svy.varcoef_hopit: Calculation of the variance-covariance matrix for a specified...

In hopit: Hierarchical Ordered Probit Models with Application to Reporting Heterogeneity

Calculation of the variance-covariance matrix for a specified survey design (experimental function)

Description

This function is an equivalent of survey:::svy.varcoef. In the original approach estfun is calculated from glm's working residuals:
estfun <- model.matrix(glm.object) * resid(glm.object, "working") * glm.object$weights
In the hopit package, estfun is directly calculated as a gradient (vector of partial derivatives) of the log likelihood function. Depending on detected design an appropriate survey function is called.

Usage

svy.varcoef_hopit(vcovMat, estfun, design)

Arguments

vcovMat	a variance-covariance matrix.
estfun	a gradient function of the log-likelihood function.
design	a survey.design object.

See Also

svydesign hopit

hopit documentation built on Oct. 1, 2022, 5:06 p.m.