Date:   	Mon, 23 Oct 2006 19:09:32 -0500
From:  	Rudy Guerra <rguerra@rice.edu>
To:  	Darlene Goldstein <darlene.goldstein@epfl.ch>, Rudy Guerra <rguerra@rice.edu>
Subject:  	lewis edits
Part(s):  	
	2	lewis_edits.pdf	application/pdf	11361.17 KB	Download

Attached are edits for Lewis. I'm sending only
those pages for which I have written edits.

One MAJOR concern I have is their discussion of
power on p.13. Power is a function of a number
of things, but ultimately power is a function
of the parameter of interest. In their discussion,
including that of Fig 1, they seem to be saying
that power is a function of the test statistic=sum
of ranks! True, one needs a test statistic, but a
power function is a function of a parameter, not a
test statistic. I hope I'm misreading things or that
the writing is in need of clarification, but I've gone
over this many times and I don't get it. I also
have misgivings about their multiple comparison
adjustments (global and suggestive, p.6). Again, it
may just need clarification.

My handwriting can be hard to read. Let me know if you
have questions. Send an email and I'll clarify.

I'm actually excited! I really like working with you.

Rudy