#statsSky
LightNews LightNews
adam-lg.bsky.social
It’s amusing that “Structural Equation Model” can refer to either:

1) a nonparametric model, y~f(x_1, x_2,… x_n), à la Pearlian DAGs, or

2) a massively parametric model, where every relationship is parameterized with a linear/factor model (the psychometrics approach).

#statssky #causalsky
February 19, 2026 at 3:02 AM
noahgreifer.bsky.social
I'm so excited to announce the first release of my newest #Rstats package, {adrftools}! This package facilitates estimation, visualization, and testing for the causal effect of a continuous (i.e., non-discrete) treatment.

🧵 1/10

#statssky #episky #causalinference
adrftools: Estimating, Visualizing, and Testing Average Dose-Response Functions
Facilitates estimating, visualizing, and testing average dose-response functions (ADRFs) for characterizing the causal effect of a continuous (i.e., non-discrete) treatment or exposure. Includes suppo...
cran.r-project.org
February 18, 2026 at 4:05 PM
f2harrell.bsky.social
#Statistics thought of the day: If you think you can find new disease subtypes by empirically clustering patients, think again: www.fharrell.com/post/cluster... #StatsSky
The Burden of Demonstrating Statistical Validity of Clusters – Statistical Thinking
Patient clustering, often described as the finding of new phenotypes, is being used with increasing frequency in the medical literature. Most of the applications of clustering of observations are not ...
www.fharrell.com
February 18, 2026 at 12:40 PM
aaronsojourner.org
#EconSky #sociology #PoliSky #PolicySky #NumbersDay #FinTwitter #PsychSky #AcademicSky #PhDSky #metasci #SciencePublishing #ResearchPublishing #stats
#StatsSky #science #Criminology #environment
aaronsojourner.org Aaron Sojourner @aaronsojourner.org · 1d
The American Statistical Association has an organizational sign-on letter urging Congress to exempt federal statistical agencies & units from Schedule Policy/Career.

Are you a member in any orgs that value accurate, timely info about the U.S.? Please consider signing
docs.google.com/forms/d/e/1F...
February 17, 2026 at 9:07 PM
amstatnews.bsky.social
Know someone doing transformative work in official statistics? The ASA is accepting applications for the new Katherine K. Wallman Award, honoring impact on federal, state, or local statistics. $2,500 honorarium. Apply by March 1: tinyurl.com/56ne8a9b #STATSSKY #STATISTICS
A gold star award on a navy blue background with gold confetti
February 17, 2026 at 6:30 PM
amstatnews.bsky.social
Announcing the ASA’s Dr. Jean Dickinson Gibbons Endowed Scholarships—an outstanding opportunity for PhD students in statistics pursuing applied research. Up to four $10,000 scholarships are awarded annually. Learn more & apply by March 1: tinyurl.com/563394j7 #STATSSKY #STATISTICS
A textbook with a grad cap and some coins.
February 17, 2026 at 5:01 PM
f2harrell.bsky.social
2/2 Our paper shows that despite clinical trials having to use larger sample sizes to compensate for the huge information loss from using binary outcomes, the resulting power is STILL LOWER than the power of SMALLER clinical trials using continuous outcomes. #Statistics #StatsSky #rct #clinicaltrial
February 16, 2026 at 6:20 PM
f2harrell.bsky.social
1/2 Nail in the coffin of dichotomania: @erik-van-zwet.bsky.social 's paper with @stephensenn.bsky.social and myself just published: onlinelibrary.wiley.com/doi/10.1002/... with extended discussion at discourse.datamethods.org/t/dichotomiz... #Statistics #StatsSky #rct #clinicaltrial
onlinelibrary.wiley.com
February 16, 2026 at 6:20 PM
f2harrell.bsky.social
Wow this is a breakthrough in validation of predictive models that fully accounts for point estimates used in predictions being only point estimates. Great work @richarddriley.bsky.social and colleagues! #Statistics #StatsSky
richarddriley.bsky.social Richard Riley (R²) @richarddriley.bsky.social · 3d
⭐ NEW PAPER ⭐ on a Bayesian approach to sample size calculations for external validation of risk prediction models - account for uncertainty in the assumed true performance of the model - plus calculate assurance probabilities & the value of information
onlinelibrary.wiley.com/doi/10.1002/...
February 16, 2026 at 12:45 PM
konsta.happonen.eu
Roses are red
Swallows, flying, are blue blurs
So sorry this is frequentist
I should have used posteriors

