 10 predictors and describing it in the methods as "HyperMOD" as if you believe everyone now needs to use it.
060
Robin Blythe @rbly.bsky.social · 28/08/2026
There is a weirdly egotistical tendency for health economists to build a regular old simulation model and then baptize it with a name like EconMOD as if it's this incredible, permanent contribution to science rather than whatever pile of assumptions they've happened to glue together.
170
Robin Blythe @rbly.bsky.social · 26/08/2026
Most people go to university not for an education but because it makes them a more competitive job applicant. Education - if it occurs at uni - is a byproduct. We shouldn't be surprised that AI cheating is rampant because it was never about learning. (1/5)
2121
Robin Blythe @rbly.bsky.social · 19/08/2026
Delivered 7 hours of guest lectures (over 2 days) on introductory stats to an Executive Masters programme, and after both lectures had to immediately go have a nap. People with teaching commitments should have their research publications weighted doubly, this is exhausting! How do you focus after?!
140
Robin Blythe @rbly.bsky.social · 17/08/2026
A humble observation: no decision-maker I've ever met has expressed an interest in estimating effects with precision. They only care about whether X works for Y and reducing uncertainty enough to address that claim. Effect size estimation is an academic-centric exercise. #statsky #episky #econsky
130
Robin Blythe @rbly.bsky.social · 15/08/2026
Just submitted a grant: estimating individual treatment effects,* using AI,** with "real world"*** in the abstract. It's going to be irresistible, grant funders are powerless 😎 * It's CATE all the way down ** G-computation, regression with interactions *** Literally the theme of the grant call
static.klipy.com
Milhouse's Flirty Wink in The Simpsons
ALT: Milhouse's Flirty Wink in The Simpsons
140
Robin Blythe @rbly.bsky.social · 14/08/2026
Heard a heated debate (among non-PhDs) claiming that the reputation of the uni where you get your PhD will be what defines your career (rather than your advisor, thesis quality, postdoc, publications, etc). This seems hilariously stupid to me but maybe I just live in a bubble?
100
Robin Blythe @rbly.bsky.social · 12/08/2026
Seoul is a great city, the people are super friendly, and the food is great. K-pop is a hard pass for me though, yuck
3100
Robin Blythe @rbly.bsky.social · 07/08/2026
Just reviewed a paper effectively stating that the study methods would only be available "upon reasonable request."
3171
Robin Blythe @rbly.bsky.social · 03/08/2026
If you're the type to eschew epistemic humility and present yourself as a "superstar" academic, or scientist-hero, you are probably also the type to have a relaxed attitude to good scientific practices like avoiding plagiarism. #academia
070
Robin Blythe @rbly.bsky.social · 01/08/2026
Me, once a month #statsky #rstats
3374
Reposted by Robin Blythe
Berna Devezer @devezer.bsky.social · 29/07/2026
It's the unseriousness of the whole thing that gets to me. What is produced as scientific research in one of the top business schools in the country. What they measure and what's being reported have nothing to do with each other. How do they not feel ashamed to circulate this report? (rhetorical)
3505
Robin Blythe @rbly.bsky.social · 28/07/2026
I'll never forget the first time I attended a #datascience seminar: the speaker had a huge dataset and ran a t-test on every combination of variables, pointed to each p<0.05, and said that was your starting point for a ML model or research paper
1111
Robin Blythe @rbly.bsky.social · 26/07/2026
Maturing is realising that most of your field's applied work is based off handwavey assumptions of 💫asymptotic equivalence💫 #econsky
160
Robin Blythe @rbly.bsky.social · 24/07/2026
The worst part is that, as you increasingly prove yourself as a capable investigator, you spend less and less time doing any actual analysis. By the time you reach a/prof, it's all grants and speaking engagements. Meat for the academic machine.
1111
Robin Blythe @rbly.bsky.social · 24/07/2026
When I made research faculty, I naively assumed I'd get to mostly do fun stuff - learning, exploring and analysing my own projects. Instead it's 95% project managing other people's grants: chasing data & IRBs, and dealing with angry primary investigators. Word of warning to any #postdoc out there!
010
Robin Blythe @rbly.bsky.social · 22/07/2026
Intense exercise in the tropics is an ongoing battle between your stomach and your heart as you try to cram as much liquid in as possible to stop your blood turning into an underseasoned ragu. At this point I should become a shareholder in Pocari Sweat with how much I've paid them
040
Robin Blythe @rbly.bsky.social · 16/07/2026
New record for me: one reviewer, 12 pages of comments. This is about twice as much as I had to address for my dissertation.
140
Reposted by Robin Blythe
Berna Devezer @devezer.bsky.social · 15/07/2026
#Statsky I need your help. If you have a consulting gig or been involved in building a (stat. modeling) consulting center, would you be willing to share your business model, sample contracts/terms/conditions, your approach to recruiting clients... esp. in an academic setting, I'd appreciate it!
1169
Robin Blythe @rbly.bsky.social · 15/07/2026
This is baby's first steps into the prediction-causality blood wars, but the more work I do in both, the less convinced I am there are real diffs. The last predictive model I built had a cross-val AUC = 0.96, and it was from a DAG + long discussion with clinicians about which Xs causally influence Y
200
Robin Blythe @rbly.bsky.social · 14/07/2026
Anyone have papers that they regret publishing? They might not be bad science, they could just be trite or useless (though some certainly might be bad). For the sake of the public record, it's good they're out there, but how do you handle no longer supporting your own work?
5132