when we have misspecified the variance function when estimating a mode
Question
When we have misspecified the variance function when estimating a model by WLS,the resulting WLS standard errors, and test statistics computed using the WLSvariances and covariances, are wrong, an issue
that can be rectified by using heteroskedasticity-robust variances and covariances with WLS, as we often employ with OLS under possible heteroskedasticity. True O False