I've made this cheat sheet and I think it's important. Most stats 101 tests are simple linear models - including "non-parametric" tests. It's so simple we should only teach regression. Avoid confusing students with a zoo of named tests. 1/n.

RT @larsloekke: Dear American friends. We agree that status quo in the Artcic is not an option. So let’s talk about how we can fix it -….

RT @iavins: Collection of insane and fun facts about SQLite. Let's go!. SQLite is the most deployed and most used database. There are over….

RT @mattansb: NEW to {bayestestR} dev version - support for #marginaleffects!. Take it for a test drive: remotes::install_github("easystats….

RT @giveffektivt: Gør mere godt for pengene ved at give effektivt. Se vores nye explainer-video og del!

There are no R packages to calculate sample size (/power) given uncertain population parmeters. I.e., like specifying distributions over mu and sigma in a t-test rather than point estimates. Must. Resist.

`job::job()` is self-help: I felt uneasy starting heavy tasks and procrastinated while they ran. Now I quite like long-running code. It feels productive having the computer churning away at a few jobs while I work. Keep those cores busy!. Website:

If you print diagnostics within the job, your RStudio jobs pane becomes a high-level record of what you did and the results.

{job} comes with RStudio add-ins too. Select some code and run it as a job:

A few example of where I routinely use job::job():. # Upgrade packages.job::job({update.packages(ask = FALSE)}). # Run checks and unit tests.job::job({devtools::test()}).job::job({devtools::check()}). # Build pkgdown site.job::job({pkgdown::build_site(preview = FALSE)}).

When the job finishes, it silently returns the results to your main session where you can inspect it:

RStudio jobs are awesome. You can see the progress on the job while you continue working in the console:

A new release of {job} just hit CRAN. Here's a three-year anniversary thread:. job::job({slow_code_here}) runs code as RStudio Jobs, keeping your console free to continue developing. (1/n).

All the probability distribution names are so Normal and Uniform. It's Poisson for Students of statistics. Let's use Cauchy names and be ready to handle the Logistics of our Exponential success!.

Do you do statistics on clinical trials in a pharma company? Do you use modern/open tools like R? (or want to?) Check out {pharmaverse} and respond to this survey! Pretty cool stuff with heavy lifting from my skilled new colleagues at @novonordisk.

I can die happy now.

NEW JOB. After 4 years at @KamstrupGroup, I'm starting as a statistical specialist at @novonordisk in April. This will mark a "return home" to working with patient studies and RCTs. So exciting!.

Revewer #2: "The authors should check all the language over the whole manuscript.". Also reviewer #2: "After the discussion section, there is conclusion section, however, it should be placed again.".

What!? Why didn't I realize this earlier?. BTW, this whole thread is 👌 @TaymAlsalti.

Money in the bank is not a giant pile of cash. Gold and diamonds is not a giant pile of cash. Someone owing you is not a giant pile of cash. Property is not a giant pile of cash.

I do indeed get variable name limit at 10,000 chars in R. Try it:. for (i in 2^(0:20)) {. cat(i, ""). varname = paste0(rep("a", i), collapse = ""). assign(varname, i). stopifnot(get(varname) == i).}.