# statistical (blogs) — RSS Amplifier Recent posts from the 20 feeds in the RSS Amplifier directory that cover statistical. Page: Feed: --- ## [The P-value Is Not a Verdict: An Interactive Consonance Curve](https://lesslikely.com/statistics/consonance-explorer.html) _2026-08-16 · Zad Rafi · Less Likely Blog_ Drag an estimate and its interval and watch the entire P-value function, every compatibility interval, and the S-value move together. ## [How to Be a Good Collaborator to a Statistician and a Data Engineer](https://lesslikely.com/statistics/good-collaborator.html) _2026-08-11 · Zad Rafi · Less Likely Blog_ Good collaboration with statisticians and data engineers is a set of concrete habits, not goodwill. What each side owes the other, with checklists. ## [Symmetric vaccine efficacy](/writing/2026_symmetric_vaccine_ef/) _2026-08-11 · Lucy D'Agostino McGowan_ Traditional measures of vaccine efficacy (VE) are inherently asymmetric, constrained above by 1 but unbounded below. As a result, VE estimates and corresponding confidence intervals (CIs) can extend far below zero, making interpretation difficult and potentially obscuring whether the apparent effect reflects true harm or simply statistical uncertainty. The proposed symmetric vaccine efficacy (SVE)… ## [Modernizing Clinical Trial Design and Analysis to Improve Efficiency & Flexibility](https://fharrell.com/talk/rcteff/) _2026-08-10 · Frank Harrell · Statistical Thinking_ UCLA Cardiology Grand Rounds 2020-10-23 | Video (better video below) Vanderbilt University Department of Biostatistics 2020-11-18 Vanderbilt Translational Research Forum 2021-11-04 | Video Consilium Scientific 2024-03-14 | Video and here Seventh Annual Janice Pogue Lectureship in Biostatistics , Population Health Research Institute, Hamilton, Ontario, Canada 2024-12-05 National Heart and Lung… ## [Parametric Models vs. Empirical Distributions and Ordinal Regression](https://fharrell.com/post/ecdf/) _2026-08-09 · Frank Harrell · Statistical Thinking_ Background In modeling data we frequently need to estimate more than a mean. We may want to estimate quantiles, dispersion, or tail probabilities. These require model assumptions to be correct when the model is a smooth parametric one. Even when estimating a mean, for inference to be accurate we need distributional assumptions we make to be satisfied. The central limit theorem offers no protection… ## [Symmetric Vaccine Efficacy: Interpretable Estimation and Inference for Vaccine Trials](/talk/asa_joint_statistical_meeting_2026/) _2026-08-04 · Lucy D'Agostino McGowan_ ## [Statistical Models Answer the Fundamental Clinical Question and Provide Clinical Trial Estimands](https://fharrell.com/post/rctmod/) _2026-07-27 · Frank Harrell · Statistical Thinking_ Goals and estimands should flow from the study design while respecting the lack of exchangeability of the participants and leading to estimators with data-verifiable assumptions. Statistical models are approximations of reality. It is not rational to fear transparent assumptions required by them. Avoiding models will almost surely lead to worse results. It is impossible to model an individual… ## [Data leakage and independence](http://notstatschat.rbind.io/2026/07/27/data-leakage-and-independence/) _2026-07-27 · Biased and Inefficient_ Derek Lowe, writing on models for protein-ligand interaction Ideally, you train your model on a big pile o’ protein data, while holding out a good piece of it to keep it out of the training set. Then once your software has machine-learned its way to happiness, you see how it performs on those cases that it’s never seen before. But here’s the problem: it is very hard to set this up in a way so that… ## [The Unifying Capabilities of Cumulative Probability Semiparametric Models](https://fharrell.com/post/cpm/) _2026-07-26 · Frank Harrell · Statistical Thinking_ Disclaimer: No data were harmed during the making of this article. Any use of binning was strictly prohibited. Background Consider the oldest statistical regression model, the linear model, named from the fact that it is linear in the parameters . Suppose that the outcome variable is continuous, and conditional on a vector of baseline descriptors has a normal distribution. Let