z40\doc\web\2000\03\sdmath.txt
From:           	"Patricia Hausman" <ph@lakeside-consulting.com>
To:             	"La Griffe du Lion" <il_leone_uno@yahoo.com>, <scottpost@aol.com>,
       	<h-bd@egroups.com>
Date sent:      	Sat, 11 Mar 2000 14:58:39 -0500
Subject:        	[h-bd] Re: Griffe on girls and math
> Griffe wrote:
> 
> The distributions of math aptitude for men and women are separated by
> about 0.3 SD. Consequently, approximately 37 percent of the top 10
> percent of the test takers will be female. The precise nature of the
> test is not important, because the rank order of finish will be
> determined by those questions, however few, that require cognitive
> ability to answer.


**********


Griffe,


This value of 0.3 SD is for undergraduate admissions.  For 
graduate admissions tests the difference favoring males averages 
0.6 SD when results from MCAT, GMAT, and a couple of relevant 
GREs are averaged.


For the GRE-math, the difference is almost .9 SD--and that value
may be a tad biased against males because fewer females than 
males take the test.  This raises the possibility that relative to
all females,  women who take the GRE-math are a more elite
group than is the case for the males.


But never mind the quibbling.  For the sake of argument let's 
assume that the F/M ratio is 1 for the GRE-math.   And let's
pretend that I'm 21 and unmarried and like to go out with 
males when I'm not busy entertaining myself with a good book
on differential equations.    What does an effect size of .9 mean?


I say it means that if I were among the top decile of the test-
takers and were invited into a room with all the others who 
also scored in that range,   I could be rather picky about who 
I went out with any night of the week.


No?


Regards,
Patti