Date:         Thu, 10 Apr 2003 09:43:24 -0400
Reply-To:     Stephanie Roll <zteph@GMX.NET>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Stephanie Roll <zteph@GMX.NET>
Subject:      error in SF-36-analysis program???

Hi,
I want to analyze SF-36 data (a 'quality of life'-questionnaire) with the
SAS program that comes with the manual (German version). Now I find 4 lines
of this program to be doubtful. In these lines values should be set to
missing if they are outside the range of valid values. They are at the
beginning of the parts that calculate the raw-subscales of
- role-physical index,
- role-emotional index,
- mental health index and
- vitality index.
(computation of the other 4 subscales (physical functioning, social
functioning, general health perception, and pain) seems correct).
The set-to-missing is done in an array. For example (role-physical index):

ARRAY RP(4) ROLPH1-ROLPH4;
DO I = 1 TO 4;
   IF RP(I) < 1 OR RP(I) > 2 THEN RP(I) = .;
END;

In the questionnaire the 4 items (rolph1-rolph4) of the role-physical index
can take values between 1 and 5.
In the above array however any value greater than 2 will be set to zero.
This is wrong, I think....

Has anyone used this program to evaluate the SF-36 and had the same problem?
Does anyone know something about where to get "official" evaluation
programmes for SF-36?
(The above might be a German problem. I don't know any US-programmes.)

Thanks,
Stephanie Roll