iiinova — where your digital self keeps evolving

Reposted by Nathaniel B

Saloni @scientificdiscovery.dev · 12/08/2026

I wrote briefly on the Ebola situation. Things could be going a lot better than they are. www.abundanceandgrowth.org/i/210835777/...

A keyhole view of the next pandemic — Saloni Dattani
I’ve been following the Ebola outbreak in central Africa, which has just surpassed 2,000 reported deaths and 4,381 confirmed cases. (The real numbers are likely higher due to underdetection.) What’s surprising about the Ebola outbreak is that we actually have an Ebola vaccine, but it was developed and tested against the Ebola Zaire virus during the 2014–2016 outbreak, while this outbreak is caused by Ebola Bundibugyo virus. The two viruses have similarities and some laboratory and animal data suggested there may be cross-protection, although the evidence is still unclear.1 But until last week, the WHO hadn’t decided to test that vaccine in this outbreak; the focus was on developing and testing new Bundibugyo-specific vaccines and antivirals instead.

In a recent STAT opinion piece, Krutika Kuppalli and Placide Mbala argue that pandemic preparedness shouldn’t be so narrow. Rather than focusing so narrowly on each virus, we should invest in broad-spectrum diagnostics, vaccines, and treatments, and should have tested the effects of the Zaire vaccine on Bundibugyo earlier; I agree.

Interestingly, the Ebola Zaire vaccine came about through innovative trials and coordination. It was developed through funding by BARDA and the DoD, tested using ring vaccination trials led by the WHO, incentivized through an advance market commitment, and stockpiled by Gavi, WHO, UNICEF and the Red Cross for future outbreaks. That stockpile has been useful in the years since then to contain outbreaks of Ebola Zaire quickly. But now that a larger, harder to contain outbreak has arrived, we’ve lost expertise and funding to tackle it, and we don’t know if the existing Ebola vaccine protects against this virus too. We could be doing a lot better.

0299 29

Reposted by Nathaniel B

Vincent Arel-Bundock @vincentab.bsky.social · 17/09/2025

Whoa—my book is up for pre-order! 𝐌𝐨𝐝𝐞𝐥 𝐭𝐨 𝐌𝐞𝐚𝐧𝐢𝐧𝐠: 𝐇𝐨𝐰 𝐭𝐨 𝐈𝐧𝐭𝐞𝐫𝐩𝐫𝐞𝐭 𝐒𝐭𝐚𝐭 & 𝐌𝐋 𝐌𝐨𝐝𝐞𝐥𝐬 𝐢𝐧 #Rstats 𝐚𝐧𝐝 #PyData The book presents an ultra-simple and powerful workflow to make sense of ± any model you fit The web version will stay free forever and my proceeds go to charity. tinyurl.com/4fk56fc8

1130087 300

Reposted by Nathaniel B

Daniel Lakens @lakens.bsky.social · 26/09/2025

You have not seen second generation p-values used anywhere because they were badly reinvented equivalence tests, and people should use equivalence tests instead. See open.lnu.se/index.php/me...

open.lnu.se

LnuOpen | Meta-Psychology

192 9

Reposted by Nathaniel B

The Review of Economic Studies (REStud) @reveconstudies.bsky.social · 17/09/2025

"New study finds salary benchmarking cuts pay gaps by 25%. Pay dispersion partly arises from firms’ uncertainty about market rates, with key implications for pay transparency policy." New paper from Perez-Truglia, Li & Cullen www.restud.com/whats-my-emp...: #econsky #REStud

0105 10

Reposted by Nathaniel B

Posit @posit.co · 12/09/2025

Posit and AWS have signed a Strategic Collaboration Agreement! This collaboration helps customers modernize their data infrastructure & accelerate their data science journey on the cloud, making it easier for teams to build and run critical data science outcomes. Read now posit.co/blog/posit-s...

0144 14

Reposted by Nathaniel B

Thomas Lin Pedersen @thomasp85.com · 11/09/2025

I am beyond excited to announce that ggplot2 4.0.0 has just landed on CRAN. It's not every day we have a new major #ggplot2 release but it is a fitting 18 year birthday present for the package. Get an overview of the release in this blog post and be on the lookout for more in-depth posts #rstats

tidyverse.org

ggplot2 4.0.0

A new major version of ggplot2 has been released on CRAN. Find out what is new here.

