Hi

James M. Clark
Professor of Psychology
204-786-9757
204-774-4134 Fax
[EMAIL PROTECTED]

>>> "Mike Palij" <[EMAIL PROTECTED]> 01-Jun-08 9:07:56 AM >>>
(2)  I'm curious as to why the chosen correlation for SAT
and GPA was rho=0.71. Sackett et al report a estimated
value of 0.55 (corrected for range and other "problems" 
and based on a sample of N=165,000) on page 218.  
Was it just because the r-squared would give a percentage 
of variance accounted for of 0.50?  Of course, the larger 
the value of rho, the greater the effect of selecting on SAT.

JC

Yes, I just know from experience that using .7071 for the shared and unique 
variance are the appropriate values, so that is what I often start with in 
playing around with some scenario.  If you want to try alternative figures, 
just substitute "rho" for the shared variance and "sqrt(1-rho**2)" for the 
unshared in the line that generates gpa scores.  I wasn't trying to match any 
real-life situation, which has lots of complications as Mike has noted.  I was 
primarily curious about the drop in SD for Y as well as X, and subsequently in 
the accuracy of the standard correction procedure. As to whether the current 
Sackett estimate of .55 corrected for range is correct, I wonder if this is one 
of the areas that would be affected by Schmidt et al's concern about the 
under-correction using the standard correction, which assumes direct 
restriction of range rather than indirect restriction?  And issues like 
drop-outs and course selection are only sometimes considered.  The intra-course 
estimate of the correlation is higher than one that ignores course selection, 
suggesting weaker students choose easier courses.  I also saw a reference to, 
but did not read, something on estimating grades for students who drop courses. 
 I don't know if these are some of the "problems" that were compensated for in 
the Sackett et al estimate.

One of the major issues in the literature is the assumption of linearity.  If 
the true relationship is nonlinear then it would seem that almost anything is 
possible.  Definitely one of those messy areas!

Take care
Jim


---
To make changes to your subscription contact:

Bill Southerly ([EMAIL PROTECTED])

Reply via email to