Skip to contents

Variance-covariance matrix for a HubbellGLM fit, optionally adjusted for shared observations via a sandwich estimator.

Usage

vcov_shared(fit, similarity = NULL)

Arguments

fit

An object of class HubbellGLM.

similarity

An optional \(n \times n\) similarity matrix encoding dependence between the \(n\) observations used to fit fit. The diagonal must be 1 and all off-diagonal entries must be in \([0, 1)\). If NULL (default), returns vcov(fit).

Value

A \(p \times p\) variance-covariance matrix.