You are here

Modifying default baseline models for computing incremental fit indices in OpenMx.

2 posts / 0 new
Last post
Ayo's picture
Joined: 04/30/2016 - 21:39
Modifying default baseline models for computing incremental fit indices in OpenMx.

Is anyone aware of anyway one can change/respecify the null models for obtaining incremenrtal fix index measures in OpenMx.

My current plan to accomplish this is to fit alternate baseline models separately and to use the chi-square values from the modified baseline models to calculate the incremental fit index values.

I have ran into an issue of not being able to fit the base line model without paths in trying to carry out my plan. if it is not possible to respecify the baseline null/baseline model, can someone/anyone suggest a way to fit models without regression paths in OpenMx.

I am obviously not very savvy with programming based on the description I have provided, so please feel free to ask for clarification if you'd like to help with my question.

Ryne's picture
Joined: 07/31/2009 - 15:12
The summary method has

The summary method has arguments for SaturatedLikelihood, SaturatedDoF, Independence Likelihood, and Independence DoF, if you want to submit values for these four statistics. mxRefModels calculates the saturated and independence models, and you can submit the output of this function or any other list of 2 MxModels to refModels in the summary object, provided the saturated model comes first and the independence second.

If you're just estimating your own saturated model, use SaturatedLikelihood and SaturatedDoF. If you want to add the independence model, I think the list is easier, but you can input all four arguments manually. As long as the models make sense to you and you can make the argument about why they are better than more conventional methods, go ahead!