mxComputeStandardError {OpenMx}

Compute standard errors

Description

When the fit is in -2 log likelihood units, the SEs are derived from the diagonal of the Hessian or inverse Hessian. The Hessian (in some form) must already be available.

Usage

mxComputeStandardError(freeSet = NA_character_, fitfunction = "fitfunction")

Arguments

freeSet

names of matrices containing free variables

fitfunction

name of the fitfunction (defaults to 'fitfunction')

Details

If there are active MxConstraints and the fit is in -2logL units, the SEs are derived from the Hessian and the Jacobian of the constraint functions (see references).

References

Moore T & Sadler B. (2006). Maximum-Likelihood Estimation and Scoring Under Parametric Constraints. Army Research Laboratory report ARL-TR-3805. Schoenberg R. (1997). Constrained maximum likelihood. Computational Economics, 10, p. 251-266.