iiinova — where your digital self keeps evolving

Adrian Olszewski @adrianolsz.bsky.social · 22/03/2025

Two examples of using logistic regression (&friends) for model-based testing hypotheses & estimation. I show the equivalence of Wald's "z" statistic for two proportions with pooled and unpooled variance vs logistic r. + AME / Rao. 1: medium.com/@r.clin.res/... 2: medium.com/@r.clin.res/... #stats

medium.com

Testing hypotheses about proportions (%) through the logistic regression (and friends) — part II

In my previous article titled “Is logistic regression a regression? It has been a regression since its birth — and is used this way every…

031 0

Adrian Olszewski @adrianolsz.bsky.social · 01/03/2025

#stats Sometimes, at work, I feel like I'm the only person who cares about the success of the study...

0120 0

Adrian Olszewski @adrianolsz.bsky.social · 29/01/2025

That's how it sometimes looks like when #stats (here: in clinical #biostatistics) meets #machinelearning with a little 𝘪𝘮𝘱𝘦𝘥𝘢𝘯𝘤𝘦 𝘮𝘪𝘴𝘮𝘢𝘵𝘤𝘩 on some topic common in both places 🙃 (happened to me twice so far).

110 0

Adrian Olszewski @adrianolsz.bsky.social · 28/01/2025

I use both depending on need. Histogram + mandatory rugs plot below for quick exploration of the distributions, as my mind finds it easier. For the assessment of "subtle issues" in assumptions - QQ. Recently I'm trying more the empirical vs. theoretical CDFs plots to see how well quantiles overlap.

000 0

Adrian Olszewski @adrianolsz.bsky.social · 28/01/2025

Far too often hist. fooled me in small data (clinical trials) when assessing assumptions, missing "holes", kurtosis and other issues. Different binning rules (FD, Sturges) only worsened it. The grey region is the pointwise CI for the theoretical N. (here as an example; any necessary can be used).

100 0

Adrian Olszewski @adrianolsz.bsky.social · 30/12/2024

- Adrian, logistic regression is NOT a regression, it's a classifier! - O'rly? I've never used it for classification while using it daily at work to test hypotheses and explore treatment effects😬 I bet #stats guys who invented it didn't know this... revelation too 🙄 #ml #machinelearning #datascience

010 0

Adrian Olszewski @adrianolsz.bsky.social · 18/12/2024

The moment, when you feel the nonsense of 𝙋𝙊𝙎𝙏-𝙃𝙊𝘾 𝙋𝙊𝙒𝙀𝙍 and yet the same concept is essential for👉 𝙙𝙚𝙨𝙞𝙜𝙣𝙞𝙣𝙜 𝙨𝙩𝙪𝙙𝙞𝙚𝙨, namely calculating the sample size: 1) to determine the safety range for dropouts 2) when there's no closed-form formula (like for the Fisher exact test) #statistics #stats #rstats

040 0

Adrian Olszewski @adrianolsz.bsky.social · 15/12/2024

Yes, that's essential for the proper interpretation. Thank you for making this comment for our readers!

030 0

Adrian Olszewski @adrianolsz.bsky.social · 14/12/2024

#statistics #stats Reading QQ plots is not easy at first. I can relate! But don't give up just because it's hard. Everything is hard until you know it and suddenly it becomes your daily tool. Practising and mastering QQ plots (and CDF, BTW) isn't a waste of time! Below I show a few examples.

2315 0

Adrian Olszewski @adrianolsz.bsky.social · 10/12/2024

#rstats #ggplot2 I must say I really *like* the gghalves package and the idea of the #raincloud plots 🥰 In my toolkit (#clinicaltrials) for 8 years. Below - various examples (from toy experiments and real reports). Luckily to me, my collaborators love them too (and even demand!).

0153 0

Adrian Olszewski @adrianolsz.bsky.social · 10/12/2024

Anscombe's quartet, Datasaurus dozen... I wanted to make my own set with same (basic) descriptive summaries. Let me introduce the Olszewski's quartet 😎 Yeah, I could easily make both kurtosis and skewness equal too, but I like the asymmetry of the nose and eyes 😂 github.com/adrianolszew... #rstats

1102 0

Adrian Olszewski @adrianolsz.bsky.social · 05/12/2024

Thank you very much for our kind support! I thought about #statistics or #stats but I was running out of characters. I've never used Twitter/X before (only LinkedIn) and I need to learn how to fit thoughts in 300 characters 😂

210 0

Adrian Olszewski @adrianolsz.bsky.social · 05/12/2024

For ~15yrs on ~weekly basis at work you've been using the logistic regression for exploring main & simple effects + interactions & testing hypotheses, and suddenly a #ML guru tells you that "nope, logistic regression is not a regression, it gives 0/1😎". And pufff! - you disappear like in Disney's💥🤦‍♂️

240 0

Adrian Olszewski @adrianolsz.bsky.social · 05/12/2024

PS: I thought you might find this orange book interesting. It was a big eye-opener to me years ago. The cover in the figure from my post on the LinkedIn, where I have a series on the rank-based methods, like MWW, KW (both special cases of ordinal logistic regression), Brunner-Munzel, ATS/WTS, ART.

