posted 03-26-2006 09:18 AM            

Effexor XR, (antidepressant)
Wellbutrin SR, (antidepressant)
Adderal, (central nervous system stimulant)
Tranzine, (anxiolytic),
Trazodone (antidepressant, used for sleep),
Haazar (for high blood pressure),
Nexium (for discomfort from a hiatal hernia),
Synthroid (Levothyroxine, for hypothyroidism),
Hydrocodone (narcotic derivative, for back pain), and
Zocor (for cholesterol).

This is a very interesting case. I'll put up more info when I get back from the gym.

r

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

posted 03-26-2006 02:04 PM         

Hydrocodone is generic for Vicodin. There is no way in hell you can get a good test with that in the examinee's blood stream.

On the other hand, the guy sounds like he is a walking Phamacy and you may want to keep him as a friend!!!!

posted 03-26-2006 02:31 PM            

I spoke with his therapist and attorney about his testability.

Sometimes the goal of testing has as much to do with what people will talk about while looking at all the polygraph stuff on our desks.

r

posted 03-27-2006 04:05 AM            

I've tested people on methadone without a problem, so oxycodone wouldn't bother me. (What I found is people on it need to sleep well and eat before the test - moreso than your run-of-the-mill examinee who might be able to get by on five or six hours.)

posted 03-27-2006 06:35 AM            

What suprises me is that other examiners included in the report, some standard boilerplate language that he was a "fit subject for the polygraph technique." I have nothing against biolerplated reports per se, but these seems an odd statement for this guy. Given these meds.

Effexor XR, (antidepressant)
Wellbutrin SR, (antidepressant)
Adderal, (central nervous system stimulant)
Tranzine, (anxiolytic),
Trazodone (antidepressant, used for sleep),
Haazar (for high blood pressure),
Nexium (for discomfort from a hiatal hernia),
Synthroid (Levothyroxine, for hypothyroidism),
Hydrocodone (narcotic derivative, for back pain), and
Zocor (for cholesterol).

He was on disability for severe depression and back pain, and had recently returned to work and been removed from disability as he had begun to look healthier.

With his improved health and activity also came increased dangerousness.

So, about a year previously, he disclosed to me that he had purchased a stainless steel .357 magnum revolver. His attorney would not allow us to force him to sell the weapon, as he is not subject to any court supervision. He did surrender the gun to his father who was previously unaware of the presence of the weapon in the home.

His ex-wife required him to attend sex offense specific treatment as a condition of his visitation with his son, as she divorced him after he sexually assaulted her older daughter. He was not charged for that assault as the victim was inarticulate and he lied to police investigators at the time of the investigation. We subsequently gained admissions through polygraph testing, after he was referred to treatment by social services. His attorney was effective at convincing authorities not to file charges based on the information gained at the polygraph (some fruit-of-the-poison-tree argument). His attorney also supported his wife's requirement for his participation in sex offense specific treatment, though that unfortunately put the burden of enmity on her.

I was tempted to call the test inconclusive, but attempted to score it anyway, as this is a dangerous subject whom I didn't want to give him a "walk" on whether he has to answer to his polygraph test results. (Had they scored truthful I probably would have called it NO.)

charts
http://www.raymondnelson.us/qc/060327.html

computer and hand scores
http://www.raymondnelson.us/qc/060327_ID.html

and report
http://www.raymondnelson.us/qc/060327_report.pdf

r

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

posted 03-27-2006 08:03 AM            

There appears to be an pneumo artifact on Chart 3/R8 contemporaneous with what looks like an ED arousal and a movement on the yellow channel. You hand scored this as a minus on pneumo what are you seeing that I didn't? It's hard to tell for sure with a compressed chart, but it looks like the ED arousal and lower pneumo might be pre stimulus onset.
I am unfamiliar with the second green tracing on your charts. I presume the yellow tracing is a movement detector.
ALSO what testing format are you using? I am unfamiliar witha format that uses 5, 6, and 8 as relevant questions.

posted 03-27-2006 09:29 AM            

Please identify for me, top to bottom of your charts, the components used in each of the channels and the location in which they are placed.

It appears as though someone forgot to activate & calibrate two components before beginning C-1.

Thanks.

END.....

posted 03-27-2006 01:34 PM            

On top of the first segment of chart that appeared when I opened the site was the following note: "Chart 1. This is the strangest electrodermal that I've ever seen." I'm referring to the segment of chart that can be seen (underneath) this statement as C-1?

When I look at C-1, I scan down the left side of the chart between SR3 and C4 and see two pneumo patterns, a third unidentified pattern and then see two flat lines/tracings being recorded onto the chart. These two lines/tracings then appear to be activated, i.e., turned on, as they continue and then begin to produce patterns starting just after C4 is answered.

One line/tracing, (the second recorded pattern up from the bottom), appears to be a cardio pattern.