#statssky
konsta.happonen.eu Konsta Happonen @konsta.happonen.eu · Feb 14
Happy Valentine's Day (p<0.05). #rstats
## A scatterplot of a heart tilted 45° to the right with a regression line overlaid. The title says Happy Valentine's Day with p < 0.05 library(tidyverse) library(ggview) library(extrafont) ## Sample from a heart curve t <- seq(0,2*pi,length.out = 400) x <- 16*sin(t)^3 y <- 13 * cos(t) - 5 * cos(2*t) - 2*cos(3*t) - cos(4*t) dd <- matrix(c(x,y),ncol=2) # Rotate 45 degrees theta <- pi*-0.25 rotmat <- matrix(c(cos(theta),sin(theta),-sin(theta),cos(theta)), ncol = 2) dd2 <- t(rotmat %*% t(dd)) colnames(dd2) <- c("x","y") dd2 <- dd2 |> as_tibble() # Check p-value lm(y~x, data = dd2) |> summary() # Publish or perish p_heart <- dd2 |> ggplot(aes(x,y)) + geom_point(size = 0.01) + stat_smooth(formula = y~x,method = "lm", color = "firebrick", fill = "hotpink") + coord_fixed() + theme_minimal(base_family = "Karla") + theme(panel.grid.minor = element_blank()) + labs(title = "Happy Valentine's Day", subtitle = "with p < 0.05", x = NULL, y = NULL) + canvas(width = 4, height = 4, dpi = 300) p_heart |> save_ggplot("sydän/sydän.png", device = png)
February 14, 2026 at 7:16 AM
datavisfriendly.bsky.social
#statssky
Oh, it has been such a long time since someone was writing about statistics and psychedelics. Rocky PCA High 🎶
February 13, 2026 at 9:48 PM
konsta.happonen.eu
So, I'd like to read something about stochastic process models, specifically brownian motion time-series stuff, preferably from a Bayesian point of view. Does anybody have any good recommendations? #statssky
February 13, 2026 at 7:34 PM
amstatnews.bsky.social
Get to know the candidates! Tune in to a virtual town hall event hosted by ASA President Jeri Mulrow in advance of the 2026 ASA election. Candidates James Cochran, Nicholas Horton, Martha Gardner, & Natalie Rotelli will make statements & take your questions. Join us: tinyurl.com/58s9n9em #STATSSKY
February 12, 2026 at 7:04 PM
ntl.no
Endelig mer press for statssky. ☁️

