Tuesday, March 31, 2009

More on Ozonoff et al. (2005)

In a previous post, I spoke about a 2005 study on the MMPI profiles of "high-functioning" autistics. In that post, I mentioned that there were aspects of the study that I couldn't comment on too much as I lacked critical knowledge.

One of the advantages of being a grad student, however, is access to university resources. At Nova, this includes an entire (small) library dedicated to psychological testing. Given that the MMPI is the most commonly used psychological test -- period -- the fact that it included a good bit of relevant documentation should be no surprise.

It took me less than five minutes to find information on the PSY-5 scales. It took me somewhat longer to photocopy the relevant pages, but a good bit of that time was spent looking for a copy machine.

The RC scales, on the other hand, weren't covered in the books that I looked through. Fortunately, however, another benefit of being a grad student is access to the faculty. I plan to ruthlessly exploit that to gain relevant knowledge.

In any event, I forgot to mention a few things in my last commentary, namely the validity scales. The autistic group in Ozonoff et al.'s study scored higher on the L, F, Fb, and Fp scales overall.

Those scales are primarily intended to detect attempts to misrepresent yourself by appearing as other than you are. While Ozonoff et al. do not discuss any of these results other than the L scale, they attribute that effect to "limitations in insight and self- and other-awareness". I believe that this explanation is somewhat contrived.

Given that this study was authored in part by the personnel responsible for diagnosing the persons involved, given that this study was authored by at least in part by personnel at the clinic that the subject pool was recruited from, given that the authors take a psychopathological approach to autism treatment, given that the test was presumably administered by the personnel in question, and given that the treatment methods based on that approach usually involve what amounts to training the autistic individual to pretend that they are neurotypical, I suspect that the real explanation is just a bit simpler.

On the PSY-5 scales, which are based around something akin to the Big 5 Personality Factors, the autistic group typically scored lower on the aggression and disconstraint scales. The aggressiveness scale is exactly what it sounds like... which is interesting, given the stereotype of autistic children as aggressive.

The disconstraint scale, on the other hand, is a measure of risk-taking, impulsivity, untraditionality, and lack of preference for routine... among other things If this sounds a bit backward, it is. Most of the PSY-5 scales are defined as the opposite of Big 5 personality traits.

Finally, the autistic sample scored higher on the introversion scale. This... shouldn't be a surprise given that a low score typically means that the person is highly social in nature. While I think that this is a real effect, I have to wonder about the possibility of they study's use of psychology students as a sample contaminating this result.

And that is all I can add until I get around to talking to someone who can help me interpret the RC scales.

1 comment:

  1. I've posted an additional follow-up at http://aspieperspective.blogspot.com/2009/04/still-more-on-ozanoff-et-al-2005.html .