By "calibrate," I mean once a component is activated, it is allowed to record a pattern; and this pattern is viewed long enough to ensure the examinee is functioning within an average range of the activity being monitored.

I'm very surprised you don't seem to know this or do you refer to this procedure in different terms?

Now, BELOW this segment of chart I've identified, is typed "Chart 2." Whatever. Two components do not appear to be activated on the chart, turned on, started, until just after C4 is answered. At about "1:20" on this segment of chart.

I'm not familiar with computerized instruments but it looks to me as those flat lines/tracings were created because the component producing the lines/tracings was not activated until after C4. Please let me know if my thinking is not accurate about this.

posted 03-27-2006 05:38 PM            

You're referring to the adjusting of the sensitivity until sensor tracings are of adequate size. (The beauty of the computer is that you can record at whatever you want and then view it at any sensitivity. You just can't print it out any differently than you recorded it. In other words, you can record at 0.1, but view at what it would have looked like if you recorded with the sensitivity at 5.0)

One of the sensors is an activity (motion) sensor. He has two. I think they are air filled, but Ray can tell you. I had one, and as the temperature of the air increases, the pressure changes, and you get a strange reading. You usually just re-center and adjust the sensitivity to fix it. I don't know why it's wacky.

The other green is the EDA (GSR).

It appears what you call activated is just him turning the sensitivity up to get a reading. It's hard to tell with the charts compressed like that. (There's probably more there in real time.) On a computer instrument, the tracings are on or off, but when on, the sensitivity could be at zero, which would give you that straight line. They could be close to zero, resulting in the line, which when compressed looks like it is flat. (And it does look flat.)

posted 03-27-2006 07:00 PM            

Believe it or not, the EDA sensitivity settings are very low (like 0.2).

other settings are standard about 1.0 for pneumos, and 10.0 for cardio and 5.0 for CM.

The dampening in the pneumos can easily be adjusted and accomodated for in the sensitivity settings, but I thought it'd be interesting to see the data without alteration.

I started the chart then he had to fidget a bit

so the test doesn't start until after about 30 seconds or so.

I crops the image to conceal the exam details and subject name, so the image starts at almost one minute.

The top two (blue) tracings are pneumos.

The bottom (yellow) is a pneumatic countermeasure sensor. I prefer the pneumatic sensor to the piezzo, as the pneumatic sensor models data more slowly and is more sensitive to low frequency movement (so the movement you see is actually mostly breathing movement. You can see occassional movement during test questions.

The red tracing is plain old cardio.

Test format is USAF MGQT with three questions. I usually review and add another IR at the end, in case I need to insert one, and 'cause I sometimes like to mutter (to myself, of course) about using those known truth questions for comparison and then say "just to run the test" to the examinee.

I am usually not too concerned about a single med, and have tested others on hydrocodone. (I have another interesting and troublesome test on a chronic pain patient, narcotic and cocaine addict, alcoholic, who now takes methadone and other narcotic derivatives for pain - he used to be a university physiology and biology professor).

My personal rule of thumb is to be cautious whenever anyone requires the daily administration of multiple prescription medications.

posted 03-27-2006 07:17 PM            

The Lafayette won't show both at once, but you can view it in either. I try to record in unfiltered (non-self centering), but other examiners hate it when I ask them to review my charts, so I usually use auto and then view it in raw. The problem is, I can't print it like that. (It'll only print as recorded.) I like that feature. I surprised Lafayette hasn't added that to their software.

You can do some neat things with Polyscore as well. For example, it'll show you a single line for the cardio - like the EDA, etc. (You can really see what it's looking at if you want to.)

Anyhow, I would have expected a dampening, but this guy apparently was a water hose. See how much you lose in self-centering mode.

posted 03-28-2006 06:42 AM            

Psychopharmacology is probably less like a game of teeter-totter (in which things can balance themselves out), and more like a game of mousetrap (with the lever, wheel, rolling ball, switch and basket that falls), meaning its not always simple to predict these the effect of the next addition.

As i indicated earlier, I get concerned anytime someone requires multiple prescription meds to function optimally. Referring agents got tired of me advising them about this and wanted their guys tested anyway. So, my response has been to include language in the test report that I consider the subject to be marginally suitable for the polygraph technique.

I think we, as a professional community have sometimes to do a better job at these things. I've seen a number of revocation complaints and terminations from treatment, in response to unresolved polygraphs (even though they say they don't do that) on subjects who take multiple meds or have serious health issues and can't pass polygraphs. In one case I worked on, the subject had taken 50 polygraphs (the PO said he would take it every month until he passed it) and completed everything in treatment that could be asked of him. He completed nearly eight years of probation without a revocation complaint, and was about to be revocated and resentenced rather than discharged at the end of his probation. No one asked any assertive questions about whether polygraph technique was being used in the most effective way - until his attorne became involved (this is not a conversation we want to have in court). Another case was a 300 lbs depressed diabetic with peripheral neurophathy (chronic pain in the hands and feet), on six different medications. So, why can't they pass their polygraphs??? One reason is they are dangerous, don't follow rules, and don't account accurately for incidents and behavior in the community. Our problem will be arguing the validity of the polygraph test results at their revocation hearings, and some attorney (and hired expert) will eventually figure this out.

