On Wed, 24 Jan 2001 13:32:56 -0400, "Richard A. Beldin"
<[EMAIL PROTECTED]> wrote:
> The transform from Z scores to probits uses a constant 6 translation,
> not 5. I don't know if that solves your problem, but it might eliminate
> some confusion.
WHOSE confusion?
I learned my probits from the classical textbooks in bioassay,
the ones by Finney; and Finney uses the add-on of 5.
Which, despite that note, is what I think everyone uses.
--------
As far as the eventual results are concerned, the probit and
the Logit are practically the same. Logit is popular because it
fits many natural models; the math models are powerful; and
you don't need a table to look up the logit (log of P over 1-P ).
Zero (or 100%) is a problem for either transformation.
Folks sometimes adjust their cell-totals by the equivalent
of "1/2 of an event."
--
Rich Ulrich, [EMAIL PROTECTED]
http://www.pitt.edu/~wpilib/index.html
=================================================================
Instructions for joining and leaving this list and remarks about
the problem of INAPPROPRIATE MESSAGES are available at
http://jse.stat.ncsu.edu/
=================================================================
