Example: Military Psychology, Criterion Validity, and Ethics.
Suppose you are part of a military program examining the appropriateness of assignments for new recruits. The goal is to eventually be able to place new personnel in positions in which they will be most likely to perform well.
To this end, you decide to take personnel aleady assign to a particular job, such as radar operator, and examine their current job performance (as measured by evaluations made by supervisors) -- the criterion. Then, you have them perform several brief tasks that you think might be predictive of job perfomance (possible predictors).
You find that one of the tasks is a good predictor of job performance. However, after further examination of the results, you find that this task is only a good predictor for some people -- white males. You results show that, for women, and for men of other ethnic origins, performance on the task does not predict job performance.
What do you do? Do you go ahead and use the measure to decide whether new personnel will be assigned as radar operators? Would there be any ethical concerns in doing so?
It would be really hard to justify testing everyone with that system and applying it for assignments or promotions. It would then be prejudicial, but is it bad to use on just white males? So long as you develop new tests for everyone else, would that be bad? I wouldn't think so, but I'm curious.
Sorry, I forgot to put that that last message was for the military group in general (Schroeder and Rumian and Me.). And, another thing, is that if we did just use the test for white males the cost of developing a test for many other groups individually might be crazy. So, to keep costs down, the best idea is to try to get an all inclusive test. But, if one can't be found, the test is valid for white males, so using it on them is not bad, is it?
This is the 1990s and you can not use tests for job performance that only predict outcomes for the majority. Disrcrimination has become a household word in the past few decades and it is not something to take lightly. If this test is only a good predictor of outcomes when white males are involved then it is an unfair predictor. Whether it predicts that minorities will do poorly when they actually would have had high performance or the other way around, it is still unfair. It is unfair to the person taking the test when it predicts they will do poorly and they really would have been good at the job. And it is unfair to the government (or any company that has a test like this) when the test predicts the person will do well when their actual performance is quite low. They need to spend the time and resources to provide adequate tests for all people, whether the tests are different or not. This will help them in the long run, especially since women and minorities could sue for discrimination if not hired because of the faulty test. Also, hiring the wrong person could be very costly.
Irvin, Karam, Freeman