What really surprises me is that more therapists haven't started asking questions, and some even resist when you try to discuss with them the limitations of the polygraph with their most unhealthy clients (we like simple answers). Actually many therapists are poorly trained in testing and measurement, and aren't such good diagnosticians (most don't need to be).

I had an interesting conference call today with a really good psychiatrist, who seems to know a lot about drug interactions. I had kicked out of a test a subject with a severe anxiety disorder that was not effectively managed (and history of addictive self-medication). Upon learning that his meds hadn't been checked in a year and were administered by a general practitioner, I sent him home and contacted his PO who ordered him to see a psychiatrist. He's also not passing polygraphs, and on his way to prison if he doesn't get it right. He might actually belong in prison, but he should at least have his anxiety disorder properly managed. I'd like to be sure his polygraph results are meaningful, and that depends in part upon his optimal functioning.

My stalker/shooter/perp (060327), was functioning optimally. The disability evaluator began to see that and insisted he return to work as he began to look and function healthier. So, he was actually functioning better, but the combination of meds is a apparently as issue for the polygraph. That doesn't always mean no polygraph monitoring to me, as there is sometimes a deterrent and informational benefit to testing these marginal subjects in PCSOT programs. I do object to the standard language about being completely suitable (fit) for the polygraph technique.

The artifact in the pneumos at R8 on Chart 3 is an answering distortion. I scored the baseline arousal against him, because I could (possibly out of meanness on my part?).

I've been running the limestone more often now. It has some quirks, but its a good instrument. I like the Axciton black background because its easier on my tired eyes, so I run the limestone in black. Black also requires less power when running the laptop on batteries (doesn't have to light the whole screen), produces smaller .jpg images. White background obviously requires less ink for printing.

Limestone recently updated (goofed up) some things, so it won't print black background to .jpg anymore, and wont allow me to rename the printed out (thereby over-writing the previous printed output - ARGH). Overall the Limestone is great.

I don't like compressed charts, but these data are so meaningless that I tried to save myself some time.

This gentleman (060327) showed up at his wife's workplace and shot her, about one week after this polygraph. I had alerted his therapist to what appeared to me to be obvious dangerous stalking, and insisted that someone assure his father had not returned his .357 to him. We basically saw this one coming, but couldn't get out ahead of the problem fast enough.

I never heard what he shot her with, but she survived a close range, point blank, assault. I had a brief telephone conversation with his attorney after the shooting, and heard a a bit about a cheap .380, but never got the facts. The ex-wife did survive, and I've not been able to find out how she is doing. I hope she is OK. Another tragedy is their son, whose father has now shot his mother.

Like so many other DV offenders, he wants to assign his behaviors and motives to the victim. His discussion indicates that he was intrusive about her personal affairs, and actually pressured her into attending counseling with him. My guess is that she did that in attempt to appease his increasing agitation towards her. The real effect of that was apparently to accelerate his hopes of reunification. She alerted his sex offense therapist to his stalking activities, but did not disclose her participation in other counseling.

This case has become a teaching example in other settings, as it has some interesting Tarasoff illustrations - though the ex-wife seems to have known more than the therapist in this case.

posted 03-28-2006 07:02 AM            

quote:

---

I thought that on the AFMGQT all relevant questions were even numbered.The 3R AFMGQT would be.
1.I
2.SR
3.C
4.R
5.C
6.R
7.C
8.R

By the way does anyone know if O.S.S. has been validated for the AFMGQT or other multiple issues tests? If so where could I find the scoring ratios.
ebv

---

Well you're probably correct on that. My exam templates get altered through use and I'll have to look it up and correct that. I'm mostly concerned about the question sequence.

Does that make the test invalid?

As I indicated earlier, I often add another IR at the end. It seems to me that with subjects who have taken 20-some-odd polygraphs the presence of the additional IR question helps in a number of ways.

Also, I have developed a habit of rotating both CQs and RQs on most PCSTOT (screening) tests, as these subjects are often focused o the next RQ.

When there is a known event or allegation, I rotate only the CQs.

I like the USAF MGQT because of it seems sensible to me. I can use two three or four RQs, though I don't like setting four comparison questions. The other thing that I don't like about this format is that the last RQ is bounded by only one CQ, so if that CQ is artifacted then there is no comparison for that RQ. That's why I rotate the RQs - without rotation it would be the same RQ every time, and my hypothetical brain tells my the question is more likely than others to yeild inconclusive scores. On single issue tests this is less of a concern, and sticking to the protocol become more important in diagnostic testing.

For four question tests I prefer the USSS MGQT - its described in the AAPP manual for anyone who wants to look.

