0
$\begingroup$

I'm new to SAS coding from SPSS point and click. I don't currently have SPSS, so SAS is my only option.

My question is: how do I get the percent variance explained and cumulative variance explained following varimax rotation?? I get the total that I would get under 'Rotation Sums of Square Loadings' but not the % variance or cumulative %.

Example SAS code below:

PROC FACTOR DATA=survey
METHOD=PRIN
PRIORS=ONE
NFACT=2
MINEIGEN=1
SCREE
ROTATE=VARIMAX
FLAG=.32
OUT=ex4comp4;
VAR pn1 pn2 pn3 pn4 pn5 pn6 pn7 pn8 pn9 pn10 pn11 pn12 pn13 pn14 pn15 pn16 pn17 pn18 pn19 pn20;
RUN;

Essentially I want to know what % variance each factor explains but all I get is the Total.

So frustrated and confused!

Thank you in advance for any suggestions or solutions KWAK

$\endgroup$
  • $\begingroup$ How do you get from 'I don't have SPSS' to 'SAS is my only option'? $\endgroup$ – Glen_b Apr 25 '13 at 9:59
  • $\begingroup$ Hi Glen, what I meant is that I have SAS on my computer right now and I am trying to do the work right now. Thanks. $\endgroup$ – kwak Apr 25 '13 at 10:02
  • 1
    $\begingroup$ I can't check right now (with SPSS but no SAS now, ironically), but the percentages you want come as default (e.g. 1, 2). My guess is your nfact and mineigen statements might have suppressed them. Start from a simpler command and see if that's true. $\endgroup$ – Penguin_Knight Apr 25 '13 at 10:12
  • $\begingroup$ I typed my answer while you were typing your comment. :-). One further note, I think the MINEIGEN and NFACTORS statements are sort of either/or, NFACTORS insists on 2 factors while MINEIGEN says keep only those with eigenvalue over 1. $\endgroup$ – Peter Flom - Reinstate Monica Apr 25 '13 at 10:16
  • $\begingroup$ Hmm, even with the following basic code i still don't get what I'm looking for. PROC FACTOR DATA=survey NFACT=2 ROTATE=VARIMAX; VAR pn1 pn2 pn3 pn4 pn5 pn6 pn7 pn8 pn9 pn10 pn11 pn12 pn13 pn14 pn15 pn16 pn17 pn18 pn19 pn20; RUN; $\endgroup$ – kwak Apr 25 '13 at 10:27
0
$\begingroup$

It is part of the default output; see the table called "Eigenvalues of the correlation matrix" and look in the two columns on the right.

EDIT with additional info:

A bit of googling found this article which shows how to get this by using PROC PRINCOMP.

$\endgroup$
  • $\begingroup$ Thank you Peter - that's what I was expecting from my results, but I don't seem to get that output. I must be missing something obvious here. When I reduce the number of factors to 2 and run the varimax rotation I receive the initial (non-rotated) table with eigenvalues and the two columns on the right. However, I do not received a second similar table following the rotated solution. My output shows the correct total (e.g. 4.886) but not the % var or cumulative %. (Correct is known because I am trying to replicate some SPSS results done previously) $\endgroup$ – kwak Apr 25 '13 at 10:25
  • $\begingroup$ Hang on! I can't figure out how to get it to come up, BUT am i correct in saying that if my total after rotation is 4.886 and I have 20 items then my percent variance is 4.886/20*100!!! $\endgroup$ – kwak Apr 25 '13 at 10:43
  • $\begingroup$ My apologies Peter- the total variance remains the same but the % variance explained by each factor may change. Now I understand how to calculate it, I have to go and figure out why so I understand it properly. Thank you for your time and efforts! $\endgroup$ – kwak Apr 25 '13 at 11:11

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.