Home > data science, modeling, rant > Workplace Personality Tests: a Cynical View

Workplace Personality Tests: a Cynical View

April 16, 2015

There’s a frightening article in the Wall Street Journal by Lauren Weber about personality tests people are now forced to take to get shitty jobs in customer calling centers and the like. Some statistics from the article include: 8 out of 10 of the top private employers use such tests, and 57% of employers overall in 2013, a steep rise from previous years.

The questions are meant to be ambiguous so you can’t game them if you are an applicant. For example, yes or no: “I have never understood why some people find abstract art appealing.”

At the end of the test, you get a red light, a yellow light, or a green light. Red lighted people never get an interview, and yellow lighted may or may not. Companies cited in the article use the tests to disqualify more than half their applicants without ever talking to them in person.

The argument for these tests is that, after deploying them, turnover has gone down by 25% since 2000. The people who make and sell personality tests say this is because they’re controlling for personality type and “company fit.”

I have another theory about why people no longer leave shitty jobs, though. First of all, the recession has made people’s economic lives extremely precarious. Nobody wants to lose a job. Second of all, now that everyone is using arbitrary personality tests, the power of the worker to walk off the job and get another job the next week has gone down. By the way, the usage of personality tests seems to correlate with a longer waiting period between applying and starting work, so there’s that disincentive as well.

Workplace personality tests are nothing more than voodoo management tools that empower employers. In fact I’ve compared them in the past to modern day phrenology, and I haven’t seen any reason to change my mind since then. The real “metric of success” for these models is the fact that employers who use them can fire a good portion of their HR teams.

Categories: data science, modeling, rant
  1. April 16, 2015 at 8:24 am

    The commoditization of human beings continues unabated, with the main goal of corporations no longer being quality or service (as when I grew up), but cost-cutting.
    Reminds me of the automated phone-answering services most businesses employ today… great cost-cutters and time-savers for them, but often destroyers of actual customer-service.


  2. DJ
    April 16, 2015 at 9:03 am
    • Smelled
      April 16, 2015 at 6:11 pm



  3. April 16, 2015 at 9:18 am

    And now, Cathy, this just crossed my screen (of possible related interest):


  4. JSE
    April 16, 2015 at 10:26 am

    Exactly what I’d expect an ENTJ like you to say.


  5. April 16, 2015 at 11:31 am

    Maybe this is why Google is doing a study to find out how humans work-I wrote about it a while back,they definitely have algorithms in check but they made changes in their own HR hiring policies:)


  6. l haughton
    April 16, 2015 at 1:08 pm

    Just finished “On being a data skeptic” and watched your PBS interview. Both made me like you a lot.

    This post though… not so much.

    But I noticed you filed it under “rant.” Ok then. I think everybody deserves to vent unscientifically now and then.

    I look forward to your new book.


    • April 16, 2015 at 2:33 pm

      How would I be scientific in a case of a secret and proprietary model suck as this?


      • l haughton
        April 16, 2015 at 3:13 pm

        You’re right. Accept my apology. While I read the original story at the WSJ I didn’t see it was “a case of a secret and proprietary model suck”.


  7. quantummechanic1964
    April 16, 2015 at 1:44 pm

    As with all tools, there have to be checks and validations. Knowing what I do, I’d be very wary of any place using these, as they are easily abused (intentionally or not), and frankly difficult and expensive to get right.

    Eons ago, as an undergrad, I worked for some psychologists developing one of these personality inventories. (At the time, the going survey was the MMPI, which was apparently developed from a population of mental health patients, and mis-applied to the typical candidate pool, for which its validity was questionable, if not laughable.) Using data on existing employees and interviewees, combined with data from larger cross-sections of the population, they developed models to bubble the most probable candidates to the top of the interview list.

    Interviews cost money, while administering a pencil-and-paper, or these days, computer-based test, are pretty cheap. Combine that with requirements to prove non-discrimination across several dimensions. So the models were checked for as many of these things as they could manage. Of course, this included designing the thresholds so that not too many good candidates were skipped, and not too many bad candidates were included.

    What was interesting about the models, to me anyway, is that they never relied completely on the test results. There was always some practical, job-related activity that also got scored, whether it was packing shapes into a defined region (ala grocery sacker), or planning a cross-country route (ala long haul trucker), or some physical abilities (ala Navy EOD divers).

