A new study demonstrates that the Examination for Professional Practice in Psychology, known as the EPPP, has differing fail and pass rates for different races, and that the difference is large enough for AfricanAmericans and Hispanics to fall into “disparate impact” discrimination, as described in Title VII of the Civil Rights Act of 1964.
The study, “Are demographic Variables Associated with Performance on the Examination for Professional Practice in Psychology (EPPP)?” is published in The Journal of Psychology: Interdisciplinary and Applied, October 22, 2018. The author, Brian Sharpless, PhD, is associate professor at the American School of Professional Psychology. To collect data, he used a Freedom of Information Act (FOIA) to request test results and demographics from the New York state board of psychology for its candidates.
Dr. Sharpless gathered data on 4892 applicants and first-time EPPP takers. He obtained “Records of all doctoral-level psychology licensure applicants from the previous 25 years with EPPP scores, gender, ethnicity, and degree type were requested.”
He found that Blacks had a failure rate of 38.50% and Hispanics had a failure rate of 35.60%. Whereas, Whites had a failure rate of 14.07% and Asians had a failure rate of 24%.
New York uses converted scores for the EPPP, from 0 to 100, with 75 as the passing score.
The differences in minority candidates’ selection rate violates what is known as the “four-fifths rule.” This means that the pass rate for minority groups fails to reach at least 80% of the pass rate for the majority group.
Typically, when a test has this impact, industrialorganizational psychologists exercise very careful methods to set cut scores, seek additional validity or research, and investigate possible replacements with less disparate impact.
Dr. Sharpless wrote, “… given the ethnic performance discrepancies and limited validity evidence, additional psychometric investigation of the EPPP appears warranted (e.g., in terms of criterion and predictive validity testing),” Sharpless also wrote. “Further, it is recommended that the EPPP Step-2 should undergo similar assessments prior to implementation.”
“Additional empirical attention should be devoted to the cut score (i.e., a scaled score of 500, roughly corresponding to 70% correct).” He noted that “…the determination of the ‘passing’ score is one of the most important, yet difficult, psychometric tasks in testing …”
And he noted that “… passage of the EPPP carries serious professional ramifications for applicants. There appears to have been limited discussion of the theoretical and/or empirical justifications for the current cut off score in the publicly available EPPP literatures…”
While regulatory boards aren’t employers, the principles of employee selection may apply. Employers using tests often fall under Title VII of the Civil Rights Act of 1964 which prohibits employers from using neutral tests or selection procedures that have the effect of disproportionately excluding persons based on race, or other protected characteristics. Test
New Study Shows That EPPP Can Discriminate,developers and users must demonstrate that the use is necessary and related to relevant characteristics.
Industrial-organizational psychologists, who help companies show business necessity when a company is paying for employee performance, with a legitimate business goals of seeking top performing employees, may point out that regulatory boards do not have this same business necessity.
Regulatory boards concern themselves with the lower end of the distribution of performers, those likely to exhibit gross negligence or in some way endanger the public. Denying a license to a candidate because they perform at the average range, or even below average, might conceivably violate that candidate’s property rights and have no impact on safety.
Dr. Sharpless noted similar issues. “… if the EPPP is found to lack acceptable validity evidence (or if a decision is made to not submit the measure to further empirical testing), then it will remain open to charges of being a potentially arbitrary barrier in an already protracted path to professional independence…” Sharpless wrote.
He notes the lack of Blacks and Hispanic psychologists and suggests that these issues could be related.
And, he said, “… psychologists have always been at the forefront of developing tests of individual differences with valid and reliable scores…” And he wrote, “A case could be made that psychology gained recognition, as well as a more coherent professional identity, through such testing efforts. Therefore, it only makes sense that we submit our own licensing exam to these same high levels of scientific scrutiny.”
Brian A. Sharpless is an associate professor at the American School of Professional Psychology at Argosy University, Northern Virginia. He received his PhD in clinical psychology and MA in philosophy from Penn State University and completed post-doctoral fellowships at the University of Pennsylvania.
To link to the article go to: https://doi.org/10.1080/00223 980.2018.1504739
Hmm…i just discovered this b/c of some similar article i saw today. Didn’t know about this before. There is a trend of illogical thinking that i see basically everyday now in the media. It’s the proportionality argument that all groups must be equal based on numbers and if not, then it points to “discrimination” – usually these are thinly veiled political diatribe articles that go very far beyond the article title and make tns of claims not supported by the data or the article. So here, i guess we have established differences in groups based on race. But all psychologists, especially researchers that aren’t being robotic woke progressive types, would put on their PHD brain that got them through graduate school. And the next thought would be…”what could account for this differences in an OBJECTIVE TEST?” And then they might brainstorm. Here i will brainstorm and think aloud for a few sentences:
1. Do certain groups purchase test prep and other groups don’t? In other words is test prep the source of fail and do those that pass have test prep courses/tutoring?
2. Are certain schools disproportionately represented in the failures? Did more of these students attend certain schools?
3. Is there a regional or state effect?
4. Is there an effect for age?
5. Are there differences in when they are taking the test relative to graduation? Or completion of hours?
6. Are there any differences in working status – full or part time?
7. Any difference in date test was taken relative to completion of PHD?
There’s been no evidence of any racial bias in the test. I think it’s logically flawed and scientifically incorrect to assume this without any data.