3 problems

Let the spatial domain be a subset of $\mathbb R^d$, and consider generalized least squares using an undersmoothed working covariance class of fixed smoothness. Let $\beta$ denote…

Let $X$ be an exposure process, $W$ an unmeasured spatial confounder, and $Y$ a response whose regression slope on $X$ is $\beta$. Say that $W$ is smoother than $X$ when its smooth…

Let $\beta$ be the regression parameter in the candidate model, and let $\hat{\beta}_n$ denote its maximum likelihood estimator based on $n$ point observations. Suppose that $[?]$…