Date: Dec 14, 2012 6:53 PM
Author: Ray Koopman
Subject: Re: Re linearity of the Axxxx SE plots – hold on t<br>	o your hat

On Dec 14, 1:07 pm, djh <halitsk...@att.net> wrote:
> In your post of 12/14 at227 pm, you wrote:
>
> ?So far the all the plots look pretty linear, with just a hint of
> positive curvature, but it's hard to say because SEs are themselves
> heteroscedastic.?
>
> The Axxxx SEplots I?m going to send at set 2 and set 3 may be relevant
> to this matter, for the reason(s) below.
>
> As soon as I started the Axxxx SE plots for set 2 (none of which
> you?ve yet received), I immediately saw something which you can see
> for yourself by plotting the two tables at the end of this post, which
> are for:
>
> a) N_1_a1_S (you already have the plot for this one)
> b) N_2_a1_S (haven?t sent this one to you yet)


I'm sending you my plots offline. The biggest difference I see is that
the SE's in set 2 are more consistent than those in set 1, especially
for L > 90 or so. Hoever, that could be just a sample-size artifact.
What are you seeing?