9851281 851

Reposted by Nathaniel B

Julia M. Rohrer @dingdingpeng.the100.ci · 11/09/2025

PS there’s a (mildly) updated version available now: osf.io/g4s2a_v2

osf.io

OSF

2131 13

Reposted by Nathaniel B

Julia M. Rohrer @dingdingpeng.the100.ci · 25/08/2025

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.

461001284 1K

Nathaniel B @bechhof.bsky.social · 21/08/2025

We can innovate with new setups though instead of having it be academic / rigorous (with heavy caveats for that adjective) vs data bloggers. Having a reproducibility norm would be critical to such a new setup.

010 1

Reposted by Nathaniel B

Gábor Békés @gaborbekes.bsky.social · 15/08/2025

If interested please test the notebooks. Examples: Describe github.com/gabors-data-... OLS github.com/gabors-data-... Random forest, SHAP github.com/gabors-data-... Panel FE, FD github.com/gabors-data-...

github.com

da_case_studies/ch10-hotels-multiple-reg/ch10-hotels-multiple-reg.ipynb at master · gabors-data-analysis/da_case_studies

Codes for case studies for the Bekes-Kezdi Data Analysis textbook - gabors-data-analysis/da_case_studies

082 8

Nathaniel B @bechhof.bsky.social · 15/08/2025

I will insist on maximizing interaction terms on both platforms

010 1

Reposted by Nathaniel B

Scottish Journal of Political Economy @scot-j-pol-econ.bsky.social · 18/06/2025

Very pleased to announce our special issue on Gender Norms, edited by @emmatominey.bsky.social and @nikkishure.bsky.social. onlinelibrary.wiley.com/doi/toc/10.1...

11010 10

Reposted by Nathaniel B

Rafe Meager (they/them) @economeager.bsky.social · 14/06/2025

this was excellent. a highly recommended way to spend 5 hours.

2316 31

Nathaniel B @bechhof.bsky.social · 13/06/2025

St Louis and Baltimore show up as the same point because they both went exactly from 74 to 56 in raw counts, which yields the same raw % decline and model predicted underlying % decline.

000 0

Nathaniel B @bechhof.bsky.social · 13/06/2025

I indulged in a bit of fun doing a first pass of a regularized model for homicide declines with ChatGPT, using @jeffasher.bsky.social's data from jasher.substack.com/p/why-i-thin... to estimate "underlying" rates of homicide decline by city.

Scatter plot of 31 U.S. cities (excluding “Total”) showing raw vs. Bayesian MAP–regularized percent change in homicides from 2024 to 2025. Raw changes are on the x-axis, regularized changes on the y-axis. Each point is labeled by city, with small arrows connecting any labels placed more than 10 pixels from their marker. A caption below summarizes the log‐ratio transformation, hierarchical Normal(μ,τ²) model with an Inverse‐Gamma penalty on τ², and the posterior‐mean back-transform to percent change.

Screenshot of Jeff Asher’s Substack titled “Murder is down a lot in big cities.” Below the title is a subtitle: “2025 vs 2024 YTD murders in top 30 most murderous cities.” A table lists four columns—City, YTD 2024, YTD 2025, and % Change—and shows, in descending order of decline, that Denver fell from 29 to 13 homicides (–55.2 %), Birmingham from 61 to 33 (–45.9 %), Cleveland 54→30 (–44.4 %), San Antonio 50→30 (–40.0 %), Las Vegas 54→35 (–35.2 %), Atlanta 50→33 (–34.0 %), Dallas 90→60 (–33.3 %), Indianapolis 70→49 (–30.0 %), and so on.

100 1

Reposted by Nathaniel B

Crystal Lewis @cghlewis.bsky.social · 13/05/2025

Thoughtfully planning your data collection efforts has a huge return on investment. As someone who has cleaned data from projects both with and without data management planning, I can safely say it is worth the effort. 40 hours of planning > 80+ hours of cleaning messy data

0162 16

Nathaniel B @bechhof.bsky.social · 13/05/2025

There's just something special about running a proper Bayesian MCMC script that I can't replace with anything else after doing it for almost 10 years

A screenshot of a Bayesian statistical model training interface showing NUTS sampler progress. The display includes information about a model with 45 subgroup columns, training on 52,380 of 68,850 rows. The sampling is being performed using PyMC NUTS (JAX-compiled) with 4 chains running in parallel. The table shows the current progress of each chain with metrics including number of draws (ranging from 601-644), step size (0.01), gradient evaluations, sampling speed (approximately 1.14-1.22 draws/s), elapsed time (8:49), and remaining time estimates.

