turn on suggestions

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

Showing results for

Find a Community

- Home
- /
- Analytics
- /
- Stat Procs
- /
- Running Factor Score with Polychoric Correlation

Topic Options

- 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
- Permalink
- Email to a Friend
- Report Inappropriate Content

12-09-2015 08:51 AM - edited 12-09-2015 08:52 AM

I'm trying to get factor scores for my data set using polychoric correlation. I am trying to follow the same precedure one would use for Pearson Correlation. The following Pearson Correlation runs fine:

**proc factor** data=Diss.Impact method=principal score outstat=Diss.Fact;

var Impact_ELLs Impact_curric Impact_acheive Impact_manage Impact_teach;**run;**

**proc score** data=Diss.Impact score=Diss.Fact out=Diss.FScore;

var Impact_ELLs Impact_curric Impact_acheive Impact_manage Impact_teach;**run;**

After I run the **proc factor**, the factor statistics table gives me the correlations, eigenvaules, patters, and scores in the Diss.Fact file. When I run **proc score**, I get a new data set with my five variables and Factor score for each case.

So, to get the polychoric correlation stats, I run:

**proc** **corr** data=Diss.Impact polychoric;

var Impact_ELLs Impact_curric Impact_acheive Impact_manage Impact_teach;

**run**;

I have created a new factor statistics table with the polychoric correlation stats but I cannot produce the coefficient score needed to run the **proc score** test correctly. I get this error:

ERROR: No observation of scoring coefficients having _TYPE_='SCORE' is found in the SCORE= data set.

How does one create the scoring coefficients when trying to produce factor score in a polychoric correlation?

I'm using SAS 9.3

Cheers-

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

Posted in reply to ssahr

12-09-2015 03:36 PM

Take a close look at the actual contents of the data set you generated with Proc factor, maybe print all or part of it.

Then take a look at the data set created by Proc Corr (which you don't show in your code).

Proc Score is expecting some specific bits that aren't in the default output of Proc Corr. You likely need to make a pass with a data step to add as a minimum a character variable named _type_ with values of SCORE.