The first phase of our interview process is a Gallup-style assessment that our HR director administers to all candidates. I listen to the interviews because I'm very interested in seeing how a candidate processes a series of open ended questions. Based on the results of the assessment, we move on to the next phase, the technical phone interview, where I weed out all the CS majors who can spend hours discussing cardinality but can't actually design a decent normalized schema.
Apopos of this story, I have my HR director remove educational background from the applications that I see; I have a heavy bias regarding education and I want to be as objective as possible. Oh, and the test is not a personality test of any type.
Anyway, this candidate, we'll call him Friedrich, failed the Gallup test resoundingly due to his desire to turn every question into a philosophical quagmire. He seemed far more interested in intellectual wanking than anything else, and I don't have time for that shit. So my HR director sent him the standard, "thank you for your time, we've decided to continue our search blah blah blah."
I get into the office this morning to find the following diatribe:
----------------------------------------------------------------------------------------------------
If you really believe that the test you gave me is in any way valid, you are not hiring well. Though I am rather disappointed in you for even considering it an important tool, I only say that to point out that it cannot give objective data.
"The problem is that Myers and her mother [who devised the test in the 1930's] did not really understand Jung at all. [Jung wrote on personality type frameworks as as grounding for functioning] Jung didn't believe that types were easily identifiable," says Malcolm Gladwell in the New Yorker <http://www.gladwell.com/pdf/personality.pdf>. That quote is a nice statement. But the real criticisms of Myers-Briggs Type Indicators
(MBTI) and other MBTI-style tests are more difficult to undrestand.
MBTI tests make four dichotomies, or personality preferences (EI, SN, TF, JP). These distinctions are bifurcated - you are either in one aspect of the type distinctions or the other (e.g., Extraversion or Introversion). What tests like MBTI try to show is the combination of categories that a person elects for themselves. Then MBTI names those results as if they were the real-world types.
That is, MBTI tries to make *real* or objective claims. However - despite the pragmatic intent of Myers-Briggs' original research - she and her mother actually came up with a system of logic. A system of logic cannot actually *say* anything about reality. (Similarly, in statistics, a group *cannot* say anything about any specific member of a group.) Logic systems can only say, "IF condition(s) 'P' actually exists in reality, and all our axioms are correct, then we can reasonably predict that conditions 'Q' will follow." Though this may
*seem* like it is predicting reality, it is not - it is only applying a set of rules onto a set of data; the set of rules *arbitrarily* define the meaning of the terms (the terms are "empty" linguistically). It's a difficult concept to grasp - I know this from seeing most people getting the idea wrong - but it is critical to understanding what use logic systems actually do play in reality.
Before any logic system can be used at all, the system must be sound (technically, it needs to be uniquely readable, expressively adequate, etc.). MBTI is not. For instance, it requires yes-no answers for the ability to generate a preferred type profile. If *any* question is unanswered, the results will be wrong. If any question is not understood, the result will be wrong. If the question is answered in two ways, the question cannot be answered, negating one of the results, and the results will be wrong. If a person scoring the exam doesn't understand the answer given and marks the wrong answer, the results are wrong. And so on.
What you are guaranteed with MBTI-style exams is one of two things:
1) Certainty, if you don't understand that MBTI cannot have certainty.
2) No useful information, if you understand what MBTI can and cannot show.
Now, I don't have any idea of what answer you put down for the questions where I answered in two ways, nor where I did not answer, but I can guarantee that your test results do not contain the fact that I am a good programmer; nor does it indicate the idea that I would greatly benefit <your company>.
You may think an ability to control the results indicates an essential dishonesty in the testee. This is not true. The MBTI-style exams are meant to show what a person *prefers* as a personality.
For this exam, I preferred to answer in a way that would exhibit creativity, an interest in learning, careful thinking, careful planning, and dedication to a job. These are traits I saw as necessary from the job description. If the tests results don't show that, it only means that I answered some of the questions incorrectly, but not dishonestly. You have to remember what I mentioned above - that the MBTI test questions are empty linguistically.
I have no idea how you could think I am not the "right" candidate, but I hope you reconsider. [emphasis mine]
----------------------------------------------------------------------------------------------------
It was no surprise to learn that Friedrich has been working to finish his undergraduate degree in philosophy since 1993.