corystat.bsky.social
LightNews LightNews
corystat.bsky.social
@corystat.bsky.social
10 followers 83 following 0 posts
Posts Replies Media Videos
Reposted by corystat.bsky.social
tradediversion.bsky.social
The most popular data set among this morning’s presentations at #NBERSI ITI is the OECD’s Inter-Country Input-Output tables. ICIO is a substitute for WIOD. www.oecd.org/en/data/data...
Inter-Country Input-Output tables
Inter-Country Input-Output tables (ICIO) provide an international statistical infrastructure that maps flows of production, consumption, investment within countries and flows of international trade in...
www.oecd.org
July 7, 2025 at 6:00 PM
Reposted by corystat.bsky.social
solomonkurz.bsky.social
New #rstats blog up!

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

This is the first in a brief series where we use {brms} to learn {Stan} code.

Many thanks to @fusaroli.bsky.social and @stephenjwild.bsky.social for their helpful reviews.
Learn Stan with brms, Part I | A. Solomon Kurz
y ~ 1
solomonkurz.netlify.app
July 7, 2025 at 3:06 PM
Reposted by corystat.bsky.social
rocb-ento.bsky.social
Check out this incredible stack exchange answer by the author of R package MatchIt on why/when matching is preferable to regression for causal inference:

stats.stackexchange.com/questions/54...
Why do we do matching for causal inference vs regressing on confounders?
I'm new to the area of causal inference. From what I understand, one of the main concerns that causal inference tries to address is the effect of confounders! For the sake of reference, let's denot...
stats.stackexchange.com
March 21, 2025 at 1:13 PM
Reposted by corystat.bsky.social
rmcelreath.bsky.social
This paper was fun. I got to do some math! The inference problem illustrated below. Because of finite samples, rarer behavior is under-sampled, so true repertoires underestimated (left). Our model corrects for that bias (right).

Colleagues using model on crows now, and we'll do chimps next.
Left: Plot showing behavioral repoertoire size (x-axis) against observed repertoire - there is a bias such that observed is often smaller than true. Right: posterior estimates from model, showing corrected inference
March 22, 2025 at 6:27 AM
Reposted by corystat.bsky.social
drk-lo.bsky.social
Good article on backing up all your data and documents www.nytimes.com/2025/02/14/b...
Back Up Everything. Even if Elon Musk Isn’t Looking at It.
Readers worried after Mr. Musk and his team were given access to federal payment systems. Here’s how to back up all your data and documents.
www.nytimes.com
February 17, 2025 at 1:11 AM
Reposted by corystat.bsky.social
adam-lg.bsky.social
"A Latent Causal Inference Framework for Ordinal Variables"
Paper: arxiv.org/abs/2502.10276
#rstats Code: github.com/martinascaud...

