Date:         Wed, 16 Sep 1998 15:13:39 -0500
Reply-To:     Melinda MayField <Melinda_MayField@MGMT.PURDUE.EDU>
Sender:       "SAS(r) Discussion" <SAS-L@UGA.CC.UGA.EDU>
From:         Melinda MayField <Melinda_MayField@MGMT.PURDUE.EDU>
Organization: Krannert School of Management, Purdue University
Subject:      repeated measures anova problem
Content-Type: text/plain; charset=us-ascii

If you can help, please email privately at
melinda_mayfield@mgmt.purdue.edu.
 
I need help figuring out how to analyze this data in SAS.  I distributed
4 forms of a research survey.  Form 1 contains (1) first a competency
statement; (2) then task statements.  Form 2 contains (1) first task
statements; (2) then a competency statement.  Form 3 contains (1) first
a competency statement; (2) then KSA statements.  Form 4 contains (1)
first KSA statements; (2) then a competency statement.  The competency
statements are always identical in all 4 forms, the task statements are
identical in forms 1 and 2, and the KSA statements are identical in
forms 3 and 4.
 
I want to analyze the effect of (1) order of competency statement
(beginning or end) and (2) question type on subject ratings  on our
scales (task, KSA, or competency).  I created scales by averaging the
ratings on the task items and KSA items within a form.  The competency
statement has its own rating. So I have 3 variables: "TASKFREQ,"
"KSAFREQ," and "COMPFREQ."  The complication is that I want to compare 3
levels of question type: task, KSA, and *competency*.  How to do this
confuses me, since the competency statements were on the same surveys as
the task and KSA statements (i.e., I have 4 forms with competency
statements, but only two with task statements and two with KSA
statements).  I can't name the competency and KSA/task scales the same
since they are in the same dataset, and the same respondents filled out
both task and competency statements OR both KSA and competency
statements.  How do I set up the ANOVA to look for differences in all 3
question type responses? I've tried using repeated measures ANOVA (using
TASKFREQ, KSAFREQ, and COMPFREQ as the repeated measures), but I keep
getting an error message that one of the variables is missing or has a
freq of zero.  I presume this is because "TASKFREQ" only appears in
Forms 1 and 2, and "KSAFREQ" only appears in Forms 3 and 4.  In case it
helps, below I've pasted the SAS code I used for the unsuccessful
repeated measures ANOVA.  How do I handle this?
 
Please email me privately at melinda_mayfield@mgmt.purdue.edu.
 
PROC GLM DATA=ALLKIND;
CLASS ORDER;
MODEL TASKFREQ KSAFREQ COMFRQ1 = ORDER / NOUNI;
REPEATED QTYPE 3 CONTRAST (1) / SUMMARY;
run;