posted 03-22-2006 01:26 PM            

------------------
but then, that's just one man's opinion

posted 03-23-2006 03:16 PM            

Just think of your college math professor.

It may be important to see a multiaxial diagnosis - including the GAF (scores under 55 are typically found amoung hospitalized persons)

Like autisitics, persons with asperger's can be either hypersensitive or hyposentive to various types of stimuli, and the constellation of features is known to be somewhat individualized and ideosyncratic.

Both autistics and asperger's patients have substantial deficits in reading non-verbal interactive social messages. They laugh too long, speak too loud, have perseverative and obsessive interests, and are variously reclusive or intrusive socially.

I'm not at all sure they represent normal polygraph subjects and there is a substantially increased likelihood that a subject will react or not react to test stimuli for reasons that we don't fully understand.

If someone is suggesting your client is aspergers but not properly diagnosed, then that should be sorted out first.

You would also want to know about any medications, as some persons with developemental disabilities exhibit an increased prevalence of mental health, compulsive, and psychotic symptomology.

It seems to me that our empirical confidence in the meaning of polygraph test results depends upon our understanding of normal physiological and psychological processes. Persons with developmental disabilities are, by definition, neurologically compromised. In measurment science we say "they are outside normal limits," along whatever measured (statistically normed) functional parameters.

Aside from that they're fun folks

r

------------------
"Gentlemen, you can't fight in here, this is the war room."
--(Dr. Strangelove, 1964)

posted 03-23-2006 06:23 PM            

No, aspergers is not the same as autism, but they sometimes appear to overlap (so to speak), which just adds that much more difficulty to the mix.

If the guy works and presents as "normal," and he hasn't been officially diagnosed, then see how he presents, and try from there. Just be careful in the pre-test, making sure you are communicating fully. Keep in mind you won't be able to "read" the person well at all, so you'll have to really listen to what he's saying.

posted 03-23-2006 10:15 PM            

It appears we'd be getting away from the sound practice of not testing if there are any significant medical/psychological issues. No doubt we all understand that simply because a person hasn't been diagnosed does not mean their problem(s) do not exist.

END.....

posted 03-24-2006 06:57 AM            

You know the whole point of the pre-test is to set up the test so the truthful can pass. There's nothing we need to do to set up the test for the liars to fail. One can therefore be much more confident of an NDI conclusion. That is one of the reasons we have asymetrical cut-offs in our scoring (DoDPI / Utah - not Backster). We make it harder for the truthful to pass, which makes it harder for the liar to slip through the cracks. Why? The cost of a false negative error is too high in most of our tests. So, in the end, an NDI results in greater confidence in the opinion than a DI would, generally speaking.

I tested a pregnant person (seven months or so) once (pre-employment), and i told the PD up front my only calls would be INC or NDI. I wasn't willing to make a DI call because I couldn't defend it with the unknown variable involved. (How could I say the baby didn't effect something? How could the baby help her pass though? It couldn't, which is why I held that position - and still do.) They understood, and I tested her. (She had a doctor's note and signed a liability waiver.) BTW, the test went fine. The baby moved once the whole time. (Maybe I was boring?)

From a scientific standpoint, one is not more "valid." That is a term we, the uninformed, use and give it a new definition. That's another reason the antis loathe us. One test, e.g., the VSA, can be "valid" but only 36% accurate. Polygraph is valid, but accuracy rates for DI and NDI are different. The NDI accuracy rates are scary at times. Because probabilities are different for both, then confidence is different too.

posted 03-24-2006 08:19 AM            

The logic of differential consideration for various types of test results is embedded throughout testing sciences, and is fundamental to the concepts of screening and diagnostic testing objectives. All tests (polygraph and otherwise) come in these two basic flavors, and the meaning of the test results is in part circumscribed by the testing objectives at the onset of testing. Read Kraphol's article from 2003 for a good intro.

