Bryan Caplan  

From the Handbook of the Sociology of Education

Bricks, Mortar, and Education... Keynes a la Mode...
[A]lthough employment tests are often biased against minorities (National Research Council, 1989), employers who use them are more likely to hire minorities than employers who lack such tests (Neckerman & Kirschenman, 1991).  Testing may be biased, but the alternatives to testing are worse.
(From Chapter 18, "Interactions Between High Schools and Labor Markets," by James Rosenbaum and Stephanie Jones)

Comments and Sharing

COMMENTS (7 to date)
Mik writes:

New study for your book on education (indicting that education causally increases IQ).

Roger Sweeny writes:

And why would employers who use tests be more likely to hire minorities than employers who don't? Perhaps because employers who don't use tests put more emphasis on diplomas and degrees, and minorities do even worse on that account.

Perhaps education, instead of being a way up, is now more likely to be part of what the Marxians call "the intergenerational transmission of inequality."

It certainly provides a way for every judge and bureaucrat to say, "Of course, I deserve to make more money than him. I'm much better educated."

John Thacker writes:

Roger Sweeney--

and surely some employers who don't use tests rely on word-of-mouth of existing employees and nepotism, another area where minorities do even worse. (Except in new businesses started by immigrants that hire "their own," as seen in such odd patterns as Chinese laundries, Indian motel managers, and the Somali taxicab drivers of Minneapolis.)

DHM writes:

Are employment tests biased against minorities? Or are they biased against the badly educated, and our public school system is more likely to defraud minorities of an education than to provide minorities with skills in literacy and basic math?

I would suggest that those who use tests are more likely to hire minorities than those who don't for an unusual reason. They mainly care about the potential employee's ability to read, write, and cipher. At some level they recognize that a minority student who has successfully learned to do those things in spite of the roadblocks public school systems place in the paths of the disadvantaged is an employee worth having.

Matt writes:

Am I missing something? How can tests biased against minorities lead to more of them being hired? Shouldn't it be that the test are biased for minorities? This is assuming no test can be perfectly objective.

John Markley writes:


He's not talking about employment tests biased against minorities compared to an unbiased test, but rather compared to dispensing with testing entirely and going with other criteria. The implication is that the sort of selection methods typically used in the absence of any test tend to be even more biased than existing tests. The alternate criteria typically used in the absence of tests commonly involve things that are positively correlated with being white such as formal credentials or preexisting social ties with someone who can help you get the job, as John Thacker and Roger Sweeney said. In addition, dispensing with testing makes subjective criteria such as how likable and charming the employer or job interviewer finds you to be a bigger factor, which means that any racist attitudes among those people will have more influence than otherwise.

Ghost of Christmas Past writes:

In point of fact, "employment tests" of cognitive abilities (or the popular pencil-and-paper "morality" tests which attempt to detect people with poor morals but also function as cognitive tests because smarter people suss them out better) are not "biased against minorities." They predict job performance equally well for applicants of all races. Such tests do, however, produce copious amounts of "disparate impact" because non-asian (mostly black and "hispanic") minorities are significantly less intelligent on average than whites and asians (of course asians are minorities, too-- and whites are in some places, such as Southern California).

Since "employment tests" are not biased, employers who rely upon them (even in part) have objective reasons to choose one job applicant over another. By using an unbiased assessment tool, such employers mute the effect of real bias (that is to say, personal prejudice) leading to less biased outcomes overall!

Naturally the American Left is deeply opposed to unbiased evaluation of job candidates because of "disparate impact" of entirely race-neutral cognitive tests, even though unbiased evaluation is actually much more advantageous to minority applicants-- for example, getting a high score on an employer's own test means a minority applicant won't have his other credentials (e.g., diploma) discounted by an interviewer who just assumes it was awarded for "affirmative action" points rather than achievement.

Comments for this entry have been closed
Return to top