130 3

Reposted by Nathaniel B

Julian Reif @julianreif.bsky.social · 10/05/2025

“First, about half the time I reanalyze a study, I find that there are important bugs in the code, or that adding more data makes the mathematical finding go away, or that there’s a compelling alternative explanation for the results.” davidroodman.com/blog/2025/05...

davidroodman.com

Appeal to Me: First Trial of a “Replication Opinion” - David Roodman

My employer, Open Philanthropy, strives to make grants in light of evidence. Of course, many uncertainties in our decision-making are irreducible. No amount of thumbing through peer-reviewed journals ...

076 7

Reposted by Nathaniel B

Johanna Rickne @johannarickne.bsky.social · 09/05/2025

7/ A series of errors create an illusion of statistically significant treatment effects of similar size in the five strategies. Examples include changing/adding/removing variables in regressions, using incorrect bandwidths, and changing clustering methods.

1464 46

Reposted by Nathaniel B

Johanna Rickne @johannarickne.bsky.social · 09/05/2025

1/ The paper claimed that the reform caused a jump in rape by 50–60 percent. This seemed hard to reconcile with the flat time trend in reported rapes around the time of the reform.

1476 47

Reposted by Nathaniel B

Leslie Root @lesja.bsky.social · 04/04/2025

I need a newsletter because I have so many words inside of me about the "they say they want to raise the birth rate and yet they are decimating maternal and infant health surveillance infrastructure [or whatever else], what a gotcha" thing

3201 20

Nathaniel B @bechhof.bsky.social · 04/04/2025

Jeet is wrong about fighting inflation hurting growth tho

020 2

Nathaniel B @bechhof.bsky.social · 03/04/2025

You still only have N=1 per unit of the running variable when its cohort, once you consider the hierarchy. Again refer to the survey example: even if you have a measure with 10,000 respondents per unit time, that doesn't negate the issue that you only have one unit of time per unit time.

100 1

Nathaniel B @bechhof.bsky.social · 03/04/2025

It's similar to using really large surveys for a pre-post comparison. Yes you're increasing the N literally in a tabular sense, but you still effectively have a few data points across time that you're measuring with more and more precision as you jack up N.

100 1

Nathaniel B @bechhof.bsky.social · 03/04/2025

I get this last point, but the distinctions you're emphasizing seem irrelevant, with the logic from the AR piece applying just as much for cohort. For example, what looks like increased N isn't that simple here, because of correlation between people of the same cohort.

100 1

Nathaniel B @bechhof.bsky.social · 03/04/2025

It's an RDD in cohort, which, when only using one cutoff (not eg one repeated year after year), has the same issues from the Annual Reviews piece you link: - requires observations far from the threshold - requires considering the time-series nature of the DGP - the McCrary test becomes irrelevant

120 2

Reposted by Nathaniel B

Pam Davis-Kean, PhD @umpamdk.bsky.social · 02/04/2025

@icpsr.bsky.social @umich.edu is hosting DataLumos, which is a crowdsource repository of government data archive.icpsr.umich.edu/datalumos/home

archive.icpsr.umich.edu

A crowd-sourced repository for valuable government data | DATALUMOS

022 2

Reposted by Nathaniel B

Saloni @scientificdiscovery.dev · 02/04/2025

Seeing the "shingles vaccine protects from dementia" claim going around again. Unfortunately the main study it's based on is just not convincing. @epiellie.bsky.social has written a great post about why:

medium.com

Shingles vaccine & dementia prevention: too good to be true?

A public peer-review

1312833 128

Reposted by Nathaniel B

Andrew Heiss @andrew.heiss.phd · 03/04/2025

hey hey my constant use of the palmerpenguins data is now relevant for my public policy classes

0326 32

Reposted by Nathaniel B

Peter Hull @instrumenthull.bsky.social · 28/03/2025

📢New WP w/ @borusyak.bsky.social We derive the most powerful recentered IVs for formula treatments & propose an algorithm for approximating them This approach yields *huge* power gains, relative to conventional simulated IV, when estimating Medicaid crowdout effects Check it out! t.co/UPg2XQBefN

16312 63

Nathaniel B @bechhof.bsky.social · 27/03/2025

I disagree, since the magnitude of the expansions she talks about would require a much bigger tax base

110 1

Nathaniel B @bechhof.bsky.social · 27/03/2025

That mischaracterizes my take here. I want them to include principles of how they'll handle tradeoffs.