denote a vector of… ## [Blog 27 – Estimands Part 3B – What Is the ITT Estimand Anyway?](https://analytixthinking.blog/2026/07/24/blog-27-estimands-part-3b-what-is-the-itt-estimand-anyway/) _2026-07-24 · Steve Ruberg · (untitled)_ It is difficult to define the intention-to-treat (ITT) effect, and thereby the ITT estimand, since the use of the ITT approach can involve treating patients with many different treatments for many different time periods. Because of this mixture of treatment effects, the ITT approach is more of a HOW for analyzing data, than a WHAT for defining the treatment effect of interest. ## [Hire AI-Powered Professionals (Sponsored)](https://crawlproof.com/a/pcUKU3bGbcgC) _2026-07-24 · **Sponsored**_ Browse public gigs, message candidates, and run video interviews — free to start. ## [Random Selection and Democracy in Athens and Asimov](https://www.sharonlohr.com/blog/2026/7/20/random-selection-and-democracy) _2026-07-20 · Sharon Lohr · Blog - Sharon Lohr_ How do methods for choosing political candidates relate to sampling principles? Here are some class activities and discussion questions. ## [A Statistical FX Factor Model](https://dm13450.github.io/2026/07/15/statistical-factor-model.html) _2026-07-15 · Dean Markwick · Dean Markwick_ Factor models attempt to explain asset returns. You can approach this in two ways: define the factors you think are relevant, or use statistical learning to build the relevant factors from the data. My previous post took the first approach. This post will use principal component analysis (PCA) to let the data tell us which factors are most relevant in an FX factor model. ## [Blog 26 – Estimands Part 3A – What Is the Question?](https://analytixthinking.blog/2026/07/14/blog-26-estimands-part-3a-what-is-the-question/) _2026-07-14 · Steve Ruberg · (untitled)_ Even with the publication of ICVH E9(R1) and so may efforts globally for explaining and implementing the estimand framework, too often statisticians, and clinical researchers in general, focus their efforts on how to handle incomplete data, which is my term for the missing or confounded data created by so-called intercurrent events. Let’s get the focus on the clinical question of interest. ## [Thomas Jefferson and Population Estimates](https://www.sharonlohr.com/blog/2026/7/4/thomas-jefferson-population-estimates) _2026-07-04 · Sharon Lohr · Blog - Sharon Lohr_ Historian James Cassedy wrote that “probably no one in America was more imbued with the statistical spirit than was Jefferson.” Let’s look at two of his statistical activities. ## [No. 25 – Estimands Part 2 – What Exactly Is an Estimand?](https://analytixthinking.blog/2026/07/02/no-25-estimands-part-2-what-exactly-is-an-estimand/) _2026-07-02 · Steve Ruberg · (untitled)_ The word estimand has exploded into the vernacular of statistical thinking for clinical trials in the last decade or so. Where did it come from and why is it so important now? ## [No. 24: Estimands Part 1 – Just Do ITT?](https://analytixthinking.blog/2026/06/28/blog-24-estimands-part-1-just-do-itt/) _2026-06-28 · Steve Ruberg · (untitled)_ Randomization and complete data on the randomized experimental units are needed for cause-and-effect inference. What happens when that logic is broken? ## [The Octagon Is a One-Trial Experiment](https://lesslikely.com/statistics/mma-single-event.html) _2026-06-23 · Zad Rafi · Less Likely Blog_ A single MMA fight is not a probability. But repeat it ten times and the better fighter’s edge starts showing. That gap is exactly what frequentism was built for. ## [The simple case of almost sure representations](http://notstatschat.rbind.io/2026/06/16/the-simple-case-of-almost-sure-representations/) _2026-06-16 · Biased and Inefficient_ The almost-sure representation theorem in probability says that if \\(X\_n\\stackrel{d}{\\to}X\\) we can find a sequence \\(\\tilde X\_n, \\tilde X\\) with the same distributions as \\(X\_n,X\\), possibly on a different probability space, so that \\(\\tilde X\_n\\stackrel{a.s.