Social Workers Test Gap: La Griffe du Lion's Fundamental Constant of Sociology Strikes Again
Print Friendly and PDF

From New Jersey Spotlight News:

Social workers rail against biased license exam


Hundreds joined meeting to complain after report confirmed their fears

Outraged social workers say a new report on the disparities of who passes their licensing exam confirms what many have long suspected.

“When I saw this report, all the 92 pages, I was so disgusted. I was so upset,” said one licensed clinical social worker at Rutgers University in Newark.

That social worker said she knew more nonwhites and older students were failing because she constantly heard from alumni who failed the exam. One of her students, a 55-year-old woman who is an immigrant, failed the exam five times before finally passing on her sixth attempt, the social worker said.

“I was outraged,” she said of what she read in the report.

Thursday’s virtual meeting with hundreds of registered participants follows the release earlier this month of a national report detailing large disparities in race and age among those who pass the licensing exam when taking the test for the first time.

Inherent bias

The report exposes the inherent bias within the test and shows, through data, the harm being done to the profession and marginalized communities, according to a newsletter from the New Jersey chapter of the National Association of Social Workers. Social work schools in New Jersey have been advocates in pushing for the release of the data from the Association of Social Work Boards, the New Jersey chapter said.

The report says that first-time pass rates have historically been highest for white test takers, averaging 83.9% between 2018 and 2021. In 2021, 85% of white test takers passed the clinical exam on their first attempt, while 46.2% of first-time Black test takers passed the clinical exam during the same period. The pass rate for multiracial test takers was 79.9%, while the pass rate for Asians was 72%. Hispanics passed by a rate of 65.1% and Native American and Indigenous peoples passed by 62.9% between 2018 and 2021.

I’m shocked, shocked to see this pattern of racial gaps on a cognitive test. In the 100+ years in which cognitive tests have a part of American life, has anybody see anything like this before?

“This whole ASWB exam, it’s institutional racism. It’s a fiasco,” said another social worker at the meeting who is an assistant professor of social work at Lehman College in New York.

An anonymous iSteve commenter responds:

This may be paywalled, but John McWhorter wrote a newsletter for the New York Times a week ago decrying the watering down of the credentialing test for social workers:

This caught my eye:

There is a petition circulating saying just that, based on the claim that the association’s clinical exam is biased because from 2018 to 2021 84 percent of white test-takers passed it the first time while only 45 percent of Black test-takers and 65 percent of Latino test-takers did. “These numbers are grossly disproportionate and demonstrate a failure in the exam’s design,” the petition states, adding that an “assertion that the problem lies with test-takers only reinforces the racism inherent to the test.”

The problem, they say, is that (1) the test is only given in English, and (2) the questions and answers are developed from a white pool of social workers.

The 84 percent caught my eye because it has meaning in statistics, so I decided to do a little mini-La-Griffe-du-Lion analysis of these three numbers below

86 is 68 plus 16, the percentage of a normal distribution above minus one standard deviation. So in other words, if 84 percent of whites passed the test, any white who scored as low as one standard deviation below the average score for whites passed, about five out of six white test takers.

So the pass line for this social work test is 1 SD below white average. Coincidentally, the average black IQ is 1 SD below the average white IQ. So you can immediately predict that in this test the black pass rate will be about 50 percent. Hispanics are about two thirds of an SD below whites, so I needed to consult a normal distribution calculator to determine that their predicted pass rate would be 63 percent.

So these results are completely expected. The predicted results closely track IQ. The results would be right on is you assume a black average IQ of 83 and a Hispanic average IQ of 91, not that far off from the standard 85 and 90 estimates.

There is no social work test that can be developed, even if optionally offered in Spanish and with questions developed by BIPOCs, that will achieve different results, just as the billions of dollars on the table (from UC, the Ivy League, and elsewhere) for an SAT replacement that will not have this gap has not resulted in any claimants (the Smarter Balanced test is the latest failed attempt). High stakes standardized written tests will always have this gap because such tests are cognitive and are IQ tests, perhaps with a little noise.

What La Griffe calls the Fundamental Constant of Sociology is the skeleton key that opens the black box of so many seeming mysteries of American life.

McWhorter thinks the getting blacks to put more books in the home and have more verbal parental-child interaction, using a larger vocabulary, will eliminate the gap. In other words, when black parents turn white and have white kids, the black-white gap will disappear.

Of course, you aren’t allowed to mention this in Wikipedia because the august conclave of Wikipedia obsessive editors have determined in their infinite wisdom that it’s all just “pseudoscience.”

[Comment at]

Print Friendly and PDF