A rule of thumb is that if the recipe is based on functionality from one of the “StatsKit packages” https://github.com/JuliaStats/StatsKit.jl/blob/master/src/StatsKit.jl, then it generally makes sense to have it in StatsPlots.
Yeah, not the best idea I’ve ever had. My point was more along the lines of “if you don’t like the decisions that have been made, you can make a package to get what you want. Others might use it too.” I like the idea of a verbose argument, and if the package maintainers are amenable a PR is definitely a better way to go
Got things started with a pull request in MultivariateStats https://github.com/JuliaStats/MultivariateStats.jl/pull/90.
To support this request:
…so better check that we don’t choose an approach/terminology which is only familiar to psychologists.
The use of the function from the psych package is widespread and I would say it is common knowledge among R users.
…ike factor analysis.
Factor analysis is also affected. The output of the model estimate is very similar.