Hi Gang,
Ah! I see now! Though, I would like to double-check something else if you wouldn't mind. I have initially reset my specification matrix as:
A1 A2 A3
A1 0 0 NA
A2 0 0 NA
A3 NA 0 0
However, this only works by selecting the "direct" option, giving me an Instaneous effects matrix that ends with "NULL" and t-values that are ridiculously strong.
Using the matrix with iterative option:
A1 A2 A3
A1 0 0 NA
A2 0 0 NA
A3 0 0 0
I get more reasonable values but with an LR overidentification notice.
My question, then, is whether or not the second specification matrix is going to be conceptually valid, given that I think that my network should be A1 <--> A2 <-- A3?
Thank you again!
~Nate