#statstab #461 The One-Sided P-Value Paradox Thoughts: "I have concluded the stronger proposition, probability of heads > ½, but I cannot conclude the weaker proposition, probability of heads > ½ or < ½" #onetailed #pvalue #logic #Royall #paradox alexanderetz.com/2015/04/20/p...

#statstab #460 Invalid Conclusions Built on Statistical Errors Thoughts: It's easy to make analysis mistakes, providing code should be the norm. #analysis #pseudoreplication #error #design #code #guide #educational steveharoz.com/blog/2024/wr...

#statstab #459 Compute Summary Statistics w {rstatix} Thoughts: Useful package to report descriptive statistics for your data and models. #rstats #r #dataanalysis #statistics #stats #apa rpkgs.datanovia.com/rstatix/refe...

#dataviz #rstats #ggplot What are people's thoughts on UpSet plots? Easier way to show multiple comparisons in one plot? Or just more plot clutter? share.google/ggr8YOogsnBt...

I took this advice to heart, and will have a few papers coming out this year (and some more informal blogs) with bespoke stats tutorials for forensic and Investigative Psychologists.

Interesting question. I can't really see a context-agnostic scenario for a study. So I guess 1. It depends on the study, 2. I guess positive is the psychological default 3. Might not matter if non directional hyp testing is the goal.

#statstab #458 There is still only one test Thoughts: Downey make a good point that all/any testing ensemble does essentially the same thing. #randomization #testing #nhst #pvalue allendowney.blogspot.com/2016/06/ther...

#statstab #457 Race as a Bundle of Sticks: Designs that Estimate Effects of Seemingly Immutable Characteristics Thoughts: The theoretical framework a researcher uses will affect the causal inference they can make. #estimand #causalinference #rubin www.annualreviews.org/content/jour...

New clout metric to chase just dropped for #2026 (I guess I need to post more controversial stuff to get my engagement up) Bluefacts.app

#statstab #456 Bestiary of Questionable Research Practices in Psychology Happy new year! ✨ Thoughts: Starting the year with a resource to make you more cynical in 2026. #QRPs #phacking #metascience #research #ethics journals.sagepub.com/doi/10.1177/...

Journal article: “the data are available on request” The data:

#statstab #455 False Discovery Rate (FDR) and q-values Thoughts: The q-value of a test is the expected proportion of false positives among all hypotheses with p-values as small or smaller than that test. #pvalues #qvalues #FDR #FWER #error www.nonlinear.com/support/prog...

#statstab #454 A very ggplot2 Christmas Thoughts: Tis the festive season. 🎄 #ggplot #rstats #xmas #xmastree t-redactyl.io/blog/2016/12...

Watch out many-analyst/many-labs. There's a new kid on the block. Seems "adversarial collaboration" is the hip new thing to do in 2026.

#statstab #453 Gradient plots {ggdist} Thoughts: Like these as an alternative to half eye plots. #bayesian #dataviz #plots #figures #rstats #r #tidybayes #ggdist mjskay.github.io/ggdist/artic...

#statstab #452 Plotting Distributions in R Thoughts: A small shiny app for plotting Distributions. Useful. #rstats #shiny #dataviz #probability #distributions #r #normal #quantiles bryer.org/posts/2025-0...

Grading and googling hallucinated citations, as one does nowadays, and now that LLMs have been around for a while, I've discovered new horrors: hallucinated journals are now appearing in Google Scholar with dozens of citations bc so many people are citing these fake things

#statstab #451 Log-normal Distribution – A simple explanation Thoughts: A simple and useful source for the log-normal distribution, which has many uses (as I will show in 2026 🤫) #lognormal #distribution #exponentional #guide #pedagogy towardsdatascience.com/log-normal-d...

Another day, and another reference to these bangers. Contemporary causal inference has fundamentally changed the way I think about control variables and mediation analyses. These can show you why: doi.org/10.1177/2515... doi.org/10.1111/spc3... doi.org/10.1177/2515...

#statstab #450 Kendall's Tau {video} Thoughts: understanding nonparametric correlations can be difficult. This video is quite ok at explaining it. #kendall #tau #correlation #nonoarametric #ranking #rank numiqo.com/tutorial/ken...

#stats PSA: In January, I will host a workshop on "e-value". If you care about confounding, causal inference, and sensitivity analyses this is *not* for you; you are thinking of "E-value". Different thing. Different workshop. Isn't stats naming fun? What are ppl's thoughts on either of these two?

#statstab #479 Uncertainty limits the use of power analysis Thoughts: Frequentists avoiding uncertainty is never good. #poweranalysis #error #uncertainty #simulation #cohend #effectsize www.researchgate.net/publication/...

Today we hit 500 followers! So what better time to showcase our wonderful community of researchers, research enablers, institutions, and research sector stakeholders, with the UK Reproducibility Network starter pack. Give them a follow too! #AcademicSky #Research #OpenResearch

#statstab #478 Stuck between Zero and One: Modelling Non-Count Proportions with Beta and Dirichlet Regression Thoughts: A few resources for proportion data. #Dirichlet #proportions #betareg #count #tutorial methodsblog.com/2019/08/06/b...

#statstab #477 Simulating data for Dirichlet regression with varying estimates Thoughts: Interesting thread about an underused model. #Dirichlet #brms #poweranalysis #rstats #proportions #stan #forum discourse.mc-stan.org/t/simulating...