1.I
2.SR
3.C
4.R
5.R
6.C
7.R
8.R
9.C

Same scoring schemes as other techniques.

This technique seems the most sensible to me. You can see that all RQs are bounded by two CQs, and it requires only three CQs. A couple of examiners in Colorado use it almost exclusively. With three questions, I've seen it like this

1.I
2.SR
3.C
4.R
5.R
6.C
7.R
8.C

So that is probably how my RQ numbers get goofed up, when I alter exam templates on the fly.

r

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

posted 03-28-2006 07:21 AM            

Adding an IR at the end, as a matter of form, may not make your data any less accurate, but it does make the format a MODIFIED Air Force Modified General Question Test. That is different than the insertion of an irrelevant where needed during a chart. I suspect that if such a test was submitted to APA for review, they would make an issue out of the fact that you didn't follow the validated testing format.

My list may not be complete and I don't have it with me now, but I don't recall the USSS test as being on the list of validated formats for PCSOT.

posted 03-28-2006 07:32 AM            

quote:

---

a MODIFIED Air Force Modified General Question Test.

---

Hmm. Good point. I'll have to think about whether I'd write that in report. Sometimes too much info is too much.

I don't like inserting unreviewed questions into a test (even IR questions), after I have told a subject there will be no other or unreviewed questions. So, I prefer to review it and have it whenever I need it.

Do you, or does anyone, have a list of validated techniques?

A couple of years ago, I saw a memo from Don Kraphol (I think) regarding this. It was pertaining to APA or ASTM, or DoDPI and this question of what are validated techniques. The content of the memo was that in strict empirical terms validation has been a big complicated thing, and that none of our techniques would appear to be validated in the manner often observed in sister sciences (I'll try and find the memo), and that we might be better off concerning ourselves with "recognized" techniques that are formulated on valid testing principles.

What makes a test valid is its construction according to valid principles (not question numbers, or other procedural rigidity).

If someone has any more current information on this, I think we'd all be interested.

Peace,

r

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

posted 03-28-2006 07:40 AM            

I suspect he either added an extra Neutral to the beginning here, or , the intro phrase is labeled as a question. I'm not sure. I have the Limestone software, but I've never used it as I've never had to look at any Limestone charts.

posted 03-28-2006 08:42 AM            

Yes, I have that same memo here someplace. Even though it lists formats as "valid," they are not all equal. (The Utah family of tests made the list.) For example, the Army MGQT is has been "validated," but it is less accurate than others, probably the AF MGQT, which I don't think did make the list. (Call it a Utah MGQT, and it's back?)

My OPINION: pick the best test for the situation at hand.

The AAPP Handbook is being revised and is supposed to be out at the April seminar. It still will list (unless it changes again) the AFMGQT as valid. The APA considers it a "recognized" technique, which now essentially means the same thing.

What Don's memo said, essentially, was to stop looking for cookie-cutter tests that we call "valid" and realize that a polygraph tests in the hands of a capable examiner is robust and valid if we follow certain essential procedures. Test format / structure is on the low end of the totem pole, and typically a matter of personal choice.

posted 03-28-2006 11:02 AM            

If, in the chart that was posted (C-1), the sensitivity remained set at zero for the identified GSR's, I would discard C-1 as what was described as the "centering" and "raw" GSR were not started until (after) C4 of C-1. Yes, even eliminating C4 for analysis there is C7 and C9, but I question the quality of the patterns.

The quick answer to the question on whether or not I would test is no. I use (and please refer to) comments I posted regarding Aspergers, medical, medications and psychological issues.

I would not test, based in part, on the total number/type of medications involved, reported "severe depression," purchase of a .357 (not a problem in most circumstances)and the sexual assault of the stepdaughter. I'd want to know more about the .357. Some of these and other issues alone might not cause me any concern, but add these all together for one examinee and I see a subject that in my opinion is not suitable for this test.

How was the depression resolved? The examinee began to "look healthier" and this was why he returned to work? I'm not interested in his work issues but "severe back pain" and medications used to relieve the pain are. I suspect there was more to his returning to work that just looking "healthier".

"rnelson." You report the subject as " - marginally suitable for the polygraph technique - due to multiple medications and potentially overwhelming affects of his reported medical and mental health concerns." But, you also state a person of this type "may also produce polygraph test data of optimal interpretable quality while taking any necessary medications," and "should always be regarded with caution." Too much CYA for me!

It's my understanding the USAF MGQT has RQ's at 4-6-8 & 10; CQ's at 3-5-7 & 9. Can the questions be adjusted as they appear on C-1?

Again, I'm not familiar with computerized polygraph or their charts. (If)each block on the chart, broken lines left to right, represent five (5) seconds of time, this examinee is showing an inhalation/exhalation cycle of between 48 & 72 per minute. Would this be accurate?

