f



AMOS output

Hello,

Does somebody have any suggestions why AMOS provides the following
message in the output?

"Notes for Model (Default model)

Computation of degrees of freedom (Default model)
Number of distinct sample moments:	55
Number of distinct parameters to be estimated:	23
Degrees of freedom (55 - 23):	32

Result (Default model)

Iteration limit reached
The results that follow are therefore incorrect.
The model is probably unidentified. In order to achieve
identifiability, it will probably be necessary to impose 1 additional
constraint."


Is it possible to have both a positive number of degrees of freedom
and a model which is unidentified? How can i fix the problem?

I will really appreciate your help. Thanks.
0
1/22/2010 3:03:07 PM
comp.soft-sys.stat.spss 5679 articles. 0 followers. Post Follow

1 Replies
5069 Views

Similar Articles

[PageSpeed] 56

On Fri, 22 Jan 2010 07:03:07 -0800 (PST), Zidavora Elena
<elena.zidarova@googlemail.com> wrote:

>Hello,
>
>Does somebody have any suggestions why AMOS provides the following
>message in the output?
>
>"Notes for Model (Default model)
>
>Computation of degrees of freedom (Default model)
>Number of distinct sample moments:	55
>Number of distinct parameters to be estimated:	23
>Degrees of freedom (55 - 23):	32
>
>Result (Default model)
>
>Iteration limit reached
>The results that follow are therefore incorrect.
>The model is probably unidentified. In order to achieve
>identifiability, it will probably be necessary to impose 1 additional
>constraint."
>
>
>Is it possible to have both a positive number of degrees of freedom
>and a model which is unidentified? How can i fix the problem?
>
>I will really appreciate your help. Thanks.


One key phrase is probably "Iteration limit reached."

I've never used AMOS, but by analogy with its simpler
cousin, factor analysis, I conclude that a solution did
not converge adequately within the given number of steps.

Presumably, you can increase the number of steps by
stating "ITERATIONS=" .  However, slow convergence
also seems to indicate that *these*  data do not provide 
a substantial difference between alternative latent 
structures, so that simplifying the model (additional
constraints)  may be a good idea even if changing
ITERATIONS  lets AMOS crank out a solution.  

You might also check your data again for such things
as miscoded data or other natural outliers that might 
underlie the difficulty in obtaining a fit.


-- 
Rich Ulrich 
0
Rich
1/22/2010 8:14:20 PM
Reply: