|
|
Parameter a function of person specif... |
|
Message/Author |
|
|
Is it possible to have a parameter, such as a factor loading, constrained to be a function of a person-specific observed covariate? For instance, is it possible to specify lamba_ix2 = a + b * IQ_i, where i is an individual and IQ_i is the observed IQ of that individual? If the covariate is categorical, this can be implemented via multigroup analysis, where the number of group is equal to the number of categories of the covariate. However, how do I do this if the covariate is continuous? Best, Guillaume |
|
|
Try the VARIABLE command option Constraint = iq; in combination with Model Constraint, defining the "new" parameters a and b, and letting the loading parameter be a function of these as you have written it. See UG ex5.23 for an application of this approach. |
|
Back to top |
|
|