@amoeba Oh and I just discovered this thread. It probably does not fulfil my expectations of a "canonical question", particularly because it does not venture into the territory of "how can I interpret the output". But it is worth taking note of, and trying to avoid being a mere duplicate of it.

In general, I think you should not be afraid of duplicating some old obscure threads with few to no upvotes and no great answers.

It's much easier to attract attention to a new question than to re-kindle attention to such an old one

I take your point, I actually thought this answer by Aleksandr Blekh was quite a nice attempt

It does fit quite a lot into a relatively short space.

I feel that if I made a focus on the interpretation then it avoids being a duplicate of that one, anyhow.

Both this and another question (with Placidia's answer) should be edited, by the way, to make them more useful. Titles could be improved and both can take some light editing. Perhaps some upvotes will accumulate in the process.

I have been recently thinking of creating a [factor-rotation] tag. I did not do it because it would need to be applied to at least a 100 questions, probably more, and that's quite some enterprise.

But this tag is definitely lacking.

I agree, "What does “varimax” mean in SPSS factor analysis?" really doesn't get to the heart of that question.

People sometimes put [rotation] tag on these question, but "rotation" is so ambiguous that it's not very useful

Heh, it's odd that the top google hits for "varimax" include several posts on this site, and they all seem to throw the "PCA" in somewhere (even "What does “varimax” mean in SPSS factor analysis?")

I think I will take pains to avoid mentioning "PCA" at all.

I was going to ask you if there were an appropriate tag. You have pre-empted me :)

Just edited this one stats.stackexchange.com/questions/113003/…

Oh that's much better.

The tag is really needed... Perhaps we should start it after all.

There's not a lot you can do with "Also, I am confused about the relationship between principal component analysis, varimax rotation and exploratory factor analysis, both in theory and in SPSS. How are they related?"

I must go. Looking forward to your varimax question :) Good night!

I'll try to ask something a bit clearer than that!

I'll likely post it early next week.

Thanks for the discussion, most useful.

Good night!

@amoeba ttnphns has used chat, since there's a user page

chat.stackexchange.com/users/42761/ttnphns

The only thing I clearly recall is seeing involvement in one-on-one discussions spilling over from comments in posts, but if I remember right that's enough for a ping to work.

@amoeba if you do write (or even see) an answer to a newer question where the combination of question and ends up being better answer to an earlier one than exists there, vote(/flag if you can't vote) the older one as a duplicate (or if not quite close enough to be duplicates but still of interest, link the new one in comments on the old one.

If the questions are effectively identical, flag to merge the old into the new.

@amoeba I think you can also explicitly invite ttnphns to Ten Fold from that user page

But I think your @ should notify.

@SIlverfish on avoiding becoming over-broad in answering a question, that's a good impulse. If there's another answer that deals with the broader issue, you can always link to it rather than deal directly (that might then involve writing a new question to ask about those aspects in order to have an answer to link to ... and then possibly answering it).

If there's several inter-related questions-and-answers that cover a topic well, that seems like a good thing, if they are linked up.

...

I found a bad wikipedia page the other day. The page on iteratively reweighted least squares is so terrible, I feel like I want to delete it. I feel like it's beyond repair. IRLS is far more general than it suggests, for starters, but even what it does is rambly and oddly set out. I don't even know where to begin with it. I wish I had time to just sit down and rewrite it, but I think the battle would be a long one.

The alternative would be to write a half-way decent exposition here but at present I really don't have time to write that either.

I was going to point someone to read about IRLS at wikiipedia, but fortunately I looked at the page. ... and just went "nope, nope, nope" ... can't send them there because they'd wonder what that had to do with the problem we were discussing.

@Silverfish: I only glanced yesterday at the thread you found (stats.stackexchange.com/questions/144966), but now actually read it, and to me it makes no sense at all. What the OP seems to have been asking, is whether varimax rotation in FA is equivalent to doing PCA on FA scores (that's my best guess to what he/she meant). The answer to this is definitely "no". And then there is Aleksandr's answer, accepted and with 4 upvotes, saying "the answer to your question is Yes".

This makes no sense.

I don't even know how to edit this question! It's too confused.