Regarding what I viewed as the reported cardio tracing on C-1, the tracing that begins just above the second pneumo tracing from the top (at SR3). In my opinion this tracing should have been centered prior to C4 being asked. This tracing was not manually adjusted to correct this problem. I don't see how any scoring/analysis can be made of a pattern as presented on C-1. This cardio pattern continues to "fall" on C-1 from SR3 to just after C9 (where I cut off my printout). Why?

Regarding the GSR, I'm not sure what to think of these two tracings/patterns. Neither are centered and not consistent with any GSR pattern I've seen. Is this due to the computerized system? Why did the two GSR tracings/patterns start out flat?

Why not call this test, any test, as it comes out after analysis? "rnelson," why attempt to " - score it anyway, as this is a dangerous subject whom I didn't want to excuse him from having to answer to his polygraph test results." Our job is to test a subject that is (suitable) for testing and render an opinion of the tests. Nothing more, regardless of (your) reasons or concerns! Why "push" a call? In my opinion, if I were the examiner I'd have to think I'm already on thin ice for conducting this test(s).

I don't like the quality of the charts but we have to work with what we have. I'm suspecting medications and psychological issues were at play here. Just a guess on my part but I suspect there is a lot more garbage to clear out from under this boys lid.

END.....

posted 03-28-2006 11:35 AM            

Why must the tonic level be constant to be scorable? We are looking for phasic responses, and they still appear to be there.

No, he's not breathing 70 breaths per minute. Count them and look at the times below. He's breathing a little fast (I'd say about 22), but not that fast.

When it all comes down to it we're looking for physiological responses that correlate with deception. It doesn't matter if he's breathing fast or slowly (within reason). If the rate slows, it's a reaction. If it increases, it isn't. If the baseline rises, it's a reaction we score, and if it drops, we don't.

That's one of the reasons CMs are so hard to pull off, even with controlled breathing there is still usually enough suppression to score the charts and not get fooled.

posted 03-28-2006 12:47 PM            

quote:

---

I always review 1 or 2 extra irrelevant questions with examinees in case they are needed.I agree that you should never ask an unreviewed question. I understood your previous comment to mean that you ended each chart with an Irrelevant question.

---

Ebvan,

This sounds like a good idea.

I like to add the question routinely as including these questions help with the misinformation campaign that the IR questions are the CQs (it doesn't always work, but word gets around the treatment programs, and its fun - it doesn't work if we run USAF MGQT with one one IR at the beginning).

quote:

---

if we add an irrelevant or a comparision question to EVERY AFMGQT chart that we run in the same location it isn't AFMGQT. It's something else. It may follow accepted principles, be as accurate, reliable, and valid as the day is long and but we really need to call it something else.

---

OK, I see the point.

I've notices that some examiners have taken to noting in the exam report that a CQT (comparison question technique) was employed.

quote:

---

What Don's memo said, essentially, was to stop looking for cookie-cutter tests that we call "valid" and realize that a polygraph tests in the hands of a capable examiner is robust and valid if we follow certain essential procedures. Test format / structure is on the low end of the totem pole, and typically a matter of personal choice.

It's a nice letter. If I can find it I can probably scan and save it as a PDF doc if anybody wants it.

---

Barry,

if you can .pdf the memo, I can make it available for others to read. Its worth reading.

quote:

---

Regarding what I report as "calibration." I calibrate (or how ever you choose to phrase it) the examinee. Yes, by activating a

properly. I view the pattern/tracings for 20-30 seconds to ensure they remain within this average range. If I'm losing pressure in the cardio or pneumos, can't center the GSR or identify any other number of problems that could occur and prevent a clear pattern from being recorded I don't test until the problem has been resolved. Too far above or below average for the examinee I don't test.

---

The EDA was responding at test onset, but I've croppped that part of the chart due to examinee name and exam number. After onset, the EDA flattened out and then came back with that craziness.

I think its possible the guy overmedicated himself that day and wouldn't tell me.

quote:

---

If, in the chart that was posted (C-1), the sensitivity remained set at zero for the identified GSR's, I would discard C-1 as what was described as the "centering" and "raw" GSR were not started until (after) C4 of C-1. Yes, even eliminating C4 for analysis there is C7 and C9, but I question the quality of the patterns.

The quick answer to the question on whether or not I would test is no. I use (and please refer to) comments I posted regarding Aspergers, medical, medications and psychological issues.

---

I think we all agree these data are largely invalid.

The computer software doesn't include the option of turning off the EDA partway through the test. So it is on throughout the test.

Additionally, on the Limestone system, sensitivity adjustments affect the entire chart - not just from the point of adjustment onward. (On the axciton you can adjust the sensitivity starting at a certain point only, and a sensitivity adjustment marker is inserted in to the chart.) Limstone inserts no marker be cause the adjustment affects the entire tracing - even the part recorded before the adjustment.

The sensitivity setting is actually very low.

I still want to know more about the .357 and whatever weapon was used. (It seems it was not the .357 from the year before.)

I also agree that it would be legitmate to not test this subject.

