Anderson's theorem states that for a symmetric log-concave measure on R^n, among all translates of a given symmetric convex set,the measure is maximized at the origin. We develop quantitative counterparts to this result for the Gaussian measure by investigating the quantity: 


\[
D_K(\theta, r) =
\frac{E[\|G + r\theta\|_K - \|G\|_K]}{E \|G\|_K},
\]
where $G$ is standard Gaussian, $K \subset R^n$ is a centrally symmetric convex body and $\theta $ is a direction. Based on a joint work with Gil Kur and Reese Pathak.