search

aiorazabala / qmethod

R package to analyse Q methodology data
GNU General Public License v2.0
34 stars 18 forks source link

reconstruct `reorder = TRUE` from `psych::principal` rather than reverse-engineer #274

Closed maxheld83 closed 5 years ago

maxheld83 commented 9 years ago

obviously, reverse-engingeering is bad as in the stop-gap #273 so if William Revelle agrees to add this feature to psych::principal that would be preferable.

maxheld83 commented 9 years ago

Here's the email I just send:

Dear Prof. Revelle,

Many thanks for the prompt addition of the rotmat output to psych::principal – this is a great feature, and made my life a lot easier developing for qmethod.

I recently noticed that psych::principal reorders the components according to their sum-of-squared loadings (formerly known as Eigenvalues, but no longer once the Principal Components are rotated). (That is documented behaviour in your package, I had just forgotten about it).

Would you consider adding an argument (say, reorder = TRUE) to psych::principal to en/disable this reordering?

The reordering causes a bunch of problems for my development of a by-hand rotation function for our package "qmethod" (you're probably not interested in the details).

We can of course reverse-engineer the reordering on our end, but I'd like to avoid that if you're open to this suggestion.

I've also posted a question on CrossValidated concerning the reasons for re-ordering.

Let me know what you think about this.

Best, Max

maxheld83 commented 9 years ago

there is now a better way to do this than is presently done: http://stats.stackexchange.com/questions/169934/should-i-reorder-principal-components-after-rotation/169991

This will also, incidentally, render the homegrown flipping from #268 unnecessary.

aiorazabala commented 5 years ago

@maxheld83 can we close this?