Date: Dec 13, 2012 2:52 AM
Author: Ray Koopman
Subject: Re: I need to correct an apparent miscommunication r<br>	egarding derivation of het H’s and L’s

On Dec 12, 7:10 pm, djh <halitsk...@att.net> wrote:
> In point (2) of your last post of 12/12 at 6:57pm ssm time, you asked:
>
> ?2. [...]Are the values whose median you get averages over all
> Lengths? [...]
>
> No. I?m sorry my earlier posts on this matter were insufficiently
> clear.
>
> To try and clarify, each of the twelve values that underlie the twelve
> H?s and L?s in this row of the het table:
>
> a1 a3 b1 b47 c1 c2
> C S C S C S C S C S C S
> Het
> 1N Aubqe H H L L H H H H L L L L 0
>
> is the slope of the regression of all values of Aubqe against all
> values of Length for each of the twelve fold x subset combinations |
> set=1, method=N.
>
> To make this statement perfectly clear:
>
> I.
>
> The twelve slope values underlying the H?s and L?s in the above het
> row are:
>
> Slopes of Regressions of Aubqe on Length (L)
> for each Fold x Subset | Set 1, Method N
>
> [...]


Got it. Plotting Aubqe against L suggests that the relation may be
heteroscedastic. However, it is complicated by the fact that the Aubqe
values have different standard errors. Could you post the same data
plus the SE for each Aubqe? Or is all that somewhere in one of those
zip files you sent offline?