Thursday, October 22, 2009

Woods v Texas (2009): Another Atkins MR death penalty decision psychometric quagmire


Woods v Texas (2009) is the most recent (October 7, 2009) Atkins mental retardation death penalty decision I've come across (now added to Court Decisions listings).  Woods denial of classification of mental retardation was previously decided in 2008 (Woods v Texas, 2008).  And, per usual with these decisions, the decision contains a quagmire of psychometric issues.  Although I've stated previously that I'm no longer shocked at some of the psychometric testimony and the courts interpretations and decisions (sometimes which border of voodoo psychometric logic), I guess I still am....but more a form of numbness to some of the material I have read.  Woods v Texas (2009) does not disappoint with a boat-load of psychometric issues that need further exploration.

As I've stated previously, I simply can not keep up with these cases.  Each requires considerable study and reflective thought...and I currently lack the time to do each justice.  Rather than sit on these decisions until I have the time, I'm posting them so others, who may have the time or motivation, may dissect them and possibly offer comments via the comments feature of this blog or, and my preference, to submit a guest blog post for publication at this site.

The above being said, I've done a very quick skim of the decision and see some of the "old familiar" measurement issues being present, and some "newbies".  Below is a sampling of issues that need to be addressed in sfuture post:

  • Flynn Effect
  • Standard error of measurement (SEM)
  • Diagnosis of MR before the age of 18.  Especially in the presence of IQ testing during the school years that produced IQ scores of 78 and 81, yet Wood's was diagnosed as learning disabled (LD) and not MR.  I've yet to delve into the whole issue of the accuracy of school-based classifications.  Having been a school psychologist for 10+ years, during the time when IQ tests where under serious fire for MR classification, esp. for minority students (the famous  Larry P. decision that sent shock waves through the psyche of many school psychologists when it came to classifying any minority student as MR even when faced with IQ test scores that were in that range), I know that many students were given the more politically easy Dx of LD to avoid the potential fallout of  labeling someone (esp. a minority student) MR.  It was simply the "easy way out"....yet now I wonder if these decisions are clouding Atkins rulings related to the third prong of the definition of MR (Dx prior to age 18).  This is another whole post or series of posts.
  • Adaptive behavior (AB) definition.   I see in the decision, that in the same breath, both adaptive behavior and maladative behavior, which are separate constructs, are discussed as they exist under the single umbrella of adaptive behavior.  This relates to the discussion of the Scales of Independent Behavior--R (SIB-R) adaptive and maladaptive index scores.  Yet another long post or series of posts.  Toss in Texas' unique Briseno twist on adaptive behavior and there is much to dissect.
  • Adaptive behavior (AB) assessment.  Aside from using an appropriate nationally normed measure of AB (SIB-R), I'm dumbfounded by the use of the SSSQ as a measure of AB, given that the SSSQ was normed on a sample of individuals with disabilities.  Woods apparently scored 95, which was in the normal functioning range when compared to other individuals with disabilities.  This would be analagous to developing WAIS-III/IV IQ norms based only on a sample of individuals with mental retardation, and then giving the test to a person and saying they are average or normal.  It would only tell you that when compared to other individuals with MR, they are better functioning than most so classified people.  Standardized psychometric instruments (IQ or AB) used in Atkins decisions  should only be based on nationally representative samples characteristic of the populatlion of the US....not a select subsample of individuals with disabilities.  This makes no sense. 
  • Competency to stand trial.  Then there is the use of an instrument I've not yet familarized myself with..the Competency Assessment to Stand Trial For Defendants with Mental Retardation (CAST-MR).  This may be an excellent psychometric tool....I need to examine it and determine how it fits in the whole Atkins three-prong defintion of MR for death penalty cases.  It has been my understanding (which I need to verify) that competency to stand trial is a separate issue from determination of mental retardation.  Again....I suspend judgment and comments until I can get up to speed on this instrument.  That being said, a skim of the decision suggests the scores were used when debating whether Woods was MR...hmmmm.....doesn't seem appropriate.
  • Short form IQ tests.  Do short form screening IQ tests (e.g., WASI) have any place in Atkins MR determination?
More than enough to chew on. 

Technorati Tags: , , , , , , , , , , , , , , , , , , , , , , , , , ,


1 comment:

  1. I agree that the CAST-MR has little to no place in a discussion of whether or not someone has MR. To the extent that the MR diagnosis is not entirely psychometric but also has room for clinical judgment, extreme performances on the CAST-MR and qualitative aspects of responses to the structured interview portion might make some contribution to the diagnosis.
    The SSSQ is a different matter. I don't know the test, but it is distinctive in this context in that it purports to be a performance measure of AB rather than a rating scale, which would seem to make it less susceptible to response bias (although potentially still susceptible to malingering). The website says there are both normal and disabled norm groups, so it is possible that it is psychometrically appropriate. As a performance test, it may also have the possibility of helping to get around the problem that prison requires few ABs. Of course, then we'd be back to the issue that performance measures look at best performance, which AB measures look at typical performance, so would a performance AB measure be measuring best or typical behavior? I'd appreciate someone taking a closer look at this instrument for those distinctive features. How well does it correspond to the 10 defining domains of AB? Any other performance measures of AB out there?
    Tedd Judd

    ReplyDelete