EEOC Compliant?

If I hear one more person tell me their materials are EEOC compliant, I think I’m going to throw up! Time after time, people tell me this or that test is EEOC compliant. I give them the benefit of the doubt, check it out and discover it is just another error-filled piece of nonsense. I think many of these folks got their heads stuck in the press when their brochures were being printed. I personally have no idea what “EEOC Compliant” means outside of two documents. These are:

  1. The Uniform Guidelines on Employee Selection Procedures
  2. The Standards for Educational and Psychological Testing

If you don’t know both of these documents by heart, I suggest you immediately quit recruiting and start an exciting career operating freight elevators. Tools of the Trade Contrary to popular rumor, I really don’t make up the things I write about. Sure I have a few personal opinions, but they refer to people who ignore science and law in favor of unsupported nonsense. Take, for example, “EEOC Compliant.” Apparently this is supposed to mean that XYZ test is “okay” to use because it “complies” with EEOC regulations. Well, that all sounds nice on the surface. But in my recollection, there is only one condition for when a test becomes “EEOC compliant” ó and that is one you would not like. Intrigued? Look at what the Uniform Guidelines and the Standards say (in a highly abridged version). The Uniform Guidelines on Employee Selection Procedures You don’t have to follow the Guidelines unless you have “adverse impact,” but the DOL recommends following them anyway.

  1. Adverse impact is defined as when the passing rate (i.e., hiring, promotion, etc.) for any small demographic group is less than 80% of the passing rate for the largest demographic group. In other words, suppose 150 aliens from the planet XNERFT applied for positions as airport shoe-sniffers (a newly created airline security job) and 75 were hired. If XNERFTians are the largest demographic group in the applicant pool, then every other demographic group is tested to see if it has at least a 40% pass rate (75/150 = 50% XNERFTian pass-rate x 80% EEOC test = 40%).
  2. “Applicant status” is still ill-defined, organizations are still expected to track adverse impact at each decision-making point. Does your applicant tracking system do that?
  3. It is okay to have adverse impact if people cannot meet job requirements based on job competencies and business necessity. That is, organizations do not have to hire unqualified people (whether they are XNERFTians with big noses, or not). But they do have to show a professional job analysis, use tests that are job related, demonstrate that test scores predict job performance, and engage in reasonable efforts to reduce adverse impact.
  4. Hiring challenges seldom get to court, but I would not take comfort in that. You may get to hire as many unqualified employees as you like, and the EEOC may eventually ask you to pay for their foolish on-the-job behavior. See

Now, I challenge anyone to show me which part of the Guidelines is a bad idea. The Standards for Educational and Psychological Testing While the Guidelines define adverse impact and best practices, the Standards define test development and test use.

Article Continues Below
  1. A test should be grounded in a sound theory of job performance. This does not mean any old training test can be transported into a hiring application. That would be like taking your favorite opposite-sex best friend home to meet your spouse. Some things are just not a good idea. Some examples of useful training (but horrible hiring) tests include personality types, communication styles, leadership styles, and so forth. Why? Because they were designed to predict differences, not job performance.
  2. A test should be reliable and valid. That means it should measure what it is supposed to measure, measure it consistently, and show scores that have a strong relationship with job performance.
  3. The test should be job related. That means it should be backed by a professional job analysis. You cannot give a knitting test to managers under the theory that managers have to “knit together” solutions. The same goes for many training tests. Fun for training, but a disaster in hiring. Conducting training does not make a person a test expert, hiring expert, or psychologist (although they may “play” one on the job). If you think this kind of ignorance is harmless, you might as well ask the training department to purchase raw materials without a proper education.

Now, I challenge anyone to show me which part of the Standards is a bad idea. What Does EEOC Compliant Mean?

  • As far as I know, the DOL does not have a test certification department. Responsible test use is still incumbent on users (not vendors). So that can’t be what they mean by “compliant.”
  • The ‘Guidelines’ clearly state that a vendor’s claims of validity or brochure data are turkey doodoo (well, they don’t say that exactly, but that’s what they mean). Anyway, claims are not “compliant.”
  • Organizations are free to commit any foolishness they desire until someone brings an adverse impact challenge. That, certainly can’t be what they mean by “compliant.”
  • If one company borrows another company’s job analysis and validation data, that is not “compliant.”

If none of these situations are compliant, then what is? Well, here is the answer most professionals folks give: You are in compliance if your company survives an adverse impact challenge and is found not guilty. Now, what’s with that vendor “EEOC compliant” stuff, again?