12 February 2011

Mismeasures of psychometry

My title is modified from Stephen Jay Gould's. My post is going to be a lot shorter than his book. He wrote it to debunk the intelligence test industry, and then he revised it a few years later and re-published into the midst of the Bell Curve controversy. My topic is only remotely related: I am going to write about personality tests, not intelligence tests.

It appears to me that the gold standard on the world wide web is Myers-Briggs. It has been awhile since I have read that wikipedia article on Myers-Briggs and it has grown enormously since the last time. It now is roughly one-fifth as long as the most popular book on the topic, Please Understand Me by David Keirsey. My two biggest criticisms of the test are covered in the article as it currently stands: 1.) some of the measures are unrepeatable, in that subjects will score as different particular personality types on subsequent tests; and 2.) it is completely self-reporting, with the inevitable distortion that causes. In the words of the current wikipedia article, "If respondents fear they have something to lose, they may answer as they assume they should." I have taken a number of versions of the test, variously scoring as INTP, INTJ, ISTP, ISTJ. I will give them the "I" and the "T", but my own opinion is the N-S and P-J axes are not universal, and personally meaningless.

My freshman psychology textbook is Gleitman. I have the fifth edition from 2000. Myers-Briggs is not in it. They have 75 pages devoted to the topic of personality, and several pages devoted to the topic of personality testing. The preferred instrument is the five factor scale, pioneered by W. T. Norman (Toward an adequate taxonomy of personality attributes, Journal of Abnormal and Social Psychology, V. 66, pp 574-583). Since three of their factors are explicitly social--agreeableness, conscientiousness, and (absence of) neuroticism--this test is even more subject to the distortions of incentives to erroneous or deceptive answers in the test subject's self-report. One of the most profound mysteries may be how Myers-Briggs seems to rule the web and Gleitman, et al. do not see fit to even mention it.

The issue there may be that Myers-Briggs originates in Jungian psychology, and the great doctor Carl Jung is conspicuously and totally absent from the Gleitman textbook. Carl Jung was not exactly a real scientist you see, having been seduced by the Black Tide of Occultism.

The oddest personality test I ever experienced was in a company training class on "Creative Problem Solving". This was a one-week class where I and twenty of my peers were exposed to such company-sanctioned activities as "brainstorming". Part of the class was a personality assessment test. The testers claimed they could sort for: right-brain versus left-brain, and top-brain (cortex) versus bottom-brain (limbic). They had a composite display of the whole class on one plot. I suffered the ghastly embarrassment of the teacher telling all my classmates that I was the only right-brained person in the room. The other thing I remember from that test is my perfect career match was supposedly physician or metal sculptor. There is absolutely no way that a statistically significant number of well-adjusted metal sculptors have ever taken their test.

My favorite personality tests are the ones on the OKCupid web site. My results indicated such arcane features of my testing self as:

I am an INTP (d'oh!)
I am an Enneagram type 5.
I am partial to Cognitive Behavior Therapy.
My philosophy is akin to William James.
The Looney Tune character I most resemble is Wile E. Coyote.
The Romantic poet I most resemble is William Blake.
The major arcana I most resemble is the Hermit.
The Shakespeare character I most resemble is Richard III.

OK I am going to stop with that one because it is way off. Prospero, maybe; no way Richard III. Murdering my two young nephews to legitimize my kingly succession is just not me. As much fun as the OKCupid tests were, I only took about twenty of them. I know one woman (she is a World of Warcraft player) who has taken like three hundred OKCupid tests. Alright, one last one:

The Arthurian character I most resemble is the Lady of the Lake. This dear readers is scientific proof that there are worse personality testing instruments than Myers-Briggs.

No comments:


Please see paragraphs 8.4, 8.5 in the Google Terms of Service document!

About Craig

My photo
Houston, Texas, United States
I have been living in the lovely neighborhood of Spring Branch in the great city of Houston since late in 2005. I started out with the idea of making this blog about my life in this neighborhood. That did not last long. Right now I am posting every five days on the alternating topics of literature, philosophy, psychology, and metaphysics. This project has been ongoing since July 27, 2010 and I believe it will continue for at least a few more months.