|
|
 |
| Clustering But No Between Group Variable |
 
|
| Message/Author |
|
|
| Anonymous posted on Thursday, September 13, 2018 - 9:24 pm
|
|
|
I have a dataset with employee responses and supervisor ratings of employee performance.
I want to account for the non-independence of the performance ratings, but there is no between-group variable. I have read a fair amount but don't see an example of how to accomplish this (all of the examples I have seen have a between-group variable).
Can I just use this, where I am saying there is no between-group variable?:
USEVARIABLES ARE clus x y;
missing = all (999);
cluster = clus;
Within = x;
Analysis: Type = TwoLevel random;
MODEL:
%within%
s | y on x;
%Between%
;
OUTPUT: sampstat; |
|
|
|
|
The answer is Yes. Just add on %Between%:
y WITH s;
to make sure the 2 variances and the covariance get estimated. |
|
|
| Back to top |
|
|