I had consulted with his attorney and therapist prior to testing. His attorney was hoping to exculpate his ex-wife's stalking allegations.

Had we not been testing the guy, he certainly would have had access to the .357, which, at point blank range might have done a bit more damage to the ex-wife than whatever he used.

quote:

---

Again, I'm not familiar with computerized polygraph or their charts. (If)each block on the chart, broken lines left to right, represent five (5) seconds of time, this examinee is showing an inhalation/exhalation cycle of between 48 & 72 per minute. Would this be accurate?

---

on this chart the vertical lines are 20 second intervals.

quote:

---

"rnelson." You report the subject as " - marginally suitable for the polygraph technique - due to multiple medications and potentially overwhelming affects of his reported medical and mental health concerns." But, you also state a person of this type "may also produce polygraph test data of optimal interpretable quality while taking any necessary medications," and "should always be regarded with caution." Too much CYA for me!

---

I know this sounds somewhat ambiguous to anyone who wants a concrete answer. But some of these things are hard to anticipate in advance. Sometimes all we can to is advise people about the various concerns.

Its not just a case of CYA. There are a lot of PCSOT consumers (therapists and supervising officers) who have very simplistic expectations from the polygraph.

As I indicated earlier, other examiners had tested this same subject, before he returned to work (while on disability) and called him a "fit subject for the polygraph technique." Sometimes we have to educate our customers about what to expect.

On the spina-bifida case - I noted the marginal test data for the PO, and she said "he's passed polygraphs before and none of the other examiners have ever said anything about this."

quote:

---

Regarding what I viewed as the reported cardio tracing on C-1, the tracing that begins just above the second pneumo tracing from the top (at SR3). In my opinion this tracing should have been centered prior to C4 being asked. This tracing was not manually adjusted to correct this problem. I don't see how any scoring/analysis can be made of a pattern as presented on C-1. This cardio pattern continues to "fall" on C-1 from SR3 to just after C9 (where I cut off my printout). Why?

---

I manually moved the tracing up.

The Limestone polygraph offers two ways to make adjustments: 1) the normal way that segments the tracing and inserts a centering marker (not a sensitivity marker) or 2) if you click the sensitivity adjustments, the whole tracing centers, without segmenting the tracing and without inserting a marker.

quote:

---

Why not call this test, any test, as it comes out after analysis? "rnelson," why attempt to " - score it anyway, as this is a dangerous subject whom I didn't want to excuse him from having to answer to his polygraph test results." Our job is to test a subject that is (suitable) for testing and render an opinion of the tests. Nothing more, regardless of (your) reasons or concerns! Why "push" a call? In my opinion, if I were the examiner I'd have to think I'm already on thin ice for conducting this test(s).

I don't like the quality of the charts but we have to work with what we have. I'm suspecting medications and psychological issues were at play here. Just a guess on my part but I suspect there is a lot more garbage to clear out from under this boys lid.

---

All good points. The data really are somewhat meaningless, and it may be best to call this type of think inconclusive.

I also, in prinicple, dislike test questions that attempt to test the limits of admitted behaviors. like these "more than X times," though we use them routinely.

By now it is obvious that this test was conducted primarily as an opportunity get more information.

Clearly, this is one of most unhealthy test subjects.

Peace,

(i'm going to go purchase a teflon-nomex suit now)

r

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

posted 03-28-2006 05:04 PM            

"So what if he starts with CQ whatever?" Are you indicating there is no problem in starting any examination w/a control question?

What is "a test for utility - not validity?"

The examiner stated the test was conducted primarily to get more information. Is this a utility test? Based on the report it appears substantial information was learned during the pre-test that was beneficial.

"So what's" and "who cares" are what will continue to create problems for this profession. As I believe a member indicated earlier, we need valid standardization and we must work off the same page.

What "phasic responses - still appear to be there?"

How many indices of deception are you looking for in each component when you analyze a test? If you don't mind, what physiological responses in C-1 can you say correlate with deception. For exammple, R5 on C-1. Do you believe there is data in C4 or C7 that can accurately be evaluated against R5? I don't! The pneumos on C-1 are about the only component I would say are within an average range and suitable for evaluation/analysis from SR3 to C9.

Personally, I've never really been impressed by statistics. They too often tell a person what they want to hear. And for this reason, no, I don't know much about them. I don't have a clue about what percentage of the population people on multiple drugs represent. What I do suspect, using common sense, is that there is a very good possibility a person on multiple drugs is probably not a person suited to take a polygraph test! This statement is in no way is intended to demean, insult or cast a negative light on anyone that chooses to conduct a test of a person using more than one drug. We all evaluate the examinee and circumstances and give it our best.

I believe the test we are discussing can be used as an example of why we should think twice about testing a person using multiple drugs. I don't know if this was the only reason the charts turned out as they did but I suspect the drugs played a large part.

END.....

posted 03-28-2006 06:15 PM            

I don't know how to do the quote thing, so I'll just drag your comments / questions up and answer them as best I can below them.

