What is the effect of increasing the sample size $n$ on the standard error of the OLS slope estimator $\hat{\beta}_1$, assuming the range of $X$ is held constant?