>
> On Wed, 27 Jan 1999 [EMAIL PROTECTED] wrote:
>
> > At the last SRCD, Dr. Sternberg was questioned on the generalizability of his
> > findings given the sample for this study. His sample consisted of Yale
> > graduate students. It is possible that, because of the high scores of the
> > students they accept, a ceiling effect of sorts is present. The predictive
> > validity of GRE's in this context may be different with average as opposed to
> > very high scorers.
> >
Well, I'm no psychometrician but I think it's more than just possible,
it's a fact, and not just for Yale.
Any time you have a distribution of scores which are intended to
predict something, such as success in graduate school, and you toss
out all those who score less than a particular criterion, the
predictive validity for those remaining will drop sharply. If Dr.
Sternberg really wants to see how good the GRE is for selecting
students, he should get his colleagues to let _everyone_ into Yale.
Then we could at last see whether GRE scores are meaningful or not.
And I agree with Annette that lurkers should give their name if they
want to do more than lurk.
-Stephen
------------------------------------------------------------------------
Stephen Black, Ph.D. tel: (819) 822-9600 ext 2470
Department of Psychology fax: (819) 822-9661
Bishop's University e-mail: [EMAIL PROTECTED]
Lennoxville, QC
J1M 1Z7
Canada Department web page at http://www.ubishops.ca/ccc/div/soc/psy
------------------------------------------------------------------------
