Bush thinks that homoskedasticity is unnatural and is going to ban any talk of it in government funded statistical studies. Only observations that are heteroskedastic will be allowed.
Cheers, Ken Hanly ----- Original Message ----- From: "Jurriaan Bendien" <[EMAIL PROTECTED]> To: <[EMAIL PROTECTED]> Sent: Thursday, February 26, 2004 8:34 PM Subject: Re: demo fervor > > <http://economics.about.com/cs/economicsglossary/g/heteroskedastic.htm> > > > > which tells us that such a beast does exist. and i thought i earned my > > math degree with higher grades than sonofabush! > > I think the definition isn't very good. Heteroskedasticity and > homoskedasticity really refer to the actual distribution pattern of a set of > observations in relation to a norm, as far as I remember. Suppose you wanted > to know what would be the best predictor of price movements. Well, you could > specify some relevant variables, and then some price vectors, and then you > could study how actual observed prices deviated from the level or path which > you have hypothesised. That is the type of analysis which Anwar Shaikh did > once in support of the 93% LTV, published in the Robert Langston memorial > volume. > > J.