So, if the objective of testing is to investigate and provide an empirically based professional opinion regarding the presence or absence of bananas in a tree (are there no bananas in the tree, or the null version: are there any, at least one, bananas in the tree), then it is not satisfactory to give a personal opinion regarding that - even if it is a banana tree (it might be a very old and barren banana tree). A thorough investigator would go to the tree and look. Are there any bananas under the tree? If there are, then it becomes a more sound empirical assumption that there are bananas in the tree (assuming the laws of gravity have not been violated, and assuming that no-one is engaging in any clandestine banana dumping underneath a barren barren tree). A more thorough investigation might involve shaking the tree (and I mean shake it rigorously, not just call in names and tease it), to see if any bananas fall out. That is essentially what some test do - stimulate or illuminate an issue to reveal its presence. Some tests are invasive - like the Heida scan which is uses radioactive tracers to investigate or test for gallbladder dysfunction. A more invasive test or investigation of our banana tree might involve climbing the banana tree for closer inspection (by hand, ladder, or high tech rock climbing gear if you love gadgetry) of any evidence of bananas in the tree.

In polygraph this stimulating, illuminating, or shaking is analogous to provoking (fun choice of words) a test subject's physiological response potential with a polygraph stimulus question. (Did you take the bananas?) One requirement is that the stimulus question be sufficiently robust that it would reliably stimulate a reaction within a subject who was involved in the issue of concern. So that if we attempt to provoke a reaction with a robust polygraph stimulus question and observe no measurably significant reactions (statistically significant non-reaction to the stimulus - a bit of a chunky concept here), then (based on the robustness of the stimulus, and normally anticipated responses from a normal functioning subject) we can begin to formulate empirically based opinions that yes, we have no bananas, or the subject has no response potential to the issue under investigation because he was not involved in taking the bananas. We went looking for something and came up empty handed (negative result). The validity of negative results is in part based upon the rigorousness and thoroughness of our efforts to find something (shaking and climbing the tree). So, investigation strategies based on don't-ask-don't-tell methods are obviously insufficient.

You can extend this logical syllogism to conditioned response theory, complexity theory, fear of consequences, guilty knowledge (or whatever your psychophysiological explanation du-jour). Just keep in mind that we loose empirical confidence whenever the empirical rationale underlying the investigation is not supported by good information from the sister sciences of psychology, physiology, predictive statistics, and test theory.

So, under normal testing situations, it is a reasonable to offer an empirically based professional opinion that yes we have no bananas, as distinct from a personal opinion, that a certain banana tree is barren or fruitful (excluding the possibility that some zealous banana thief has done very tidy work). What we want is to be able to offer a professional opinion that we searched as rigorously as we could and came up empty-handed, and that our empty-handedness is unlikely to be due to chance or accident.

I don't completely agree with the idea that negative results (yes, we have no bananas) are more valid with developmentally disabled persons than positive results.

It is important to understand that human psychology and physiology is quite complex and somewhat variable, and that the polygraph depends upon the normal functioning of psychological and physiological mechanisms.

To push the banana metaphor further, climbing the tree (or chopping it down) might be the most robust form of investigation as it allows the observation of physical evidence. In polygraph we most often investigate issues for which there is no physical evidence. So, hypothetically, if there were some religious or moral taboo against climbing banana trees, and all we could do was shake the tree (imagine a very tall tree in which the presence of bananas were obscured by foliage) to stimulate evidence of the presence of bananas.

Consider that persons with Autism or Asperger's disorders (or other developmental, or neurologically based abnormality) may be analogous to a banana tree with some genetic abnormality, causing it to produce unusually small and lightweight bananas (like the mentally retarded person's limited conceptual knowledge and memory recall skills) or unusually thick banana stems (like the developmentally disabled person's cognitive rigidity in the prefrontal cortex), causing bananas to remain in the tree even when vigorously shaken. Negative test results have now lost some empirical meaning.

In another set of problems. I recently reviewed a test in which an examinee was found non-deceptive, in which the test questions pertained to the behavior of the alleged crime victim ("did she take the bananas???," "did she give you the bananas???," or "does she have the bananas???"). I've seen a few of these - usually done at the request of some defense attorney - that have turned out to be false negative tests (as confirmed by the subject's later confession), and I'm now highly suspicious of this type of question. I know of no sound theoretical rationale why I could attach my polygraph junk to one person's body to investigate another person's behavior. Has anyone ever seen a telekinetic polygraph? (Maybe those voice stress guys, but that is a whole 'nother rant.) This would be like shaking or climbing some neighboring tree, to determine the presence or absence of bananas in particular tree under investigation - even if the neighboring tree were a banana tree.

