posted 01-29-2003 11:46 AM            

The latest buzz is of course the NAS report. We have already discussed that and got some good information and ideas. One of the problems as stated in the report was that more research was needed on screening tests.

Questions:

1) How would you set up a research project
for employment screening tests.

a. would it be a mock situation?
b. If a mock situation how would you
make it important to the subject
or would yo have to?

2) Could research be done on a body of
pre- employment screenings already done?

a. How could you establish ground truth?
Even with admissions in the post
test you would not know if he said
everything.

b. Would it be enough to say admissions
were made, so the deceptive result was
confirmed? Maybe

This is something we all need to think about. I am not a researcher and would probably not set a research project up properly but it is true that we need more research in this area.

I don't think it will be acceptable much longer to say we are using a technique that has been validated and researched for specific tests so we know it will work for screening. We are going to have to come up with a way to research different techniques directly related to screening tests.

So lets here everyone's ideas!

Jack

------------------

posted 01-29-2003 11:25 PM         

As you indicate more research is needed in the area of screeing tests, something "everyone seems to know, but does nothing about." Personally, since DODPi and other government agencies utilize screening as they do (and receive some funding), I am rather disappointed a research project was not done in the past--not does there seem to be one on the horizon. However, research is never too late.

What could be interesting is to include a member of the NAS board to help design/ participate in the research project.

As a quick comment,a lab/mock study should probably be performed first, as it may be the easiest to deal with- say pre-employment screening. The major problem of this type of study, as you also noted, is the establishment of the Control group (because we all know establishing 'ground truth' "ain't that easy."

I often think there are enough Ex-cons out there who admit their guilt and their case has an overabundance of corrobarrative evidence- (and who have gone straight enough afterwards)- that would be willing to particpate in such a study for a little reimbursement. Maybe fill out a "pre-employment" for a ficticious law enforcement agency wherein they exclude their crimes of conviction.

But where and how do we come up with a 'control' group (the known standard so to speak) ?"

At this point I don't have anything else I can really add to your post, as a lot more thought obviously has to go into it.

I just wanted to jot a note so that you knew someone was taking time to read the messages.

posted 01-30-2003 08:02 AM            

I wrote the post because it seems to me that if we are ever going to gain acceptance in the scientific field we have to play by their rules. I am perfectly willing to help in any way I can but have no background or education in research.

I do not buy it when people not in the profession say examiners are biased and any research done by them is flawed. We are just as interested in seeing the job done right as anyone else, maybe more. It seems to me that anyone who has a theory and researches those beliefs, if ethical, would strive to be completely unbiased. After all, good research should stand the test of replication. If the results of an experiment can't be replicated the study is worthless.

Someone tell me how we can properly research pre-employment screening tests. Personally, I believe these test can be accurate. However it is not the test but the information gained in the process that is important.

Jack

posted 02-02-2003 09:45 PM            

Ground truth is much easier to obtain in a laboratory setting but even here it can be elusive. One problem with using a confession as ground truth is that it is thought to taint the independence of the results. Inasmuch as it is standard practice that the examiner whom administered the exam also conducts the post test interrogation, the results may be claimed to be based in part on the information received in the post test. Another area that may be less prone to criticism is the use of other conclusive forensic evidence to establish ground truth (i.e. DNA and fingerprints).

The amount of variables possibly present in any given polygraph examination may be unending (given the fact that the test is administered on a human and by a human). For instance, analog research has shown that innocent subjects may be attempting countermeasures even when not instructed to.