More research like this on other psychometric scales is needed (although the IAT should just be abandoned regardless). Researchers and practitioners should know the precision of their tools. If someone is 1pt over a cutoff how reliable is this for classification and treatment?

#statstab #476 In linear regression, when is it appropriate to use the log of an independent variable instead of the actual values? Thoughts: Great thread explaining the (few) instances when log transforming makes sense #logarithm #transformation #data stats.stackexchange.com/questions/29...

How often do you see in an intro section the authors making a clear distinction bw causal papers they cite and observational ones? All seem to be magically causal. 🤔

I've been exposed to enough peer review that I think we can look at two classes of reviewers: 1. Constructive peer review 2. Adversarial peer review I'm looking at this mostly from a psychology/methodology perspective (but wonder what other fields experience) 🧵 1/

While adoption of open science/research hadn't been ubiquitous, I think many researchers just realised the limitations of old approaches and said "ok, I need to do slightly better if I want to not fool myself with my studies". This is having a clear benefit.

#statstab #475 {tinyplot} A lightweight extension of the base R graphics Thoughts: Cool package for easy plotting of simple data. #dataviz #ggplot #rstats #r #plots #figures #graphs grantmcdermott.com/tinyplot/ind...

#statstab #474 Linear Models with Heterogeneous Coefficients Thoughts: Sometimes you need more complicated models even if identification gets messy. #heterogeneity #modelling #nonlinear #economics #econometrics vladislav-morozov.github.io/econometrics...

#statstab #473 Synthesis without meta-analysis (SWiM) in systematic reviews Thoughts: An interesting proposition for when you don't have good data for a meta-analysis. #synthesis #metaanalysis #systematicreview www.bmj.com/content/368/...

#statstab #472 Practical Bayesian Inference in Neuroscience: Or How I Learned To Stop Worrying and Embrace the Distribution Thoughts: Good intro to bayesian inference. #bayesian #likelihood #tutorial www.biorxiv.org/content/10.1...

#statstab #471 Give Your Hypotheses Space! Thoughts: "each hypothesis requires its own model" + "Only interpret the output for your exposure of interest" #causalinference #modelling #hypothesis #tutorial #confounding #mbias brian-lookabaugh.github.io/website-bria...

#statstab #470 {nmfspalette} palette according to NOAA Fisheries branding Thoughts: Idk how I came across this package, but it has some ok colour combos. #ggplot #dataviz #color #colour #themes #fish nmfs-ost.github.io/nmfspalette/

#statstab #469 Significance tests, p-values, and falsificationism Thoughts: An "interesting" thread of philosophy of science and debating. #discussion #debate #pvalue #falsification #nhst discourse.datamethods.org/t/significan...

#statstab #468 Why Summaries of Research on Psychological Theories are Often Uninterpretable Thoughts: psychology uses significance testing backwards relative to hard sciences. #meehl #theory #criticism #falsification #popper #methods #nhst #psychology scienceplusplus.org/metascience/...

#statstab #467 Hypothesis testing, model selection, model comparison some thoughts Thoughts: An excellent (but too short) discussion on bayesian inference. #bayesian #bayesfactor #modelselection #inference #NBHT #BF #ROPE #primer discourse.mc-stan.org/t/hypothesis...

#statstab #466 Bayesian workflow: Prior determination, predictive checks and sensitivity analyses Thoughts: Having a good bayesian work flow can be challenging with complex models. #priors #bayesian #sensitivityanalysis #posterior #ppc #brms pablobernabeu.github.io/2022/bayesia...

✒️ New blog post by Mircea Zloteanu @mzloteanu.bsky.social 💡 Post-publication reviewing - some reflections reproducibilitea.org/jc/2025/11/2...

#r #bayes Q: is there a way to easy plot a contrast density using just the {easystats} package library? I just want a simple stats_halfeye() plot from estimate_contrasts() 🤷‍♂️

Moral: Be careful what you ask for.

#statstab #465 How to embrace variation and accept uncertainty in linguistic and psycholinguistic data analysis Thoughts: An accessible paper on communicating your results with nuance. #bayes #bayesian #uncertainty #error #bias #guide #tutorial sites.stat.columbia.edu/gelman/resea...

A new paper shows that you can use LLMs to fake survey data - not just successfully bypassing all reasonable checks used by survey researchers but also allowing you to instruct the LLM to fake public opinion in any direction you choose. It’s a strong challenge to any trust we want to have in […]

#rstats #dataviz Q: Does anyone have any code they can share for easily making a plot like this (sensitivity to priors for your model)?

#statstab #464 Plotting p-check interaction {brms} Thoughts: Annoyingly #brms doesn't natively allow plotting for interactions (that I know of). The forum has a solution. #ppc #posterior #bayesian #modelfit #diagnostic #rstats #r #stan discourse.mc-stan.org/t/plotting-p...

I spent 3 days meticulously trying to specify and justify my priors, only to end up with normal(0,1) being very appropriate...

#statstab #463 {modelbased} Understanding your models Thoughts: A deceptively simple case study on how to understand and report your model. #rstats #modelling #easystats #r #reporting easystats.github.io/modelbased/a...

It was an interesting and (semi-) fun experience! Based on the feedback most teams provided so far, it does seem there has been an improvement in how well research is documented and computationally reproduced.