
Re: CV of square
Posted:
Nov 14, 2012 2:56 PM


Rich's rule works well for the case where T is a chisquare variable, so that S is a chisquare variable. Here is the ratio CV(T)/CV(S) for df=1,2,...,10. The ratio is typically close to 2, as Rich predicts, and it gets closer to 2 as df increases.
{1.87186, 1.91306, 1.93475, 1.94796, 1.95678, 1.96308, 1.96778, 1.97143, 1.97435, 1.97672}