The problem is that the absence of reaction becomes empirically meaningless as there is no sound empirical explanation why someone would react to stimulus questions about another person's behavior - and if there were such explanation, then we better be worried... because everyone who failed polygraph could proffer the excuse that someone else took the bananas.

Now that the banana tree metaphor is exhausted, however there are also some important problems when interpreting the empirical meaning of the presence of polygraph reactions when we deal with persons known to be neurologically compromised...

but more on that later,

'cause I've got to get some work done.

Peace,

r

------------------
"Gentlemen, you can't fight in here, this is the war room."

posted 03-24-2006 01:25 PM            

#2. RMELSON All this talk about bananas was very illustrative. I really mean that.
If the bananas ever existed and if we identify the culprit, may I suggest we make Daiquiris with the proceeds of the crime.

posted 03-24-2006 05:23 PM            

Thanks for the assertive and personal position on that. I think its always worth listening to the most conservative voice in the room, whenever we are goofing around with other people's lives.

I have a nephew who is autistic, and I still have teethmarks on my arms from a college job working with autistic adults (most of whom had previously lived in state hospitals for 15 to 20 years until a massive deinstitutionalization effort during the 1980s).

I like Barry's conservative approach to the exceptional testing scenario - in which the absence of reactions would can be regarded as not due to chance or accident (extraneous factors), while any unexplained or inconsistent reactions would be interpreted more cautiously due to the difficulty of prevailing in a arugment about the causes of those reactions.

While my earlier post attempted to map out why the absence of reactions might offer reduced empirical validity, it was also pointed out earlier that there are also viable reasons why the presence of signifant reactions may have complex or extraneous causes with developmentally disabled persons who are hyperreactive to some forms of stimuli, while underreactive to others.

When dealing with severely behaviorally disordered autistics it is very important to train staff members never to hold or twist a joint, as they sometimes dissociate pain and will not stop aggressing even after a limb or joint is broken or dislocated. You just can't arm-bar or figure-four lock them the way you can a normal person.

Some of the really positive effects of deinstitutiionalization efforts are that we observe far less institutionally driven behavioral violence (including fewer self-injurious behavior patterns) amoung these special populations than we have in the past. I haven't seen a young developmentally disabled person whose teeth were pulled due to biting behaviors since the late 1980s (sadly, this was common at one time, as was the use of aversive behavioral interventions like spraying amonia in the face). Plus, we now know that with early intervention developmentally disabled persons can learn far more and live much more normal lives than we had previously thought possible.

So, change is a good thing. Understanding, patience and tollerance for exceptional persons is also a good thing.

Now for a less conservative argument (even if only for the simply joy of argument).

POLYBOB

All else aside, I think that you may want to evaluate each client individually, with consideration for the goals of testing. I imagine there are some exceptional folks for whom the polygraph might work just fine, depending on the severity and complexity of the disorder.

If you or someone else determines his disorder does not preclude his suitability for polygraph testing, I'd suggest you consider advising people that he might be considered at best marginally suitable for the polygraph technique, (similar to the way that Barry discribed his caution around the suitability of his pregnant subject), and I'd probably be unwilling to give anything other than a qualified or cautious opinion regarding any test results. The test results would probably mean what they appear to mean, but I wouldn't want to have to argue or fight to the gravel (or gavel) regarding the empirical validity of the results. Despite that limitation, the information gained might be very useful, and in some situations (PCSOT) there may still be some deterent value to polygraph monitoring. As Gail Ryan (who seems mostly opposed to polygraph) once pointed out, it may be that conducting an intrusive test is less uncomfortable and less unsettling in the long-run than weeks or months of belabored discussion and questioning.

However, BILL2E is absolutely right that some people are simply not good subjects.