"So what if he starts with CQ whatever?" Are you indicating there is no problem in starting any examination w/a control question?

No, I meant it doesn't matter which CQ he puts in the first CQ position. The first question should always be a buffer as the examinee is going to respond to it (an OR).

What is "a test for utility - not validity?"

A test for "utility" simply means it is a test for its usefulness. That is, the test has some value to the examiner or those involved. The costs of errors is presumed to be lower in a test in which utility is the goal. Srceening exams are utility tests. So the answer to your next question is yes.

"So what's" and "who cares" are what will continue to create problems for this profession. As I believe a member indicated earlier, we need valid standardization and we must work off the same page.

I disagree. Debating things that don't matter and claiming they do is the problem. What is "valid standardization"? If it means what I think it's impossible.

How many indices of deception are you looking for in each component when you analyze a test? If you don't mind, what physiological responses in C-1 can you say correlate with deception. For exammple, R5 on C-1. Do you believe there is data in C4 or C7 that can accurately be evaluated against R5? I don't! The pneumos on C-1 are about the only component I would say are within an average range and suitable for evaluation/analysis from SR3 to C9.

Yes, I believe there is data that can be evaluated in C4 R5 C7. I don't know what "accurately evaluated" means. If it means can I follow validated scoring criteria, then yes I can. How confident I am is another question, and that's where statistics comes in to play.

Anyhow, the amplitude increase - the mess that it is - is greater in the CQ. That's the standard to score it. Now you might not like the quality - I don't - but this is a test for utility, and there's data there to be mined.

If you look at the downward trend in the cardio you'll notice that stops in C4 - on time - and increases slightly. It appears to do the same in R6, but is concealed by the EDA. Also notice it stays elevated when he clearly wants to drop his pressure (note the downward trend that pauses for 30 seconds or so after the RQ). At CQ9 it drops. Why is arousal greater in the RQ? We can come up with lots of reasons and sound like the anti crowd, but are the RQs (or is a particular RQ), overall, more salient than the CQs? If so, we presume, deception.

Personally, I've never really been impressed by statistics. They too often tell a person what they want to hear. And for this reason, no, I don't know much about them. I don't have a clue about what percentage of the population people on multiple drugs represent. What I do suspect, using common sense, is that there is a very good possibility a person on multiple drugs is probably not a person suited to take a polygraph test! This statement is in no way is intended to demean, insult or cast a negative light on anyone that chooses to conduct a test of a person using more than one drug. We all evaluate the examinee and circumstances and give it our best.

Science and common sense often clash. Lykken says the CQT doesn't work because it offends his common sense. We dismiss him because he ignores the science. You want to take science out of the argument and replace it with tradtion and what seems right to you. How do you know that your common sense is correct? You can't unless you test your hypothesis using the scientific method. You need to analyze your data to determine if what you've observed was the result of X or just chance alone.

Polygraph would not be where it is today were it not for statistics. We wouldn't know where optimal cut-offs are were it not for statistics. We wouldn't know (we might suspect) a number of things about polygraph were it not for statistics. I think it's wrong that many sit back with that attitude and reap the rewards of a few other's work.

Let me take a stab at this "valid standarization" nonsense, and as I mentioned, I might not really understand where you're going with this one, but let me make this point as it comes up all to often.

We know that certain researchers have tested polygraph, and they found that it discriminates liars from truthtellers at better than chance rates (because of their employment of statistical analyses). Now, many seem to think that just because the researchers used format A, for example, that format A is "valid." Let's say format A looks like this:

N SR C1 R1 I1 C2 R2 C3 R3 I2

If you read the studies, you'll often find that the researchers used the same RQs for everybody. They also used the same CQs. So, technically, any test that uses that format, but uses different CQs, RQs, or a different order of presentation (that is C2 first, instead of C1) has conducted a different test, and that differs from the validation study.

The question is, has the CQT been studied under enough different conditions and found to still work (after all the staistical analyses were conducted and verified our suspicions)? If so - and it has - what variables are those that are importand? I believe we are at the point at which we have demonstrated there are things more important than the debates we continue to have, e.g., which "format" is superior to others. If you talk to or read the works of the scientists / researhers out there you'll find they are tired of having these debates too. People don't understand the scientific method, research methodology, statistics / probability, so we continue to argue nonsense.

We've got to end the double-talk. When I went to polygraph school, we were taught it is unethical to run a self-made test. It will still say that in the AAPP handbook. However, we were also told if we wanted to cahnge things to just call the test a General Questions Test, which as most know, lets you create a CQT that looks a lot like all those "recognized" techniques, but might have a couple extra neutrals or look the the "Colorado" format above.

Additionally, Honts and Raskin told me themselves that their chapter in Kleiner on polygraph validity applies to ANY CQT format that is otherwise properly conducted, which I won't go into here. They explain it well enough. How can they say that? They've got a lot of data to base it on. If we all better understood what they do, then we wouldn't be fighting amongst ourselves over things that don't matter. Many things do matter, I agree, but this format X is valid and format Y isn't, is, for the most part, hogwash.