#stats
Ordinal variables, such as on the Likert scale, are common in applied research. Yet, existing methods for causal inference tend to target nominal or continuous data. When applied to ordinal data, this fails to account for the inherent ordering or imposes well-defined relative magnitudes. Hence, there is a need for specialised methods to compute interventional effects between ordinal variables while accounting for their ordinality. One potential framework is to presume a latent Gaussian Directed Acyclic Graph (DAG) model: that the ordinal variables originate from marginally discretizing a set of Gaussian variables whose latent covariance matrix is constrained to satisfy the conditional independencies inherent in a DAG. Conditioned on a given latent covariance matrix and discretisation thresholds, we derive a closed-form function for ordinal causal effects in terms of interventional distributions in the latent space. Our causal estimation combines naturally with algorithms to learn the latent DAG and its parameters, like the Ordinal Structural EM algorithm. Simulations demonstrate the applicability of the proposed approach in estimating ordinal causal effects both for known and unknown structures of the latent graph. As an illustration of a real-world use case, the method is applied to survey data of 408 patients from a study on the functional relationships between symptoms of obsessive-compulsive disorder and depression. Keywords: Causal inference, Causal diagrams, Latent graphical models, Directed acyclic graph-probit, Ordinal data.
February 17, 2025 at 3:59 AM
Reposted by corystat.bsky.social
dschraff.bsky.social
Important points to improve the review process for quantitative papers. I found this recent @thejop.bsky.social paper really useful as well www.journals.uchicago.edu/doi/10.1086/...
What Good is a Regression? Inference to the Best Explanation and the Practice of Political Science Research | The Journal of Politics: Vol 0, No ja
www.journals.uchicago.edu
February 17, 2025 at 3:53 PM
Reposted by corystat.bsky.social
solomonkurz.bsky.social
Check out this sweet new blog post from Julia, which also references one of my old blog posts (solomonkurz.netlify.app/blog/2022-06...).
January 22, 2025 at 8:24 PM
Reposted by corystat.bsky.social
jebyrnes.bsky.social
OK, here is a very rough draft of a tutorial for #Bayesian #SEM using #brms for #rstats. It needs work, polish, has a lot of questions in it, and I need to add a references section. But, I think a lot of folk will find this useful, so.... jebyrnes.github.io/bayesian_sem... (use issues for comments!)
Full Luxury Bayesian Structural Equation Modeling with brms
jebyrnes.github.io
December 21, 2024 at 7:49 PM
Reposted by corystat.bsky.social
mattansb.msbstats.info
Hey #StatsSky, what are you favorite papers to cite when you need to justify something that is obvious (I once had a reviewer ask we justify the use of logistic regression on a binary outcome)
or when you need to push-back on silly reviewer requests (e.g., asking for p-values in table 1)?
January 6, 2025 at 8:15 AM
Reposted by corystat.bsky.social
rmkubinec.bsky.social
I'm just saying this is syntatically correct #rstats code
January 16, 2025 at 4:38 PM
Reposted by corystat.bsky.social
7 steps to junk science that can achieve worldly success
statmodeling.stat.columbia.edu/2025/01/17/7...
7 steps to junk science that can achieve worldly success | Statistical Modeling, Causal Inference, and Social Science
statmodeling.stat.columbia.edu
January 17, 2025 at 2:48 PM
Reposted by corystat.bsky.social
andrew.heiss.phd
The course websites for my Spring 2025 causal inference and data visualization classes (both with #rstats) are live!

evalsp25.classes.andrewheiss.com

datavizsp25.classes.andrewheiss.com
Program Evaluation for Public Service – Program Evaluation
Combine research design, causal inference, and econometric tools to measure the effects of social programs
evalsp25.classes.andrewheiss.com
January 8, 2025 at 6:11 PM
Reposted by corystat.bsky.social
f2harrell.bsky.social
1/2 Major release 6.9-0 of #RStats rms package now available on CRAN w/complete re-write of the binary/ordinal logistic regression function lrm. lrm began in 1980 as SAS PROC LOGIST so it was time for a re-do for iterative calculations. Details are at fharrell.com/post/mle
Statistical Computing Approaches to Maximum Likelihood Estimation – Statistical Thinking
Maximum likelihood estimation (MLE) is central to estimation and development of predictive models. Outside of linear models and simple estimators, MLE requires trial-and-error iterative algorithms to ...
fharrell.com
December 20, 2024 at 1:29 PM
Reposted by corystat.bsky.social
andrew.heiss.phd
New blog post! Have you (like me!) wondered what the ATT means and how it's different from average treatment effects? I use #rstats to explore why we care about the ATE, ATT, and ATU + how to calculate them with observational data! #polisky #episky #econsky www.andrewheiss.com/blog/2024/03...
Before we calculate these different treatment effects with the realized outcomes instead of the hypothetical potential outcomes, let’s look really quick at the practical difference between the true ATE, ATT, and ATU. All three estimands are useful for policymaking! The ATE is −15, implying that mosquito nets cause a 15 point reduction in malaria risk for every person in the country. This includes people who live at high elevations where mosquitoes don’t live, people who live near mosquito-infested swamps, people who are rich enough to buy Bill Gates’s mosquito laser, and people who can’t afford a net but would really like to use one. If we worked in the Ministry of Health and wanted to know if we should make a new national program that gave everyone a free bed net, the overall reduction in risk is −15, which is probably pretty good! The ATT is −16.29, which is bigger than the ATE. The effect of net usage is bigger for people who are already using the nets. This is because of underlyi Mirrored histogram showing “weird” parts of the population: treated people who were unlikely to be treated, and untreated people who were likely to be treated Mirrored histogram showing pseudo-populations of treated and untreated people that have been reweighted to be more comparable and unconfounded
March 21, 2024 at 1:50 PM
Reposted by corystat.bsky.social
dingdingpeng.the100.ci
Currently working my way through @julianschuess.bsky.social and @pselb.bsky.social's "Graphical Causal Models for Survey Inference" for our journal club and so far it is a thing of beauty! Very clear language, clear arguments, meeting readers where they are etc
journals.sagepub.com/doi/10.1177/...
November 8, 2024 at 4:13 PM
Reposted by corystat.bsky.social
tompepinsky.com
On the choices that Americans like me now face, inspired by 20 years of research and experience on nondemocratic politics around the world.

