I basically concur with everything wolf.rauch said here, and would like to discuss some alternatives that might be available to you.
My understanding is that AMOS had had FIML (full information maximum likelihood) for continuous data for at least ten years before it was acquired by IBM -- see http://www.smallwaters.com/amos/faq/faqa-missdat.html, and that is an old FAQ by one of the original developers who left the project around 2000. If you are willing to ignore the ordinal nature of your items, you can just use this method, and won't bother figuring out multiple imputation.
If you don't like this solution, and you want to retain the categorical nature of the data, you would need to find the chained equations method with ordinal links (if SPSS has it at all). If SPSS only imputes draws from a multivariate normal distribution, then you are back to the situation of ignoring the ordinal nature of the data, and in no way better off than with AMOS' FIML. (I've no clue what's available in SPSS, you'd have to figure it out. In the end, everything would be fruitless if AMOS does not support multiple imputation -- and that, again, I don't know.)
If you are willing to consider Stata, there is a chance you'd be able to conduct your analysis in it, with all the bells and whistles of both multiple imputation for ordinal data using either Patrick Royston's ice
or official mi
, and then the new sem
suite. Alternatively, you could run gllamm
to obtain FIML estimates for ordinal data (although it would probably take eternity to converge).
In general, structural equation modelling (SEM) with all observed variables is typically called path analysis.
One of the main motivations for SEM is to attempt to model relationships between latent variables. By including items rather than the composite score and modelling items as indicators of a latent variable you are able to assess relationships between latent variables.
In particular, with items rather than the composite score
- you can assess your measurement model
- you can get an estimate of relationships between latent variables (i.e., adjusting for measurement error).
Various middle grounds also exist including:
- item parcelling: i.e., you create two or more parcels of items from your 11 items, and include these parcels as observed variables for a latent variable.
- incorporate error of measurement into the model with observed variables.
It is not "invalid" to include a composite variable in SEM. However, it is in some sense invalid to say that inferences based on the observed composite variable are representative of the relationship between theorised latent variables. Most of the time, you'd want to adopt one of the other approaches (i.e., including items, including item parcels, or include measurement error).
Best Answer
Don't use a composite variable in a structural equations model. Instead, add a latent variable that the manifest variables (i.e., the items) all measure. The composite variable isn't what the items measure, it is just (presumably) a better measure of the latent variable in question. If you were to add a composite variable, it would decrease the ability of your model to extract the information in your data, as it forces the loadings to be 1, whereas the latent variable will allow the loadings to be estimated from the data.