Zach Madaj

A fist bump a day, makes a fans day 😅 🎥: @MLB

I've found this a really nice resource for explaining statistical theory (via @seabbs). Sometimes people starting out have asked questions like 'should I calculate a correlation or do a regression?' The answer: one is a special case of the other. #3_pearson_and_spearman_correlation" target="_blank" rel="nofollow noopener">lindeloev.github.io/tests-as-linea…

@biobenkj routledge.com/Hierarchical-M…

WHY do we divide by n-1 when computing the sample variance? I've never seen this way of explaining this concept anywhere else. Read on if you want a completely new way of looking at this.

Sergei Bobrovsky's best two-game stretch in Florida: games one and two versus Toronto.

7777777777772777777777777 is a prime number.

mathematicians are absolutely deranged

#Statistics thought of the day: A different way to do sample size calculations is to find N such that using a much more skeptical prior doesn't matter: #sec-design-pn" target="_blank" rel="nofollow noopener">hbiostat.org/bayes/bet/desi… @vandy_biostat #clinicaltrials #stattwitter @EdgeforScholars

A paper in @Nature reports a method of converting male (XY) mouse stem cells into female (XX) cells to produce functional egg cells. These egg cells were then fertilized and around 1% of the resulting embryos produced healthy offspring. go.nature.com/3ThGEXq

Holy shit. GPT-4, on it's own; was able to hire a human TaskRabbit worker to solve a CAPACHA for it and convinced the human to go along with it.

when they tell you all the data is in one big color-coded excel file

If you're alone today on Valentine's Day, remember: nothing can impute the missing data in your heart. #rstats #pydata #datascience

I love statistics. I also love making statistics more boring by acquiring data that can be analyzed with the simplest methods.

2 useful papers today on how you can get basically any answer you want from regressions using arcsinh(y) or log(y+1) I made these graphs to show the problem: The mean of the transformed var (black lines) varies like crazy by what units you choose for y 🧵 on my own take-aways

#Statistics proposal of the day: The best general-purpose R^2 measure is the adjusted Maddala-Cox-Snell R^2 using n = effective sample size, implemented in the #rstats Hmisc package. hbiostat.org/bib/r2.html @vandy_biostat #rmscourse

📽️ NEW VIDEO 📽️ "Sample size calculations for external validation of a clinical prediction model" - what sample size is required for a validation study aiming to precisely estimate model performance? - 30 min broad overview 👇 #stats #MachineLearning youtu.be/1bY4puru34Q

my neighbor decided to make the undead army of skeletons more scary by adding some math facts

Pay attention: Big data loses to good data. “A survey of 250,000 respondents can produce an estimate of the population mean that is no more accurate than an estimate from a simple random sample of size 10… data quality matters more than data quantity” nature.com/articles/s4158…

#Statistics thought of the day: If you must use any cutpoints for continuous variables, only use them for the 3rd derivative of how the variable relates to outcome, i.e., use cubic splines #statstwitter @vandy_biostat @VUDataScience hbiostat.org/rmsc/genreg.ht…

@sgivan Winner's curse

"The literature must include publication of null results from well-powered studies ... Otherwise, publication bias toward significant differences skews the evidence base.” spectrumnews.org/news/null-and-…