Read and don't look away.
Living through the Next American Political Order: Institutions Will Comply, and You Will Be Made Complicit
I woke up this morning to see that Donald Trump had been elected president again. Like roughly half of American voters, this is not the outcome that I had hoped for. Politics, like history, is chao…
tompepinsky.com
November 6, 2024 at 2:05 PM
Reposted by corystat.bsky.social
mzloteanu.bsky.social
#statstab #205 Equivalence Tests Using {marginaleffects}

Thoughts: Easily testing for "no effect" is important. Here's a tutorial by @carlislerainey using @VincentAB r pkg.

#equivalencetests #equivalence #TOST #marginaleffects #tutorial #nullresults #r

www.carlislerainey.com/blog/2023-08...
Equivalence Tests with {marginaleffects}
Reproducing the Clark and Golder (2006) example from Rainey (2014)
www.carlislerainey.com
October 18, 2024 at 4:26 PM
Reposted by corystat.bsky.social
dingdingpeng.the100.ci
I’ve said it before (as have others) and I’ll say it again, there’s no such thing as a non-causal mediation analysis.

journals.sagepub.com/doi/10.1177/...
October 17, 2024 at 6:44 PM
Reposted by corystat.bsky.social
edgarmerkle.bsky.social
blavaan 0.5-6 was recently released to CRAN. It has always done Bayesian structural equation modeling, but in recent years we have added some functionality for ordinal observed variables and for two-level SEM (random intercepts). More info here:
Bayesian Latent Variable Analysis
Fit a variety of Bayesian latent variable models, including confirmatory factor analysis, structural equation models, and latent growth curve models. References: Merkle & Rosseel (2018) <doi:10.18637/...
ecmerkle.github.io
September 20, 2024 at 7:10 PM
Reposted by corystat.bsky.social
stephenjwild.bsky.social
My minor contribution to the social media -> mental health debate. What's the estimand in time reduction RCTs? With a slightly unorthodox use of DAGs.

#stats

sjwild.github.io/blog/2024/09...
Causal inference, social media, and mental health, part 1
This is part 1 of a ? part series
sjwild.github.io
September 21, 2024 at 9:00 PM
Reposted by corystat.bsky.social
dmcneish.bsky.social
New paper w/ Roy Levy

These 2 path diagrams are conceptually different (regression with a latent predictor that has 3 indicators vs. a factor model with 4 indicators) but they are mathematically equivalent.

journals.sagepub.com/doi/10.3102/...

/1
Two path diagrams are shown. On the left side, the conceptual idea is a latent variable with three indicators predicting an observed outcome. On the right side, the conceptual idea is a latent variable with four indicators.
September 25, 2024 at 9:29 AM
Reposted by corystat.bsky.social
rmcelreath.bsky.social
One un-questioned research practice (UQRP) that does as much harm as QRPs is ignoring sampling bias and selection effects. This is a nice paper that shows one way to be transparent and analytical about the problem. "Graphical Causal Models for Survey Inference" doi.org/10.1177/0049... #stats 🧪
Figure 3. Three graphs where auxiliary variable must be used to recover , using external data on . Left: blocks all paths between and . Intermediate: unobserved confounders do not impede recoverability as long as they do not directly influence . An impact of on is also inconsequential. Right: may be influenced by , instead of the other way around. This too is inconsequential. However, in the intermediate and right graph, adding any additional arrow between or and impedes recoverability.
September 30, 2024 at 7:13 AM
Reposted by corystat.bsky.social
paulbuerkner.com
Drafts of the first chapters of my brms Book: Applied Bayesian Regression Modelling Using R and Stan are online: paulbuerkner.com/software/brm... Check it out and let me know what you think!
Paul Bürkner - The brms Book
paulbuerkner.com
October 1, 2024 at 2:36 PM