000 0

Adrian Olszewski @adrianolsz.bsky.social · 05/12/2024

Testing for stochastic superiority means that they will be sensitive to anything that "induces" the superiority. Effectively it translates also to dispersions (only in some patterns, as you suggested) and shapes. But far too many people (ab)use them for comparing medians naively (only holds for IID)

000 0

Adrian Olszewski @adrianolsz.bsky.social · 04/12/2024

I'm totally and happily! surprised. I thought I'm the only one in the whole Universe who's been doing 𝗱𝗮𝘁𝗮 %>% 𝗰𝗵𝗮𝗶𝗻_𝗼𝗳_𝗰𝗮𝗹𝗹𝘀() %>% 𝗺𝗼𝗿𝗲_𝗰𝗮𝗹𝗹𝘀() -> 𝗿𝗲𝘀𝘂𝗹𝘁 for over a decade. Yet most of the time, reactions of those who saw it varied between "you weirdo" and 🤮. Thanks for making my day!

120 0

Adrian Olszewski @adrianolsz.bsky.social · 04/12/2024

Worth remembering: rank-based tests like Kruskal-Wallis, Mann-Whitney (-Wilcoxon), ATS/WTS - do NOT compare medians unless IID holds (=same shape & dispersion). Otherwise, stochastic superiority is assessed. Use quantile regression (+ Wald's testing) or Brown-Mood instead. #statistics #datascience

Three boxplots with equal medians yet Kruskal-Wallis and individual pairwise Mann-Whitney comparisons report very low p-values. They do not compare medians unless the sample data are IID.

1101 0

Adrian Olszewski @adrianolsz.bsky.social · 04/12/2024

My "hello world" post 🫡 - a few slides about a viewpoint called 𝙌𝙪𝙚𝙩𝙞𝙡𝙞𝙨𝙢𝙪𝙨 (after Adolphe Quetelet), which initiated the "cult" of the normal distribution I've been observing for years. www.slideshare.net/slideshow/ch... Or a longer text: www.2kmm.pl/blog/On-the-... #statistics #ml #datascience

slideshare.net

Challenging the cult of the normal distribution in science

Challenging the cult of the normal distribution in science - Download as a PDF or view online for free

150 0

Adrian Olszewski @adrianolsz.bsky.social · 04/12/2024

If you're interested in #RStats in #clinicaltrials, check my two older (2020-2) presentations from the #rinpharma conference on the numerical validation of R. I work in a 100%R CRO, my daily challenges might be interesting. www.researchgate.net/publication/... and www.slideshare.net/AdrianOlszew...

researchgate.net

(PDF) Numerical validation as a critical aspect in bringing R to the Clinical Research

PDF | Validation of the R statistical package has become a hot topic since 2015, when the FDA issued the Statistical Software Clarifying Statement,... | Find, read and cite all the research you need o...

040 0

Adrian Olszewski @adrianolsz.bsky.social · 04/12/2024

Mee too, since I found this quote. PS: but I'm guilty of another crime - for more than decade I've been using the right arrow with pipes 😅 (data %>% chain_of_verbs %>% more_verbs -> result).

100 0

Adrian Olszewski @adrianolsz.bsky.social · 04/12/2024

A very important remark! This is essentially the idea of combining statistical discernibility and practical relevance, like in non-inferiority, equivalence and superiority (aka "minimum meaningful magnitude") studies (common in #clinicaltrials)

110 0

Adrian Olszewski @adrianolsz.bsky.social · 04/12/2024

This package is so underrated, while exposing a rich collection of awesome "hacks"! For example, I can control "breaks" for Y axis in each panel independently. Multiple colour scales, point-paths, advanced guides + axis features (eg nesting), stat_theodensity() working with faceting and many more!

010 0

Adrian Olszewski @adrianolsz.bsky.social · 03/12/2024

So good to see you here, Justin!

000 0