Re: AMOS error: moment matrix is not positive definite

Posted by Rich Ulrich on Feb 28, 2012; 4:36am
URL: http://spssx-discussion.165.s1.nabble.com/AMOS-error-moment-matrix-is-not-positive-definite-tp5520824p5520841.html

Well, the fundamental thing that you are doing wrong is what the
message tells you -- You are trying to do an analysis without a
real covariance matrix (that is, one generated from a large
enough N to make make it positive definite).

I don't know what sort of errors it would be, that Amos might
be able to work around.  But apparently your problem is worse.

a) What are you using for covariance/correlation?  Describe,
or maybe show it, too.
b) Fix it.

 --
Rich Ulrich

> Date: Mon, 27 Feb 2012 23:22:28 -0500

> From: [hidden email]
> Subject: AMOS error: moment matrix is not positive definite
> To: [hidden email]
>
> Hello,
>
> I am trying to run a path analysis with AMOS but I am getting the error
> message:
>
> Minimization unsuccessful.
>
> The sample moment matrix is not positive definite. It could fail to be
> positive definite for any of the following reasons:
> 1. The sample covariance matrix or correlation matrix contains a data
> entry error
> 2. The observed variables are linearly dependent
> 3. The sample covariance matrix was computed from incomplete data using
> the method pairwise deletion
> 4. The sample correlation matrix contains correlation coefficients
> other than product moment correlations.
>
> The message tells me to allow non-positive values for the matrix, I did so,
> but still I am getting the error message
>
> Could anyone please let me know where am I going wrong. It would really be
> of great help.
> ...