I am comparing the outputs of rotated factor patterns on stata and sas.
The interesting thing is, the pre-rotation factor patterns and
eigenvalues were identical between stata and sas.
BUT, after the varimax rotation, situation changed. The loadings are different.
Does anyone know what's going on?
Is that difference contributed to rounding?
Thanks.
Below are the codes and outputs.
=================================================================
For tha stata part
code:
factor v1 v2 v3 v4 v5 v6 v7 v8 v9 v10 v11 v12 v13 v14, pcf factors(2)
rotate, varimax
Rotated factor loadings (pattern matrix) and unique variances
Variable Factor1 Factor2
V1 0.85940 0.26300
V2 0.31170 0.88180
V3 0.52800 0.14110
V4 0.79810 0.04850
V5 0.68580 0.09280
V6 0.64250 0.23170
V7 0.69860 0.38670
V8 0.11190 0.40230
V9 -0.01900 0.90870
V10 0.82890 0.01060
V11 0.81530 0.16690
V12 0.28750 0.53900
V13 0.16590 0.21880
V14 0.59080 0.11800
=================================================================
As for the sas:
code:
proc factor rotate=varimax nfactor=4 out=PCA2;
var v1 v2 v3 v4 v5 v6 v7 v8 v9 v10 v11 v12 v13 v14;
run;
Rotated Factor Pattern
Factor1 Factor2
V1 0.84533 0.30509
V2 0.26776 0.89613
V3 0.52043 0.167
V4 0.68035 0.12658
V5 0.79476 0.0878
V6 0.63031 0.26312
V7 0.67866 0.42068
V8 0.09186 0.4073
V9 -0.06384 0.90667
V10 0.82737 0.05152
V11 0.58429 0.14706
V12 0.80606 0.20691
V13 0.26058 0.55251
V14 0.15487 0.22671
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
