QM1- CLT

[azure]

Just wondering, will taking a larger sample size always result in greater accuracy/better approximation? Will a larger sample size always result in a more "normal distribution?" Are there any exceptions?

Having a bit of trouble understanding CLT.

[lilaznkev1n]

Sorry this reply is kinda late.
An exception is that it may not be true when sampling from two different populations and determining which one gives a better approximation. For example, Y has 3000 observations while X has 2500. It may turn out that Y is extremely non-normal, whereas X is only slightly non-normal, in which case assuming normality for X would give a better approximation of true parameters although X has less observations.
However, if both X and Y are distributed in approximately the same manner, then it is true in saying that a larger sample size would provide more accuracy when assuming normality.

[azure]

Hey, thanks for the reply! Understand it now