Auto-suggest helps you quickly narrow down your search results by suggesting possible matches as you type.

Showing results for

Find a Community

- Home
- /
- SAS Programming
- /
- SAS Procedures
- /
- Differences in output (confirmatory factor analysi...

- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Highlight
- Email to a Friend
- Report Inappropriate Content

12-20-2010 07:47 AM

Dear all,

I am running an identical confirmatory factor analysis in SAS 9.2. on different computers. What I found out is that I obtain different SAS-output depending on the computer on which I am doing my analysis. It is possible that different releases of SAS version 9.2. causes my problem?

What I see is that the output is structured in a different way and tables have other names. The most important differences between both outputs which I have found are different values for AIC and CAIC, different representation of Lagrange multipliers (not Phi, Gamma and Beta) and t-values which are also represented for the standardized factor loadings. Does anyone know what's the cause of these differences?

I am running an identical confirmatory factor analysis in SAS 9.2. on different computers. What I found out is that I obtain different SAS-output depending on the computer on which I am doing my analysis. It is possible that different releases of SAS version 9.2. causes my problem?

What I see is that the output is structured in a different way and tables have other names. The most important differences between both outputs which I have found are different values for AIC and CAIC, different representation of Lagrange multipliers (not Phi, Gamma and Beta) and t-values which are also represented for the standardized factor loadings. Does anyone know what's the cause of these differences?

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Highlight
- Email to a Friend
- Report Inappropriate Content

12-24-2010 05:47 PM

You would have to specify which proc you are using and with which parameters. The following may or may not be relevant:

If you run PROC TCALIS (experimental) using the FACTOR statement to run an exploratory factor analysis, the Standardized Rotated Factor Loading Matrix and the Rotated Factor Correlation Matrix might be incorrect if you specify ROTATE= a rotation that is oblique. (The problem affects oblique rotations only - orthogonal rotations are not affected.)

Art

If you run PROC TCALIS (experimental) using the FACTOR statement to run an exploratory factor analysis, the Standardized Rotated Factor Loading Matrix and the Rotated Factor Correlation Matrix might be incorrect if you specify ROTATE= a rotation that is oblique. (The problem affects oblique rotations only - orthogonal rotations are not affected.)

Art