Implications of Profiling for Asperger's Assessment

Post Reply New Topic

Hi guys!

I've been interested in the area of forensic psychology. More specifically lie detection.

Studies have shown that with the use of a criterion based content analysis (an analysis made up of different criterion which has been shown to correlate with truth telling) it significantly increases the likelihood of detecting a liar to 68%. Whilst this is all well and good, it still means that 32% or almost 1 in 3 cases will result in a false positive or true negative.

If this use of profiling is resulting in error in 1 out of every 3 cases, then who is to say that the 'profiling' techniques used by psychologists and psychiatrists are significantly more successful?

Further studies have shown that when deceivers were informed of the criterion they were able to score just as high as truth tellers even with the assessment conducted by a criterion based content analysis expert and the use of a discriminate analysis.

We know that a lot of people who are going to psychologists or psychiatrists looking for confirmation of their self-diagnosis having done research. What does this mean for them?

no one has anything to say.
is it because i go on about lie detection assessments too much?

Interesting post, and notion. So are you suggesting that those who self diagnose may then be influencing the official diagnosis by basically taking on the characteristics they have researched? If so, that sounds like a placebo effect, and I think it probably would or could result in a false positive diagnosis for some.

If you mean what the poster above meant than yes, I can see how it is possible. A person who does that though would probably have SOME kind of disorder in the first place which makes them want to do that. Maybe a personality disorder of some kind. I don't know how accurate a lie detector would be on me because I have Tourettes so I don't know if my tics would interfere with the test! I would love to try one and answer some really obviously fake questions while in full tic!

It means something if 100% of self-dx'ed people are deliberate and conscious liars. If 0% are liars then it means nothing. The truth is likely somewhere in between.

It also means that there are a lot of spoiled, over-privileged diagnosed people, if that's what they're worried about.

Lie detection seems a different process. A liar is actively trying to hide the truth. While some being evaluated might lie about things, it seems more likely that someone going to a therapist wants an accurate diagnosis and won't be motivated to lie.

Ahhh yes. I think some of you are kind of missing my main point.

Because of the abstract nature of disorders such as Asperger Syndrome and other neuro-psychiatric disorders, a DSM criteria has been made. Psychologists and psychiatrists see if an individual fits this criteria. This fits into the definition of profiling.

Profiling used in other settings such as lie detection has been shown to be only effective 68% (which forensic psychology reports to be significantly high). However it also means that 32% or almost 1 in every 3 will result in a false positive or true negative (otherwise known as error). Given that we know that error in lie detection profiling occurs in 1 out of every 3 cases, who is to say that profiling (or an evaluation) performed by a psychiatrist is significantly more effective in terms of error.

What criteria are used for lie detection? Aren't those criteria more subtly expressed than the ones used for ASD diagnosis?

criteria based content analysis (CBCA) assess lying purely by content of what is said.

some of the criteria associated with truth-telling include:
- spontaneous correction
- reproductions of conversation and descriptions of interaction
- unstructured production
- detailed characteristics
- acknowledgement of subjective mental state and admitting lack of memory etc.

in 92% of experimental studies, truth tellers received a significantly higher CBCA score than that of liars.
however, when it came to profiling only 68% of cases where able to differentiate a liar from a truth teller.

the rate for verbal and nonverbal assessments were noted to be worse.
CBCA assessments are admissible in court proceedings in some areas in America and Europe and is known to be the gold standard in lie detection.

I don't think I'm missing the point.

Take something like an I.Q. test. These are relatively stable over time and from one practitioner to another. They are reproducible to the extent that they can be considered objective. Regardless of whether or not the measure intelligence, they measure something with far greater reliability than a 30+% error rate.

Or hallucinations in schizophrenia. Hallucinations are easy to identify, When someone says there are spiders crawling on my arm and there aren't any, it's a hallucination.

What you are suggesting seems highly dependent on the type of evaluation. It would not surprise me that some aspects of psychological evaluations are just as error prone as lie detection. But I think you can't generalize to the entire field.

to be fair IQ isn't really profiling, no psychologist is given a list and said compare people to this criteria

the criteria for asperger's is one which is not that black and white.

i forgot the whole criteria but here are some i do remember:
- failure to develop friendships to developmental level
- lack of social and emotional recipocy
- lack of appropriate use of gestures and non-verbal ability (or something)
- pre-occupation with particular object

are they really less subjective than:
- spontaneous correction
- reproductions of conversation and descriptions of interaction
- unstructured production
- detailed characteristics
- acknowledgement of subjective mental state and admitting lack of memory etc.

???

So then nobody is autistic because the evaluations are too subjective?

I still think you are comparing apples to oranges. You are taking a specific methodology for determining lies and generalizing its flaws to the diagnostic process for autism. But a professional evaluation for autism isn't even a single test but rather a series of interviews, tests and historical information.

Can subjectivity enter the picture and result in a faulty diagnosis? Sure. Especially in the hands of an unskilled practitioner.

Let's change the question. Would an expert in detecting lies be able to improve beyond the 30% error rate or is this error rate intrinsic to the technique? Would an expert in autism have a lower rate of error?

when did i ever say that nobody is autistic - no need to get defensive.

68% was that of experts for lie detection in meta-analysis by vrij. if it was done by an unskilled person or other professionals such as police officers and psychologists without extensive experience, they performed at chance rate i.e. 50%.

for more information: you may want to check -
vrij, A. (2008). detecting lies and deceit: pitfalls and oppurtunities. chichester: wiley.

I was being facetious, not defensive. Basically, profiling as you define it is invalid. So if the only way to evaluate autism is invalid then autism as a diagnosis is invalid. Hence autism does not exist. This is rhetorical hyperbole.



This just sidesteps the point. You are focusing on a specific methodology and generalizing it. But the diagnosis of autism does not rely on any single test or technique.

no i did not define profiling as invalid. i said only 68% were able to correctly catch a liar using CBCA. 92% of CBCA studies however concluded that CBCA scores were significantly higher in truth tellers than in liars. there needs to be a distinction between investigation of the construct and using this knowledge to 'profile' certain individuals.



The process of CBCA:
- interview
- evaluation of competency
- evaluation of statement quality
- evaluation of statement reliability
- validity checklist
- evaluation of motivation related content

I guess I have no idea what your getting at.