100 1

Nathaniel B @bechhof.bsky.social · 27/03/2025

I hold it against many candidates. The wealth tax isn't close to enough to fund all the program expansions listed on the page. The policies don't have to be pie-in-the-sky in practice, but having concepts of a plan for delivering them is not too much to ask IMO.

100 1

Nathaniel B @bechhof.bsky.social · 27/03/2025

There's no detail at all other than a 2% wealth tax! There's a wishlist, which very much has 2019 Corbyn Labour manifesto vibes. It's a question of whether there's any understanding of tradeoffs at all, which comes across as non credible.

100 1

Nathaniel B @bechhof.bsky.social · 27/03/2025

Lack of detail on how the proposals could be financed probably the biggest one, as part of a broader issue of seeming not to care about details (like in theory you could have a $25/hour minimum wage if you're serious about subsidies, but it's way off to just make it statutory and be done with it)

100 1

Nathaniel B @bechhof.bsky.social · 26/03/2025

Standardization and division of labor FTW

000 0

Nathaniel B @bechhof.bsky.social · 26/03/2025

I agree with that, but having read the policy pages on Abughazaleh's website, I'm not optimistic about how much of an improvement the replacement would be overall. I'm hoping she gets better about that stuff over time.

110 1

Reposted by Nathaniel B

Joram Mutenge @jorammutenge.com · 24/03/2025

Always remove whitespace before doing a unique count of values to get an accurate count. In @pola.rs, there's a function to do just that! #100DaysOfPolars #datasky

Image showing polars code for data analysis.

063 6

Nathaniel B @bechhof.bsky.social · 24/03/2025

Do a weighted average of ideological extremity and partisanship extremity and you get pretty much there yeah

010 1

Reposted by Nathaniel B

NBER @nber.org · 23/03/2025

Healthcare employment in the US has skyrocketed since 1980. Healthcare is a middle-class jobs engine, but "manufacturing-to-meds" transitions are not saving the Rust Belt, from @gottliebecon.bsky.social, @nealemahoney.bsky.social, Kevin Rinz, and Victoria Udalova www.nber.org/papers/w33583

0164 16

Reposted by Nathaniel B

Crystal Lewis @cghlewis.bsky.social · 18/03/2025

Trying out dplyr::consecutive_id() for a project today. #rstats It generates a unique identifier that increments every time a variable (or combination of variables) changes.

612527 125

Nathaniel B @bechhof.bsky.social · 14/03/2025

Let it be known far and wide that @haroldpollack.bsky.social follows up (including custom cup!) when you answer his call for tracking down clean data. He really is the top notch conversationalist you know from his many podcast/media appearances

Nathaniel Bechhofer and Harold Pollack posing for a selfie with the custom cup that Pollack gave as a reward for linking a spreadsheet with 2024 presidential election results by Congressional district.

020 2

Nathaniel B @bechhof.bsky.social · 10/03/2025

I also did see this on Twitter x.com/arachno_comm...

x.com

200 2

Reposted by Nathaniel B

Carlisle Rainey 👨‍💻📊📚 @carlislerainey.bsky.social · 10/03/2025

"The Need for Equivalence Testing in Economics" from Jack Fitzgerald (@jackfitzgerald.bsky.social) Preprint: osf.io/preprints/met... We know that "not significant" does not imply evidence for "no effect," but I still see papers make this leap. Good to see more work making this point forcefully!

4289 28

Nathaniel B @bechhof.bsky.social · 10/03/2025

I think @kavi.bsky.social or @xenocryptsite.bsky.social should be able to point you to it already done

100 1

Reposted by Nathaniel B

Simon Willison @simonwillison.net · 08/03/2025

Here's the handout for my "Cutting-edge web scraping techniques" workshop at #NICAR2025 this morning github.com/simonw/nicar... Plus some extra notes on the custom software I built to support the workshop: simonwillison.net/2025/Mar/8/c...

simonwillison.net

Cutting-edge web scraping techniques at NICAR

Here's the handout for a workshop I presented this morning at [NICAR 2025](https://www.ire.org/training/conferences/nicar-2025/) on web scraping, focusing on lesser know tips and tricks that became po...

214618 146

Nathaniel B @bechhof.bsky.social · 08/03/2025

I think this @aoc.bsky.social proposal to have a hard cap of 10% for credit card interest is a very bad idea; the only decent argument for it is a paternalist one that basically assumes people with poor credit should not get credit cards. We can do better helping those people tho in targeted ways.

011 1