To choose one model you should use Cox test of separate hypothesis (see
Pesaran 1974 On the general problem of model choice Review of Economic
Studies 41,p153-171)
Basilio
2010/11/10 Fitzmaurice, Ann E. <[log in to unmask]>
> Hi
>
> I have three variable sad (% independent variable) and eith mmr or logmmr
> as the dependent
>
> When I use regression using the logmmr compared to mmr, I get a much
> higher r squared ,indicating that the model fits better
>
> I am also interested in countries which are outwith 1.95 sd deviations and
> this is where I get a problem
>
> Using mmr as dependent I get a list of countries outwith 1.95 sd (either
> plus or minus) , but get a different list when using the log of the mmr
>
> Can anyone clarify this, what I am trying to do is find out which
> countries are estimated to do better or worse than a published estimate for
> a series of attributes, sad being one of these
>
> Comments appreciated
>
> Ann
>
>
>
> The University of Aberdeen is a charity registered in Scotland, No
> SC013683.
>
> You may leave the list at any time by sending the command
>
> SIGNOFF allstat
>
> to [log in to unmask], leaving the subject line blank.
>
You may leave the list at any time by sending the command
SIGNOFF allstat
to [log in to unmask], leaving the subject line blank.
|