...just a few more .02 worth of opinion...

and daquiris sound great, but I've had enough bananas for one day.

r

------------------
"Gentlemen, you can't fight in here, this is the war room."
--(Dr. Strangelove, 1964)

posted 03-25-2006 07:15 AM            

Again, part of the foundation of our procedure is that we do not administer an examination if there are any significant medical/psychological issues.

In my opinion, diagnosed or not, an examination should not be administered.

posted 03-25-2006 03:12 PM            

Hypotheticals and things are fun for argument and learning (and I tried to softened the technicaleze with the bananas), but in the end someone has to make a decision.

The decision to test comes down to a dichotomous choice - to test or not to test; that is the question. Whether it is nobler science to suffer the slings and arrows of our colleagues when we test a subject whom others might not, or to take up arms against a sea of second opinions and by opposing educate ourselves.

OK, Shakespeare I am not, and I can't do iambic pentameter.

As this actual case example illustrates, the dichotomous or black and white choice to test or not test is sometimes preceded by the need to navigate some very grey space first.

Hopefully, these discussions give us more information with which to make these decisions.

The real question is whether the polygraph technique will work with this client. This question is compounded by the goals of the referring agent, what types of action will or will not result from a positive or negative outcome, and whether we can assure reasonable empirical validity of the test outcome.

It may be helpful for Polybob to meet this test subject, and talk further with the referring agent, before finalizing a decision about whether or not to test this subject.

Peace,

r

------------------
"Gentlemen, you can't fight in here, this is the war room."
--(Dr. Strangelove, 1964)

posted 03-25-2006 06:08 PM            

Poly761,

Why don't we test those with any "significant medical / psyhcological issues"? I think we do all the time. We can test a psychotic if the person isn't presenting as such at the time. As I mentioned, I tested a pregant woman - a no-no in many minds. How many of us have run tests on people that we only agreed to do after a doctor cleared the person? We avoid medical issues because we fear injuring the examinee, something we want to avoid. We usually don't say no because the person is known to be unfit (by "unfit" I mean the test won't work) for polygraph.

How many people have we tested who are on an anti-depressant? If you read Matte's book, he thinks testing the clinically depressed is crazy, and many agree with him. How many people have come in with a list of their medicines and we wonder what pharmacy is big enough to stock all the stuff?

It comes down to this: after safety concerns, we must ask ourselves if there is a legit reason to believe we would be unable to create the delicate pscyh set in the examinee. If we can't set up the test to do so, then there's no sense in running it. If we think we can - and that's all we can do in soem cases - then we go for it.

Right now there is no research to support that a person with Aspergers is not a suitable candidate for psychological reasons. We are inferring that is the case based on what little we know of the potential candidate. Years ago some said - based on nothing more than their "educated" guess - that psychopaths are psychologically unfit for testing as they could easily "beat" the examiner. (Some still think that.) We now know that is not the case.

I've questioned whether I should run a test during the pre-test, but since I wasn't sure, I've kept going. That's one of the reasons we have the acquaintance test: to see if the person has an ability to respond. I've sent people off at that point due to meds that seem to flatten everything, and I've sent people home for a good night's sleep after seeing they're physiologically spent.

In the end it gets back to very basic polygraph theory: is one question type clearly more salient than the other. I've already mentioned some reasons I might be more reluctant to make a call one way or the other. The question here is whether you can create that psych set. The only way to really find out is to try. However, since there is no research to support that Aspergers subjects are good polygraph candidates, a DI call could cause you some headaches, depending on what a DI will mean in this case.

Good luck, and let us know what happens.

posted 03-26-2006 06:10 AM            

One reason (I) don't test a person with a (significant) medical issue is due to the fact the physiological parameters we measure might be distorted. Distortion can occur because of the problem (or) medication used to assist in dealing with the problem. Of course, the pre-test should identify any problems/medications we know that could distort the results/impair the examinee. In addition, a calibration test prior to each test should be able to identify if in fact the medication/medical problem prevents proper calibration of each component.

The second, and probably for me, the more important reason I don't test any person with a (significant) medical issue is that I choose not to accept the potential for liability attached. I won't place the liability on my client and didn't place it on my agency.