}{\\to}\\tilde X\\) (where the a.s. is with respect to the distribution of \\(X\\)). General versions of this theorem are a bit tricky to prove,… ## [AI and Survey Sampling Problems](https://www.sharonlohr.com/blog/2026/6/10/ai-and-sampling-problems) _2026-06-12 · Sharon Lohr · Blog - Sharon Lohr_ How does Gemini perform on problems in survey sampling, and how might one use it in the classroom? ## [Using large language models to enhance clinically-driven missing data recovery algorithms in electronic health records](/writing/2026_using_large_language/) _2026-06-02 · Lucy D'Agostino McGowan_ Objectives Electronic health record (EHR) data are prone to missingness and errors. Previously, we devised an enriched chart review protocol where a “roadmap” of auxiliary diagnoses was used to recover missing values. Still, chart reviews are expensive and time-intensive, limiting the number of patients whose data can be reviewed. Now, we investigate the accuracy and scalability of a… ## [Realtime AI Arbitrage Terminal (Sponsored)](https://crawlproof.com/a/bNUBcwUg8ZIA) _2026-06-01 · **Sponsored**_ Realtime auto-trading across exchanges with scored setups, backtesting, and risk controls. ## [Implications of the Draft FDA Bayesian Guidance](https://fharrell.com/talk/bguide/) _2026-05-21 · Frank Harrell · Statistical Thinking_ Events AstraZenica Global Statistical Forum and Washington Statistical Society 2026-05-21 Slides Video ## [Thoughts About the Roles of AI for Statistics](https://fharrell.com/talk/ai/) _2026-05-19 · Frank Harrell · Statistical Thinking_ Events Regression Modeling Strategies Course 2026 2026-05-19 Slides ## [The Joys of Free Cloudflare](https://dm13450.github.io/2026/05/18/the-joys-of-free-cloudflare.html) _2026-05-18 · Dean Markwick · Dean Markwick_ I’ve been tinkering around with the free tier on Cloudflare and have managed to churn out a couple of side projects. Of course, I had a little help with various AI systems, but it was assisted rather than vibe-coded. ## [Mind the Gap: Causal Inference is Not Just a Statistics Problem](/talk/keynote_agstat/) _2026-05-12 · Lucy D'Agostino McGowan_ ## [Causal Inference in R](/talk/agstat_2026/) _2026-05-11 · Lucy D'Agostino McGowan_ ## [Causal Inference in R](/workshops/agstat_2026/) _2026-05-11 · Lucy D'Agostino McGowan_ ## [About](https://joeconway.com/about/) _2026-05-10 · mail@joeconway.com (Joseph Conway) · Joseph Conway_ I’m a long-time PostgreSQL community member and contributor. I authored PL/R , the PostgreSQL procedural language for R, and have spent most of my career working on the database from the inside out. I currently work at AWS . The most up-to-date professional summary is on my LinkedIn profile . Elsewhere Code on GitHub Past and current talks under Presentations Notes and downloads for PL/R… ## [Hello from the new site](https://joeconway.com/posts/2026/05/hello-from-the-new-site/) _2026-05-10 · mail@joeconway.com (Joseph Conway) · Joseph Conway_ After many years on a hand-edited HTML5UP template, joeconway.com is reborn as a Hugo site. ## [PL/R Notes](https://joeconway.com/plr/notes/) _2026-05-10 · mail@joeconway.com (Joseph Conway) · Joseph Conway_ User documentation, how-tos, and tips for PL/R. The primary source repository is on GitHub . Important notes when using PL/R Define R\_HOME so that $R\_HOME/bin is in your $PATH . Ensure all R library DLLs and any other dependent shared libraries are also in your $PATH . A particularly good walkthrough, especially for Windows users, lives at BostonGIS . Examples A graphing example using files PL/R… ## [Miss(ing) Congeniality: The Cost of Incompatible Imputation Models](/talk/boston_university_2026/) _2026-04-28 · Lucy D'Agostino McGowan_ ## [Trade Smarter With IBKR (Sponsored)](https://crawlproof.com/a/w7JrUe9w7XEH) _2026-04-28 · **Sponsored**_ Join Interactive Brokers through a referral and access low-cost global trading. ## [The Role of Congeniality in Multiple Imputation for Doubly Robust Causal Estimation](/talk/distinguished_lecture_at_innovations_in_design_analysis_and_dissemination_2026/) _2026-04-20 · Lucy D'Agostino McGowan_ ## [A Fundamental FX Factor Model](https://dm13450.github.io/2026/04/19/A-Fundamental-FX-Factor-Model.md.html) _2026-04-19 · Dean Markwick · Dean Markwick_ I’ve been reading The Elements of Quantitative Investing to branch out from my usual high-frequency finance to something slower or mid-frequency. Factor models are a big part of this quant topic, and I’m trying to