|
![](http://www.statmodel.com/discussion/icons/csc/1px.gif) |
Constraints for Correlated Uniqueness... |
![Previous](http://www.statmodel.com/discussion/icons/csc/navgen_arrow_left.gif) ![](http://www.statmodel.com/discussion/icons/csc/1px.gif)
|
Message/Author |
|
|
Hi there, One factor model reflected by the same 6 categorical items measured over 3 time intervals, each 3 months apart. Data arrange in Long Format. E.g., T1_B BY v1 v2 v3 v4 v5 v6; T2_B BY v7 v8 v9 v10 v11 v12; T3_B BY v13 v14 v15 v16 v17 v18; !Time One Uniquenesses WITH Time Two & Three Uniquenesses v1 v2 v3 v4 v5 v6 WITH v7 v8 v9 v10 v11 v12; v1 v2 v3 v4 v5 v6 WITH v13 v14 v15 v16 v17 v18; !Time Two Uniquenesses WITH Time Three Uniquenesses v7 v8 v9 v10 v11 v12 WITH v13 v14 v15 v16 v17 v18; Hypothesis One: Correlated Uniquenesses become smaller over successively larger time intervals. That is, the size of the correlated uniqueness between Time 1 and Time 2 will be larger than those observed between Time 1 and Time 3 etc. Lets say that I expected the correlated uniquenesses to decrease in a quadratic fashion as one test; and then as a second test I expect them to halve in size each time. Note that I have more than 3 time waves. Question 1. How might one impose such constraints? Many thanks for you help.. |
|
|
Sorry in my earlier post, 'WITH' cited above, should read 'PWITH' |
|
|
I think you mean wide format. See Example 6.17. |
|
Back to top |
|
|