I understand a "medical release" of a pregnant woman (clears) the pregnant woman or person with other medical issues for examination. And, I also understand if "something" were to occur during the examination process I'd probably see my name listed as co-defendant in a lawsuit against the doctor who signed the release and is being sued for malpractice. I prefer not to take the chance that nothing will happen to the examinee during an exam.

While I don't know how many exams are conducted of persons with significant medical/psychological problems, I don't agree with your statement that testing a person with problems such as these is done " - all the time". I would appreciate any information you can provide indicating polygraph schools are teaching this is an acceptable procedure.

I'd be more inclined to (consider) an exam of a person with psychological issues. Do we, as examiners, know when (a psychological disorder is affecting a person as is obvious as when) a psychotic "is presenting?" No doubt an exam could be administered, but, how confident would we be of our opinion? Again, whether or not to administer the exam is based on specific facts that would help support our decision to administer the exam. The same information used to support and defend our reported examination results.

END.....

posted 03-26-2006 10:17 AM            

Some are testable, some are not.

As I've stated before, criminals are sometimes not healthy individuals. There may be times when becomes dangerous to others for us to completely turn our backs on some of these case. It might be wise for us to learn to address and discuss these special cases in an ethically and empirically responsible fashion.

I have a very interesting case that I worked on a short time ago.

I'll start another thread and get the test materials ready for review if people are interested.

The case has a tragic outcome.

r

------------------
"Gentlemen, you can't fight in here, this is the war room."
--(Dr. Strangelove, 1964)

posted 03-26-2006 05:51 PM            

As far as the pregnant person goes, you probably would be named as a defendant, but you've got a good defense. If you fail to test her (she was a job applicant), then you're open to an ADA suit, and for that you have little or no defense as you've got to make reasonable accomodations. If the doctor thinks she's fit - and only a doctor can make that medical decision - then she's fit as far as the law suit is concerned. What do you do with a job candidate who is bi-polar but "passes" the psychological portion of the job testing? That is, the psycholgist determines that with meds, the person is fit for duty. (There are bi-polar, PTSD, depressed, etc. police officers - and candidates - out there.) What do you do with

How do we know if a psychotic person is presenting as psychotic? Are you just pulling my leg? If the examinee is hearing voices or is seeing things that aren't there, then you can feel pretty confident the person is having a psychotic episode and you're not going to want to continue the conversation. If, however, he presents as you and I would expect the average examinee to present, then why wouldn't you test him?

How do you determine what is "significant" and what isn't? It sounds like a very subjective test. Even the psychological community has a hard time makind those decisions. They typically consider a psychological problem significant if it impairs the person's ability to function as he or she desires. What's significant for one, might not be for another.

Let's just stick to the basics: does the person present as one who is able to understand and follow our instructions? If so, does he or she appear to have the ability to establish a psych set on one question type over the other, if so, then we would appear to have a good candidate. The next question we must ask it whether the examinee has the physiological ability to respond to the questions we present. That is where medications and diseases might interfere.

posted 03-26-2006 08:22 PM            

Those are very good points, regarding people who are actively psychotic. I think what might be a little more complicated is testing individuals with a history of psychoticism, though not presently active. Anyone who works in prison settings probably sees people with such histories, and many people in prisons take multiple medications.

I think you can apply some clinical commonsense to some of these situtations. One assumption might be that persons who function optimally while taking prescription medications might also produce polygraph test data of optimal interpretable quality while taking any necessary medications.

There is curently no published research or theoretical rationale suggesting that any medications would contribute to erroneous polygraph examination results. That is the easy part of the equation. The complication is that the polygraph depends in part upon the normal functioning of a person's psychological set (including conceptual understanding memory recall, and ability to interpret contextual relevance or saliency), and the normal function of a person's neurolophysiological mechanisms (including brain pharmacology, and sympathetic functions).

