Mon, 08/16/2010 - 23:02

I received this error when trying to run a latent growth curve model:

Error: The expected covariance matrix associated with the FIML objective in model 'MZ' is not of the same length as the dimnames provided by the objective function.

I used the traceback() function to try to determine the error and received the following output:

9: stop(msg, call. = FALSE)

8: updateObjectiveDimnames(.Object, job, model@name, "FIML")

7: genericObjModelConvert(flatObjective, job, model, namespace,

flatJob)

6: genericObjModelConvert(flatObjective, job, model, namespace,

flatJob)

5: translateObjectivesHelper(job, submodel, namespace, flatJob)

4: translateObjectivesHelper(model, model, namespace, flatModel)

3: translateObjectives(model, namespace, flatModel)

2: runHelper(model, frontendStart, intervals, silent, suppressWarnings,

unsafe, checkpoint, useSocket)

1: mxRun(lgcACEModel)

I am not sure how to interpret this message, or how to correct the script. Any help would be greatly appreciated! I have attached my script, and can upload my data if that is helpful.

Thank you!

I do not think my scrip attached, so I am trying again.

The expected covariance matrix must have row and column names. In this case, you are using the dimnames = selVars argument of the mxFIMLObjective() as a shorthand for placing row and column names on the expected covariance matrix. Your data must contain column names, and the column names should match the dimnames of the expected covariance matrix (although they don't have to be in the same order). You should assign a variable to the value

`data.frame(dzData,dzDefs)`

and then use the`colnames()`

function to determine the column names.