A bias-corrected covariance estimator for improved inference when using an unstructured correlation with quadratic inference functions

Notable bias can exist in the empirical covariance matrix of parameter estimates obtained from the quadratic inference function method that incorporates an unstructured working correlation. We therefore derive a bias correction. Via simulation, we show that the proposed correction leads to appropriate standard error estimation. (C) 2013 Elsevier B.V. All rights reserved.