Some medications do have exaggerating or dampening effects (sometimes refered to as sympathomimetic and cholinergic effects) that will affect the quality of polygraph test data. Some meds have extrapyramidal effects (though those are most older antipsychotics) that cause tic-like movements as side effects (which is worse tics, or elvis in a flying saucer.) Other meds, such as athma (steroid) inhalers, and others may have gluco-cortico effects that affect energy and activity levels.

The most likely result of the disruptive side-effects of medications can be expected theoretically to be an increased likelihood of obtaining inconclusive results (garbage in, garbage out kind of logic).

In testing science test methods are normed (and sometimes stratified) for applicability to certain populations through the data modelling and decision fitting with samples that are respresentative (through random sampling or match sampling methods such as political pollsters use) of the population for which the test method is intended. Data modelling and decision fitting tasks are ideally performed on different representative samples or a curious problem of overfitting results.

Several empirical principles are employed to assure the applicability of a test method to various populations. One principle is to include in the sample populations persons representing different diversity elements in proportions similar to that of the general population that the test methods will be intended for (i.e., persons of various age, gender, socioeconmic status and ethnicity in proportions similar to the population). If we intend a test method to be applicable to persons with serious mental health problems we would want our development sample to include persons with serious mental illness, in proportions similar to the general population. This alone does not assure that the test will work as well for those mentally ill minorities, and additional steps are necessary.

Sometimes test developers will establish an entire test method for a special population (as in the Gilliam Autism Rating Scales that are useful for teasing out differentials between Autism, Asperger's and other developmental disorders). Other times, a test method is devoped on a cross-sectional sample that is intended to be representative of the broader population (as in the Weschler and Kaufman intelligence tests). Kaufman, in his lastest iteration of his intelligence tests for children, has chosen to take additional steps to actually construct normative strata for various ethnic and socioeconomic groups. This is very interesting for comparing the IQ scores of a kid from an empoverished part of town to norms from other kids from similar neighborhoods (so a kid can be average for a certain sub-culture), but doesn't solve the problem that the kid might still be outside normal limits for the general population or another neighborhood. So sub-group norms are interesting, but don't always tell us much.

The general rule of thumb is that normative data interpretation rules and normative decision threshold apply to normal persons from the intended sample or population. Applying normative data interpretation and normative decision thresholds to exceptional persons or known outliers is somewhat more speculative. The results of an english language spelling test become meaningless when the test is given to a french speaking student.

Polygraph, as it exists in the realm of testing science, is no different. Our concern is whether or not persons with serious mental illness and multiple medications are represented in the design and development samples upon which the test was constructed - and whether there is empirically based theoretical rationale indicating that the test methods and physiological and psychological principles apply similarly to those persons. Our concern is also whether the same decision thresholds apply to persons whose psychophysiological response characteristics might be dampened or exaggerated.

So, to pose a simple question, can we argue that persons who require the medical or psychiatric administration of multiple prescription medications, just to get through a workday or schoolday, are normal persons who are represented adequately in the test design and decision thresholds underlying the polygraph technique? I suppose, depending on whom you are arguing against, that this could be a tough argument to win.

In the case I mentioned in the other thread, my stalker (on mulitple medications) showed up at his ex-wife's workplace and shot her, while in the parking lot. Exceptional persons can be every bit as dangerous as non-exceptional persons.

Peace,

------------------
"Gentlemen, you can't fight in here, this is the war room."
--(Dr. Strangelove, 1964)

posted 03-26-2006 09:37 PM            

I'm not qualified to make a determination of a significant psychological disorder. But, if a person arrives for an exam, identifies a past psychiatric/psychological problem and states they are aware they continue to have the problem I may not test. If a person arrives and states they have been depressed and have recently considered suicide (but not been seen by an MD) I would not test. Each and every examinee is evaluated based on their history, past and present.

If there is no research showing effects on a persons ability to respond there definitely is reasearch/literature (PDR) and your local pharmacist that will identify the effects a drug(s) will have on the person. As we know, barbiturates, amphetamines, narcotics and other drugs can affect the parameters we measure, i.e., increase/decrease heart rate, blood pressure, respirations, etc.

I believe the answer to your question is obvious. No, we should not test a person on more than one medication (if) it is known they can/will alter physiological activities we measure.

END.....