What does Alice in Wonderland have to do with psychological testing?

Dear Cecil: I have never read the book Alice in Wonderland and am afraid to do so. I have read that during psychological tests for law enforcement agencies a usual question is whether you have read/like this book. In Christopher Whitcomb’s book Cold Zero about the FBI’s Hostage Rescue Team he mentions that one of the questions in a test was if he had read Alice in Wonderland. I had a friend who interviewed for the Denver Police Department and they asked if he liked Alice in Wonderland. Have you ever heard this? If you have read the book what does that say about your character? thewazir, via e-mail

Cecil replies:

Let’s clear up one thing off the bat: the question isn’t whether you’ve read Alice in Wonderland, it’s whether you liked it. (To be precise, you’re supposed to respond true or false to some variant of the statement “I enjoyed reading Alice in Wonderland.”) This slightly sinister query was in the original edition of the Minnesota Multiphasic Personality Inventory (MMPI), a well-known personality test commonly, if inappropriately, administered to job applicants. It was dropped from a newer version of the test, MMPI-2, which was published in 1989, but still appears in another test, the California Psychological Inventory (CPI). Your paranoia about such things is understandable – what’s a question like that supposed to gauge? Whether you like fantasy? Hookah smoking? Little girls?

I’ve gotten several answers. William Poundstone in Bigger Secrets (1986) writes that, on the MMPI, “Liking the story suggests femininity in a man; disliking it suggests masculinity in a woman.” However, retired University of Minnesota psychology professor James Butcher, one of the prime movers behind MMPI-2 and somebody who ought to know, tells me the Alice question was experimental and never measured anything. Possibly Poundstone got the MMPI and the CPI mixed up – a spokesman for the company publishing the latter says the answer among other things contributes to the “femininity/masculinity” scale, which measures “sensitivity vs. action.” An element of sensitivity is an interest in the arts, presumably including literature, hence Alice. Does answering “true” make you too sensitive to be a cop? Not to worry. Dozens of items contribute to the F/M scale, giving you ample opportunity to demonstrate you’re butch.

Strange as some of the questions sound, tests like the MMPI and the CPI have long been considered more scientific than other psychological evaluation tools. In creating the MMPI, the older of the two tests, psychologist Starke Hathaway and psychiatrist J.C. McKinley purposely shunned theory in favor of an empirical approach: They administered sample questions to groups of people with identified psychological disorders and compared their responses to those of a control group. Questions that reliably distinguished different types of pathological respondents from “normals” and each other were included in the test.

The CPI, developed by Harrison Gough, used a similar method, except that the idea was to assess personality traits such as dominance, sociability, etc., among normal individuals. Test responses in the benchmark group were correlated with personal assessments by trained observers, family, coworkers, and so on. Gough had gotten his doctorate in 1949 from the University of Minnesota, where the MMPI was developed, and he and the MMPI guys were drawing from the same pool of sample questions, so that’s why some items appear on both tests. Which test did the people you mention see the Alice question on? Hard to say – the original MMPI was available until 1999.

The thinking behind the MMPI and CPI sounds reasonable, but how well the tests work in practice is a matter of debate. The original MMPI was especially criticized in part because of the large number of culture-bound, intrusive, or just plain goofy test items. Some statements drew widely varying responses having nothing to do with personality – for example, research showed that, in Alabama, when confronted with “I like Lincoln better than Washington,” black people went with Abe while whites preferred George. Some items were so dated as to be incomprehensible – “I used to like drop-the-handkerchief” referred to a kid’s game no one’s played since bobby socks went out. Another problem with the MMPI was the inadequate statistical basis for the scoring. For example, the calibration for whether you had homosexual tendencies – this was in the day when homosexuality was considered a sickness – was based on a sampling of 13 gay men. The normals against which the supposedly psychopathic cases were contrasted were 724 “Minnesota farmers,” minimally educated rural types who were hardly representative of the general population.

MMPI-2 supposedly rectifies such defects, eliminating the more dubious items and providing more reliable norming. The CPI has been revised a couple times as well. But plenty of seemingly weird stuff remains on both tests. What to do if a job’s on the line? The sensible advice is to answer honestly. The tests contain items intended to detect if you’re faking it, and while these aren’t all that difficult to suss out (HINT: Nobody always reads the editorials in the newspaper, never tells white lies, etc.), it’s probably not worth trying to out-psych the psychs. That said, the results of even the most scientific personality test are subject to interpretation, and who knows what the HR geeks at your local police department are looking for? Ain’t saying I recommend this, but the cynical advice would be to put yourself in the mind-set of a small-town Minnesotan with an eighth-grade education circa 1942.

Cecil Adams

Send questions to Cecil via cecil@straightdope.com.