> On Tue, 12 Aug 2008 14:44:42 EDT, Jack Tomsky <jtomsky@ix.netcom.com> wrote: Consider the following linear model Y_i=beta0+beta1*X+eps_i where eps_i~Normal(0, sigma^2) We want to test Ho: beta1/sigma=1 I know how to conduct tests on linear combinations of beta0 and beta1, but I do not know how to conduct tests on beta1/sigma
> Consider the following linear model Y_i=beta0+beta1*X+eps_i where eps_i~Normal(0, sigma^2) We want to test Ho: beta1/sigma=1 I know how to conduct tests on linear combinations of beta0 and beta1, but I do not know how to conduct tests on beta1/sigma. I am guessing that it is possible to reparameterize and rewrite Ho. Maybe Y_i/sigma=beta0+beta1*X+eps2_i
On Aug 8, 12:50 pm, isabelle...@fastmail.fm wrote: Consider the following linear model Y_i=beta0+beta1*X+eps_i where eps_i~Normal(0, sigma^2) We want to test Ho: beta1/sigma=1 I know how to conduct tests on linear combinations of beta0 and beta1, but I do not know how to conduct tests on beta1/sigma. I am guessing that it is possible to reparameterize and rewrite Ho. Maybe
"Einar Andreas Rodland" <einarro@ifi.uio.no> wrote in message news:g8or1u$c70$1@readme.uio.no... However, I still challenge you to actually do the calculations, say with mu=1/sqrt(5), sigma=1, for a few different values of N, and explain the results...if you can. Just for the record, in case someone ever wants to know the results, and since I already have them... Computations run
On 22 Ago, 15:52, "lic...@hotmail.com" <lic...@hotmail.com> wrote: On 22 Ago, 12:43, "Einar Andreas Rodland" <eina...@ifi.uio.no> wrote: "Jean-Claude Arbaut" <jeanclaudearb...@orange.fr> wrote in message news:48ae99f8$0$967$ba4acef3@news.orange.fr... Einar Andreas Rodland a écrit : I'm not much interseted in Luis' troll, but anyway, I wonder : maybe