get a deeper understanding by following the book and applying the process to FX data. ## [Stage vs phase, again](http://notstatschat.rbind.io/2026/04/17/stage-vs-phase-again/) _2026-04-17 · Biased and Inefficient_ As I have said before, the distinction between multistage and multiphase surveys is mostly important as a shibboleth for survey statisticians. The two=phase quantities \\\[\\pi\_i^\*=P(i\\text{ sampled}|\\text{phase 1})P(i\\in\\text{phase 1})\\\] are not the same as the marginal sampling probabilities \\\[\\pi\_i=P(i\\text{ sampled})\\\] but one uses them in exactly the same formulas, so the difference doesn’t… ## [Questions We Forget To Ask When Designing an RCT](https://fharrell.com/talk/qrct/) _2026-04-16 · Frank Harrell · Statistical Thinking_ Events JHU Trial Innovation Center DIDACT Symposium 2026-04-16 Slides ## [Exploring the MET API with Python - Francisco Goya’s Artworks](https://www.codingthepast.com/2026/04/16/met-api-with-python.html) _2026-04-16 · coding-the-past_ Discover how to retrieve data from the MET API using Python. Convert complex JSON data into pandas DataFrames and create a visualization of the most frequent terms in Francisco Goya's artwork tags. ## [Are predictive models enough?](http://notstatschat.rbind.io/2026/04/13/are-predictive-models-enough/) _2026-04-13 · Biased and Inefficient_ In one of the social media discussions about causal inference the suggestion was made that predictive models are all you need: a good predictive model gives you all the conditional distributions you could want, and you don’t need any special causal inference stuff. I think there’s something to this point of view, but there are a few limitations. The first is that causal inference theory (eg causal… ## [Causal Inference in R](/talk/workshop_on_experiments_at_neoma_business_school_reims_france_2026/) _2026-04-09 · Lucy D'Agostino McGowan_ ## [Causal Inference in R](/workshops/workshop_on_experiments_at_neoma_business_school_reims_france_2026/) _2026-04-09 · Lucy D'Agostino McGowan_ ## [Causal by Design](https://fharrell.com/post/causal/) _2026-04-08 · Frank Harrell · Statistical Thinking_ Background Two research methods are being used with increasing frequency: causal inference and target trial emulation . There are many complex situations where a formal causal calculus such as that developed by Judea Pearl is needed to allow one to infer than an effect is caused by a specific variable such as the use of a treatment or exposure to a specific agent. Practitioners of causal calculus… ## [Causal by Design](https://fharrell.com/post/causal/) _2026-04-08 · Frank Harrell · Statistical Thinking_ Background Two research methods are being used with increasing frequency: causal inference and target trial emulation . There are many complex situations where a formal causal calculus such as that developed by Judea Pearl is needed to allow one to infer than an effect is caused by a specific variable such as the use of a treatment or exposure to a specific agent. Practitioners of causal calculus… ## [EM-DAT, the world's disaster memory, is at risk](https://statsandr.com/blog/em-dat-the-world-s-disaster-memory-is-at-risk/) _2026-04-07 · Stats and R_ I do not usually write posts that are calls to action. But sometimes, something important enough comes along that it would feel wrong to stay silent. This is one of those times. What is EM-DAT? EM-DAT, the Emergency Events Database, is the world’s most widely used and trusted global database for tracking natural and technological disasters. It has been maintained since 1988 by the Centre for… ## [Calculating the moments of a loss function](http://erikdrysdale.com/loss_moments/) _2026-04-05 · Erik Drysdale_ In this post I describe how to calculate the first and second moments of a loss function for a machine learning (ML) model. The expected value of a loss function is known as its risk, and I refer to the second central moment as the loss variance. While these quantities are not perfectly knowable in the real world, in the simulation setting the researcher has full knowledge of how data is generated… ## [Bayesian Neural Networks in {tidymodels} with {kindling}](https://statsandr.com/blog/bayesian-neural-networks-in-tidymodels-with-kindling/) _2026-03-20 · Stats and R_ What Are Bayesian Neural Networks? Standard neural networks learn fixed weights during training and produce a single point estimate for each input — with no sense of how confident the model is. Bayesian Neural Networks (BNNs) replace those fixed weights with probability distributions (Neal 2012). Instead of learning a single value per weight, the network learns a mean and a variance, and samples… ## [Making Sense of the DXY](https://dm13450.github.io/2026/03/10/Making-Sense-of-the-DXY.html) _2026-03-10 · Dean Markwick · Dean Markwick_ My day job is in quant trading, but there’s another fascinating world: quantitative investing. While I focus on latencies and execution, quant investors are busy building the most efficient portfolios and ensuring they extract pure alpha. Not one to stay in my lane, I’m using this blog post as an opportunity to dive into the world of quant investing and level up my knowledge. ## [AI vs. Undergraduate Statistics Students](https://www.sharonlohr.com/blog/2026/3/8/ai-vs-statistics-students) _2026-03-08 · Sharon Lohr · Blog - Sharon Lohr_ I submitted some of my undergraduate class assignments to Gemini. How did it do? ## [New in the survey package](http://notstatschat.rbind.io/2026/02/25/new-in-the-survey-package/) _2026-02-25 · Biased and Inefficient_ Version 4.5 of survey is on CRAN now. There are a lot of little changes and a few new features. Thanks to Stas Kolenikov we have Bell-McCaffrey standard errors for svyglm. The standard svyglm standard errors are based on sums of squares of PSU-level residuals. As Bell and McCaffrey point out These sums of squares tend to be too small for two reasons: residuals are generally smaller than true… ## [You can do more for neural networks in R with {kindling}](https://statsandr.com/blog/you-can-do-more-for-neural-networks-in-r-with-kindling/) _2026-02-19 · Stats and R_ Why this post matters Neural networks in R are no longer niche. Today, we can choose among: {nnet} for classic, small-scale neural nets, {neuralnet} another classic neural nets package besides {nnet}, {keras} / {keras3} for the Keras API (typically with Python backends such as TensorFlow/JAX/Torch), {torch} for native-R deep learning with explicit model and training control. So why discuss another… ## [Paper: 'Effectiveness of pneumococcal conjugate vaccines against invasive pneumococcal disease in Vietnamese children prior to national introduction: A matched case-control study'](https://statsandr.com/blog/paper-effectiveness-of-pneumococcal-conjugate-vaccines-against-invasive-pneumococcal-disease-in-vietnamese-children-prior-to-national-introduction-a-matched-case-control-study/) _2026-02-17 · Stats and R_ I am happy to share that an article I contributed to has just been published in Vaccine (Truong et al. 2026): Effectiveness of pneumococcal conjugate vaccines against invasive pneumococcal disease in Vietnamese children prior to national introduction: A matched case-control study 🔗 https://doi.org/10.1016/j.vaccine.2026.128349 Background Streptococcus pneumoniae remains a major cause of severe… ## [nycOpenData: A unified R interface to NYC Open Data APIs](https://statsandr.com/blog/nycopendata-a-unified-r-interface-to-nyc-open-data-apis/) _2026-01-27 · Stats and R_ Guest post by Christian Martinez, developer of the nycOpenData package in R. nycOpenData: A unified R interface to NYC Open Data APIs I am pleased to announce the release of nycOpenData, an R package providing convenient, tidy access to dozens of datasets from the New York City Open Data platform. The package is designed as part of an open-science and reproducible-research effort, with the goal of… ## [Do predictive models need to be causal?](http://notstatschat.rbind.io/2026/01/26/do-predictive-models-need-to-be-causal/) _2026-01-26 · Biased and Inefficient_ This post is partly because I’m about to start teaching generalised linear models and partly to avoid doomscrolling Do predictive models need to be causal? At first glance the response is something like “Of course not; are you high?”. If you have a model \\\[\\log P(Y=1)= \\alpha +\\beta\_X X+\\beta\_Z Z\\\] then whether the model is usefully predictive has almost nothing to do with whether one of…