PSPP - News: Factor Analysis and Rotations
[Submit News]
[13 news in archive]
Factor Analysis and Rotations
Item posted by John Darrington <jmd> on Fri 21 May 2010 08:44:28 AM UTC.
The latest PSPP master version now supports factor analysis/principal components analysis including factor rotations. Varimax, Quartimax and Equamax rotations are available. Currently oblique rotations are not supported.
The factor command also outputs a "scree" plot and loading matrices.
Comments:
| problems with PSPP factor (posted by Ton Vogels, Mon 09 Aug 2010 01:37:48 PM UTC) |
I tried the new factor analysis and encounterd some problems
I suppose this is caused by a problematic handling of missing values, because when i removed observations with missing values all loadings remaind < 1.00.
However, when i compared the results with those of R-factal there were marked differences, both when using the correlation matrix and the covariance matrix.
I was not yet able to compare the results with those of SPSS, by I wonder how reliable the results of PSPP are. |
| RE: problems with PSPP factor (posted by Ton Vogels, Tue 10 Aug 2010 12:58:42 PM UTC) |
I guess I was to sceptical about PSPP. I just checked the PSPP results with thos of SPSS and they are nearly identical.
|
| RE: problems with PSPP factor (posted by Juanma Belda, Thu 28 Apr 2011 04:29:15 PM UTC) |
Its neither a problem of SPSS or R. In fact both are doing different things. In SPSS you are doing a factor analysis and the decomposition is slightly different that when you are doing a principal component analysis. |


