"Room for improvement" refers to the potential incremental validity of a
predictor (e.g., performance during an interview). If the interview has no
impact on the quality of employees recruited, then it has little incremental
validity. Obviously, if baserates are high--anyone can run the french-fry
machine--then an interview would provide no more value than a stethoscope.
An interview might be more helpful for a job that requires adaptability,
intelligence, and interpersonal skills (college professor?). However, given
that the baserates for these characteristics are so low, would you want to
go directly from the stethoscope to the interview stage? Time and cost
would limit the number of applicants that would be seen and, as these
resources run out, you might not ever get to the genuinely qualified
applicants. So, how do you raise the baserates closer to the ideal 50%
level? Ask for transcripts, letters of recommendation, cvs, and statements
of career goals. With all of this relatively cheap information, an initial
screening can be done so that you are reasonably sure that the final pool of
applicants to be interviewed includes an adequate number of qualified
people.
This strategy is also used in medical testing. Someone who experiences
occasional chest pains will usually not be scheduled for an immediate
angiogram. More likely, an exercise EKG would be performed to see if the
chance of a serious problem justifies use of the more expensive (and
invasive) procedure.
--
Marie Helweg-Larsen wrote:
> I have an I/O related question about baserates. Baserates refer to the
> percentage of applicants who would be successful if all of them were
> hired. Obviously this number can vary anywhere from 0% to 100%
> depending on the job. I also understand that a baserate of 50% gives
> the most room for improvement if implementing a specific selection
> device. What I don't understand is the significance of this number. It
> sounds as if you don't have much control over the number. So other than
> just noting the statistical fact that there is most room for improvement
> around 50% what practical implications does it have? Would you want to
> have a selection device (say for college admissions) so that the
> baserate was 50%? It doesn't seem to make much sense to me.
> Help please!
> Marie
********* http://www.coe.uca.edu/psych/scoles/index.html ********
* Mike Scoles * [EMAIL PROTECTED] *
* Department of Psychology * voice: (501) 450-5418 *
* University of Central Arkansas * fax: (501) 450-5424 *
* Conway, AR 72035-0001 * *
*****************************************************************
