Date: Nov 28, 2012 7:49 PM
Author: Halitsky
Subject: Finally! Pay-off for all that work I did with the "A" matrix returned<br> by Ivor Welch's module!

Thanks for this:

sqrt[ var(a1) + 4*var(a2)*(mean_x)^2 + 4*cov(a1,a2)*mean_x ],
with df = n-3.

I hope you're pleased to know that:

a) I recognize all the terms of that formula;

b) I compute and report all of them already, as part of the existing
code;

I'm also very glad that a cov(i,j) for i<>j finally showed up - it
makes the
matrix manipulation of Ivor's "A" all worthwhile!

Will modify the code to report out the average slope SE next to the
average slope
and repost the last two sets of results with SE's included.

Then, I will "ring the changes" across all the remaining cells for c
on (u', u'^2).

Then, I will do the 2-ways across lenInt and fold and per lenInt and
fold.

And then, if the p's seem to warrant it, I will do the Bonferroni
correction.

From a purely technical point of view, I think I may have merely
progessed into the second half of first semester undergraduate
statistics. But from what I've learned about how a statistician
actually "thinks" by watching your moves, I feel like I've been in a
graduate seminar. It's one thing to know the rules - it's entirely
another to know how and when to apply them.