Wednesday, July 29, 2009

Cthulhu approaching on his Harley

Well, that's what I think it looks like, anyway. Your mileage is by definition invited to vary -- the image is Plate 4 of the Rorschach Inkblot test.

A Wikipedia user -- Dr. James Heilman, of Moose Jaw, Saskatchewan -- posted all ten of the Rorschach plates on the site last month, and now all hell is breaking loose in the "psychology community."

Apart from the obvious financial implications -- one company has been selling the plates to practitioners for the better part of a century and they currently go for $110 and up; now those practitioners can just crank up the office printer on some good card stock instead -- some psychologists complain that the publication of the plates "jeopardiz[es] one of the oldest continuously used psychological assessment tests." Those who expect to be assessed using Rorschach, they say, can now know what's coming and game the system to produce the results they choose, rather than giving the assessor real insights into their minds.

Color me skeptical of psychologists' grounds for complaint. This particular kind of "assessment" strikes me as very similar to palmistry, astrology or "cold reading" parlor trickery in the first place. Yes, information can be extracted from the responses, but that information really comes down to nothing more than how the patient's responses plot against a bell curve of a data set of prior responses.

Sure, the psychoanalyst can figure out whether or not the patient is "normal" (i.e. tends to respond toward the center, rather than one or more standard deviations out toward the edges, of the bell curve) in some sense. But when it comes to what "abnormal" responses mean ... well ... I've never heard hard, testable claims, not even correlative ones like "some significant X% of those who answer Y to Inkblot 4 are subsequently found to be ax murderers," or "some significant A% of those who answered B to Inkblot 9 were subsequently diagnosed with tumors in their cerebral cortexes." That would at least be of value to, say, epidemiologists.

It's a set of blots, guys. That's all it is -- and the responses to those blots aren't meaningful objective pieces of information, they're just starting points for you to riff on. After nearly a century of use, any "trade secret" shiny has long since worn off. Get over it and get a new prop if you think this one is compromised.

[Note: I picked up on this topic at Techmeme.]

No comments: