Here are some new timings for Bayesian proportional odds models, using 4 cores (4 chains) with cmdstan and my R rmsb package blrm function. Non-Dirichlet priors result in divergences after 300 intercepts, and slowness, so timings for them are not shown for > 300 distinct Y-values. #Statistics

P-value = P(data MORE extreme than mine | H0). Justification for this is largely missing. See fharrell.com/post/journey

Computing tool of the day: I've needed to have an elegant way to stitch together pieces of Quarto books to make a unified html file for teaching. In collaboration with Claude I now have a great yaml-driven tool for this. #RStats

There is also an "st" command line option to get a symbol table at the end of the R code listing. #RStats

Python script for pretty-printing of R code files: github.com/harrelfe/rsc...

Got a personal email out of the blue for a ghost worker. Shameless.

Goal-Driven Flexible Bayesian Design presentation updated w/comparison of performance of frequentist group sequential designs: hbiostat.org/bayes/design . Frequentist approach takes far too long to make a decision by controlling something that is NOT an error prob. #Statistics #StatsSky

Nice work, and these linear models are special cases of semi parametric ordinal regression with no normality assumption - hbiostat.org/rmsc/ordsurv

I just posted a critique of this paper at discourse.datamethods.org/t/critique-o... where I hope others will add their thoughts #StatsSky #EpiSky #Statistics #rct

Here’s the example: hbiostat.org/bbr/images/a...

Ordinal semiparametric cumulative probability models have an amazing variety of statistical methods as special cases, including proportional odds and proportional hazards models - hbiostat.org/rmsc/ordsurv... #Statistics #StatsSky

#Statistics thought of the day: The Central LIMIT Theorem hbiostat.org/bbr/htest.ht... #StatsSky

I lost Galben in 2019 and still think about him a lot. We had a deep connection.

The #rstats qreport package for assisting with report writing using Quarto has a new function makegvflow that allows users to easily make graphviz flow diagrams by specifying plain text with indenting to define the hierarchy. An example is at hbiostat.org/rflow/rforma...

Nashville is a winter wonderland today

3/3 The third way to make code more readable is to use Quarto's wonderful code annotation syntax. When you hover over one of the circles, the annotation pops up. #rstats

2/3 The second way to make compact code more readable is to include comments providing little diagrams to dissect the code #rstats

I continue to add #rstats how-tos to my online R Workflow e-book at hbiostat.org/rflow

Lots more yaml that applies site-wide, e.g.

#statistics thought of the day, from hbiostat.org/bbr/change#s...

Spike histograms work for all sample sizes and for extreme frequencies of tied values.