robVcov {drgee}

Robust Variance Calculation

Description

robVcov and robustVcov calculates the asymptotic variance for Z-estimators.

Usage

robustVcov(U, d.U.sum, id = NULL)
robVcov(U, d.U, id = NULL)

Arguments

U

A n x q matrix of the estimating equations evaluated at the estimated model parameters, where n is the number of observations and q is the number of estimating equations.

d.U.sum

The sum of the jacobian of U evaluated at the solution to the estimating equations, with rows corresponding to the estimating equations and columns corresponding to the model parameters. The number of model parameters is assumed to equal the number of estimating equations such that d.U.sum is a q x q square matrix.

d.U

The mean of d.U.sum taken over all observations.

id

A factor with levels corresponding the clusters in the data. Default is NULL in which case all observations are considered to be independent.

Details

For robust variance estimation, see van der Vaart (2000).

For clustered data, the rows in U are added clusterwise resulting in a cluster robust estimate of the variance.

Value

The estimated covariance matrix.

Author(s)

Johan Zetterqvist, Arvid Sjölander

References

van der Vaart, A.W. (2000), Asymptotic Statistics, Cambridge University Press, pp. 52–53.