    Also, the generic test had a “validation” marker, a set of questions that, if you scored below, I think it was 70%, marked you as an outlier, and the test models were invalid. Each customized model had an additional validation marker, which might replace, or be used alongside of the generic one. Here’s a place for abuse — outliers might always get skipped, when in fact they might be the next genius intern.

    The goal of all of this was to save the clients (that is, the employers) time and money by skipping the interview for people very unlikely to be hired, and hiring people likely to stay longer (either not quit soon, or not get fired).

    But it also cost a lot of money to develop the customized versions, and only made sense for very large companies with high turnover rates, especially those that had a high training cost for new hires, where washouts were very expensive.

    How to game the system? If the test is based on a well-studied test, you can probably take it by applying to several other employers, or perhaps find a university study that’s using it. Read some papers on how the generic test was developed and used. Find out how the questions cluster along different dimensions. Figure out what the job needs, and decide how to answer based on that. Ask other candidates that applied to the same company how they answered. Better yet, if you know someone who was interviewed, answer the questions as they would.

    The grunts working for the psychologists knew the answers to that test so well, they could all pass themselves off as anything they wanted to, at least on paper.


  8. Gordon
    April 16, 2015 at 2:50 pm

    Aren’t you only dealing with one part of the issue when you say that turnover has gone down b/c economic uncertainty has reduced workers’ willingness to leave a bad job, etc.? Presumably some portion of turnover is also due to companies firing their staff, and this number might have declined thanks to better selection policies. I have no idea whether it has or not, but the possibility has to at least be in the consideration set.


  9. Brent
    April 16, 2015 at 4:18 pm

    We crazy humans are so complex that it’s going to take a while to crack this nut. In a couple decades we’ll look back and realize we were curing diseases with leeches. But we have to start somewhere (though I also hate the types of statements to which you refer).

    What you have to consider is how often someone gets hired because they gave the interviewer a semi-boner, had a confident smile, or were able to find a tenuous connection between the two of them. As bad as most of these personality ‘tests’ are, most can’t help but be better than the irrational gut of the even shittier manager hiring the worker for the shitty job.


  10. April 16, 2015 at 6:39 pm

    You are probably right Workplace personality tests are nothing more than voodoo management tools. They don’t really empower employers though. We shouldn’t forget the human interviews kinda suck too.


  11. toto
    April 17, 2015 at 1:56 am

    Thought you might like this : https://www.youtube.com/watch?v=hTkSlJRprAc


  12. Zathras
    April 17, 2015 at 8:21 am

    Yes, these tests are nothing more than voodoo management tools. Which makes them no better, but no worse than the rest of the suite of HR management tools used these days.


    • April 17, 2015 at 8:23 am

      It’s worse if all of them use the same tools and a given person is thus shut out of all jobs.

      On Fri, Apr 17, 2015 at 8:21 AM, mathbabe wrote:



  13. Zathras
    April 17, 2015 at 8:29 am

    On a separate topic, if you were to write the following article Cathy, what quotes would you include?


    • April 17, 2015 at 8:30 am

      “I am afraid of you. Everyone here is afraid of you.”

      On Fri, Apr 17, 2015 at 8:29 AM, mathbabe wrote:



  14. Moeen
    April 17, 2015 at 12:07 pm

    Have any of these companies tried giving these personality tests to their current employees to see how many of them would pass? It seems silly to use a model if it rejects people they have hired and are doing well.

    Also want to mention the link does not seem to be working for me.

    On a tangentially related note, you may be interested in this article that talks about combining data science with behavioral science to “nudge” people’s behavior:


    While the author mentions potential downsides in the conclusion, I wonder if there’s more that he is missing.


  15. The Quiet One
    April 17, 2015 at 7:20 pm

    I had to take one a few years ago for an engineering job. I felt like all the questions were in the answer yes or no, have you stopped beating your wife yet and how many times a day do you wash your hands style. That and lots of Myers-Briggs type questions. They never once asked about education or experience and I just found it exasperating.


  1. No trackbacks yet.
Comments are closed.
%d bloggers like this: