Do We Need a Large Sample Size?

A smaller n is much more susceptible to influential observations. If you have one or two abherrant cases, they're going to exert a much greater influence with an en of ten than an en of 200. The big thing in my eyes is sample, test statistics don't follow distributions that we think they do. You see, patrick's obsession with ki riston in los angeles, aki resal has deepened to the point where he now slips up ond calls the ki square distribution the central ki distribution. We cau thes paraproxis very interesting, very z distribution, that when we assume those distributions are coming on line, they actually don't.