Last Friday I suggested that we do Ch. 3 and a small part of Ch. 4 for 
this week, but I think it may be better to just plan to do Ch. 3 and look
at the "homework problem" from Ch. 2 (Problem 5 on pp. 38-39), and then
spend 2 weeks going through Ch. 4 (so spreading over 3 weeks what I
originally thought we should try to do in 2 weeks).

It's unclear to me what the best format is for the meetings, but I propose
that we try something a bit different this coming Friday.  Let's start
with a discussion of Ch. 3 (and then go to the Ch. 2 data analysis problem
afterwards).  Let's see how it goes if we have different people serve as
discussion leaders for the various sections.  (Note: Some sections are
rather short and there may not be much to discuss.)

Each discussion leader can start by giving a 2 to 5 sentence description
of what the section is about, and then perhaps mention a few specific
points to discuss (and we'll discuss them as we go along).  Try to have at
least one question that you know the answer to ask the others who may not
have read the section as carefully, and picked up on a subtle point
pertaining to it.  Also, try to have a related question to ask the others
that you don't know the answer to.  (My guess that Harrell being unclear
will allow you to identify at least one such question.)  I'll make a point
to have one question for each discussion leader about the section.  I
think by discussing each section a bit in turn will allow us to pick up on
some points that a quick reading may have left undiscovered, and that the
questions and answers will help us all engage the material.  I believe
that the main point of a group like this is for each of us to try to get
truly comfortable with the material in the book.  (The more things you let
slip by now, the harder it will be to quickly learn new material in the
future.)

I'll make the following discussion leader assignments, giving the longer
sections to the seasoned Friday afternoon members, and the shorter
sections to the newbies.  (I'll let Jill off of the hook this week
completely, since her formal statistics background is less than what the
rest of us have.  But I expect that Jill will be able to come up with some
good questions --- like asking the discussion leader to put some passage
into his own words, hoping that they will be clearer than Harrell's.)

Sec 3.1 Charles P.
Sec 3.2 James W.
Sec 3.3 Mark L.
Sec 3.4 Jim S.
Sec 3.5 Arndt L. (This is a longish section, and perhaps it would be
           good to outline it --- help us break it down and identify 
           the main points.  (You can have an easier section next time.)) 
Sec 3.6 Allen K.
Sec 3.7 Jim G.
Sec 3.8 Don G.

The rest (as well as those listed above) can take a look at the problem
from Ch. 2.

Also, everyone can ask about points in the first 2 chapters that are
bothersome if we didn't get to them last week.

Let's see how the discussion leader idea works --- if you don't like it we
can consider some other format.

CDS