NTL og Fagforbundet krever fortgang i statlig eid og drevet skytjeneste. Stortinget vedtok dette i 2021, men regjeringa har åpna for kommersiell løsning. Nå er saken løfta igjen, med forslag om mer fart og mindre avhengighet.
February 12, 2026 at 4:19 PM
konsta.happonen.eu
I used to think Bayes factors are suspicious because of how they're very sensitive to the prior. But what if that's a feature, not a bug? If the results of an analysis depend heavily on small changes in methods, maybe the results aren't robust enough to warrant interpretation? #statssky
February 12, 2026 at 8:36 AM
drjeffreyb.bsky.social
Not surprised at the dearth of null results. Modeling suggests significant results are 10 to 100 times more likely to be published. Structural bias as few celebrate or get tenure on confirming the nul. Publication bias is ascendant in virtually every discipline, not just poly sci! #statssky #episky
It must be very hard to publish null results Ryan C. Briggs, Jonathan Mellon, Vincent Arel-Bundock* 2026-02-11 Publication practices in the social sciences act as a filter that favors statistically significant results over null findings. While the problem of selection on significance (SoS) is well-known in theory, it has been difficult to measure its scope empirically, and it has been challenging to determine how selection varies across contexts. In this article, we use large language models to extract granular and validated data on about 100,000 articles published in over 150 political science journals from 2010 to 2024. We show that fewer than 2% of articles that rely on statistical methods report null-only findings in their abstracts, while over 90% of papers highlight significant results. To put these findings in perspective, we develop and calibrate a simple model of publication bias. Across a range of plausible assumptions, we find that statistically significant results are estimated to be one to two orders of magnitude more likely to enter the published record than null results. Leveraging metadata extracted from individual articles, we show that the pattern of strong SoS holds across sub-fields, journals, methods, and time periods. However, a few factors such as pre-registration and randomized experiments correlate with greater acceptance of null results. We conclude by discussing implications for the field and the potential of our new dataset for investigating other questions about political science.
February 12, 2026 at 4:35 AM
amstatnews.bsky.social
Nominate a colleague for the ASA Founders Award! Individuals selected will have served the association for an extended time in various leadership roles and contributed significantly to the ASA. Deadline is March 1. #STATSSKY www.amstat.org/your-...
A gold trophy.
February 11, 2026 at 3:00 PM
editoratlarge.bsky.social
Statisticians beware #StatsSky
thefinancenewsletter.com Andrew Lokenauth | TheFinanceNewsletter.com @thefinancenewsletter.com · 8d
Microsoft released a study showing the 40 jobs most at risk by AI:
February 11, 2026 at 1:08 PM
amstatnews.bsky.social
ASA’s Ron Wasserstein is also the cohost of the “Practical Significance” podcast. Each episode highlights work shaping statistics & real lives—and your ideas matter. Watch Ron’s latest video, contact him with ideas at ron@amstat.org, & listen to the latest episode: tinyurl.com/bd53233f. #STATSSKY
February 10, 2026 at 2:32 PM
amstatnews.bsky.social
Calling all students! Want to use your statistical skills to make a difference? StatCom is a student-run group providing pro bono statistical consulting to community organizations. Learn more at a StatCom info session Friday, Feb. 27, at 2 p.m. ET. tinyurl.com/2p2pykrh #STATSSKY #STATISTICS
February 9, 2026 at 6:30 PM
ajdaajournal.bsky.social
#PsychSciSky #DrugSky #PharmSky #MedSky #AddictionSci #NeuroSky #EpiSky #StatsSky #HealthPolicy @tandfresearch.bsky.social
February 9, 2026 at 5:18 PM
amstatnews.bsky.social
Who should attend #SDSS2026? Statisticians & data scientists pushing the edge of AI and beyond. Collaborate across disciplines, learn practical tools, grow your network, and help shape responsible AI in Milwaukee, WI, April 28–May 1. Register early & save! tinyurl.com/2pbece9b #STATSSKY
Advancing AI, statistics, and data science together.
February 9, 2026 at 3:00 PM
konsta.happonen.eu
Lighting the Bayesian bat signal: #stan #rstats #statssky

A latent construct C is informed by multiple variables X like so:

X₁, X₂ ~ C

I want to use C to also predict Y,

Y ~ C

but without C being informed by Y. What's the cleanest way to implement this in Stan?
February 7, 2026 at 9:49 AM
f2harrell.bsky.social
Semiparametric models reduce to empirical CDF when there are no covariates. They substitute an assumption about relative distribution shapes for a much stricter assumption about absolute distribution shapes used in parametric models. hbiostat.org/ordinal #Statistics #StatsSky
Ordinal Regression
hbiostat.org
February 6, 2026 at 12:27 PM
adam-lg.bsky.social
"Markov Random Fields: Structural Properties, Phase Transition, and Response Function Analysis." An interesting-looking review of a common tool in spatial statistics

Paper: arxiv.org/abs/2602.02771
#statssky
This paper presents a focused review of Markov random fields (MRFs)--commonly used probabilistic representations of spatial dependence in discrete spatial domains--for categorical data, with an emphasis on models for binary-valued observations or latent variables. We examine core structural properties of these models, including clique factorization, conditional independence, and the role of neighborhood structures. We also discuss the phenomenon of phase transition and its implications for statistical model specification and inference. A central contribution of this review is the use of response functions, a unifying tool we introduce for prior analysis that provides insight into how different formulations of MRFs influence implied marginal and joint distributions. We illustrate these concepts through a case study of direct-data MRF models with covariates, highlighting how different formulations encode dependence. While our focus is on binary fields, the principles outlined here extend naturally to more complex categorical MRFs and we draw connections to these higher-dimensional modeling scenarios. This review provides both theoretical grounding and practical tools for interpreting and extending MRF-based models.
February 5, 2026 at 12:53 AM