[IPAC-List] Guessing on the SAT/ACT

[Patrick McCoy]

Whether or not a program or test uses Z scores, T scores, etc. one might use a correction for guessing or not.  The decision to correct or not should not depend on the type of score disseminated.

Seems to me that an important consideration in deciding  one way or the other , in addition to the one mentioned in previous posts (i.e., possible differential proclivity to guess across groups), might be whether there is evidence that such corrections increase validity.  I don't ever recall encountering evidence that validity increases.

If validity does not increase why would one want to uses a technique that is not well received?!

Patrick McCoy,
Ottawa, Canada

 


>>> "Mark Hammer" <Mark.Hammer at psc-cfp.gc.ca> 5/3/2010 11:04 am >>>

Again, I hasten to add that there may be certain assumptions about the nature of the test underlying the appropriate use of such corrections/adjustments. (where "appropriate" means correcting for guessing  with no other untoward impact on test-scores)

Most scholastic tests are speed tests that end up using deviation scores.  A great many employment tests are not converted to deviation scores, but are used in raw form.  Is the impact of corrections likely to be different under each circumstance, hence contra-indicated under one of them?

Here I defer to my colleagues whose statistical/psychometric knives are in a sharper state than my own.

Mark Hammer

_______________________________________________________
IPAC-List
IPAC-List at ipacweb.org 
http://www.ipacweb.org/mailman/listinfo/ipac-list