The reason I am for the "making your own test is unethical" statement is because examiners who don't fully understand polygraph science would be creating junk - not because a format itself was bad, but because they failed to understand the essential principles of polygraph - the principles that make any test "valid."

posted 03-28-2006 06:37 PM            

Its a large image .pdf - about 1.8 megs so beware if you are using a slow dialup.

Thanks Barry, I didn't have time to look for it yet.

I think we might all be arguing from the same side of the desk here.

Everyone likes to see test data with consistent tonic levels and the empirical logic that reactions are meaningful is much more robust than in this test, in which there is a very strange fluctuation in tonic levels (while the subjects consciousness did not appear to fluctuate).

quote:

---

Personally, I've never really been impressed by statistics. They too often tell a person what they want to hear. And for this reason, no, I don't know much about them.

---

"There are three kinds of lies: lies, damn lies, and statistics."

--variously attributed to Mark Twain, Benjamin Disraeli, and others

- I vote for Twain

Whether we like or understand statistics personally does not matter. The fact is that without statistics the polygraph would have no validity. The polygraph is intended to investigate the veracity of issues of concern, for which their is no physical substance. Epistemologically speaking, truth is not a thing (such as a chair) or event (such as driving a vehicle in front of an ex-wife's home). In some case, there is physical evidence (shootings for example), meaning the shooting event is not in question, though whom is responsible might sometimes be in question (not is this case as the subject fled and drove his car off the road later that night). The measurement and interpretation of amorphous phenomena is a statistical problem.

In the case of the polygraph, the underlying statistics could include common z-score methods to compare individual scores to a normative sample, and might also include standard errors of mean difference to establish a statistically significant difference between the theoretically bimodal mean scores of a development sample of both truthful and non-truthful persons.

Thankfully, we examiners don't have to trouble ourselves with that brain-damage, and just count reactions until we get a certain number. To be valid, that number must be describable according to predictive statistics, which apply to a representative sample. These are common methods for any graduate or post graduate stats course in the social sciences.

quote:

---

I don't have a clue about what percentage of the population people on multiple drugs represent. What I do suspect, using common sense, is that there is a very good possibility a person on multiple drugs is probably not a person suited to take a polygraph test!

---

I agree in part and disagree in part.

It seem a weak argument that people who require multiple meds to get through a workday represent normal persons for whom the polygraph technique was intended. And, Barry, I'm not at all sure the development samples upon which our techniques were developed included such persons in proportions similar to the population. I think Poly671 is arguing my point that these people are probably best (conservatively) considered to be known outliers to the normative sample and intended population.

The problem is the criminals are sometimes (sometimes often) not normal healthy people.

I suspect that those of us who work in PCSOT program are asked all the time to test people on multiple meds.

Prisons don't hesitate (in my view, though I recently argued with a prison psychiatrist about this) to put people on meds, as they do reduce the risk of stress exacerbated institutional behavior problems.

If we turn these test down it becomes a bit marketing problem. now doesn't it. As ugly as that is, lets not fool ourselves about this. POs and therapists who desire simple answers will simply find and examiner who will tell them so-and-so is a fit subject for the polygraph technique.

Some of stated objectives of PCSOT programs are to get information that would otherwise be concealed, and to deter problems by increasing the likelihood of detection and consequences - this is in addition to the decision-support value of the test results. Some people, like Kim English, have suggested that any of these goals is sufficient to employ polygraph in PCSOT programs.

The informational goals of polygraph in PCSOT programs is not that different from the common practice of police examiners polygraphing (and clearing) statutory rape suspects on the use of violent force, as an opportunity to gain a confession regarding "consensual" sexual contact with a minor.

The empirical problems associated with these tests are real, and are probably best accounted for by our development of the vocabulary and ability to have rational and empirically responsible conversations about these issues.

The ethical complications of these tests are actually simpler, and has mostly to do with what bad things might happen to a person if he takes the test. The anti-polygraph community beats a regular and loud drum about polygraph harming people, and they want to foster this impression. In this case example (060327) nothing bad happened to the guy from taking the test. Nothing bad could happen to him. He was referred to an anger management group, and there was an effort to assure his limited access to firearms. The polygraph didn't prevent all tragedy in this case, but cannot be expected to. The polygraph simply got information - which did not harm the subject. He was not charged or incarcerated for taking the polygraph. To suggest that polygraph testing harms people is to foster the anti-polygraph argument. Now, if someone were to suggest revocating and resentencing this subject based on his polygraph results, I would object to that - as it shifts responsibility for a professional decision onto the test. Tests don't make decisions, they give information. Decisions are made by courts and thoughtful professionals.

So, the ethics of disclosure and deterrent objects are probably less complicated that the police example above, in which the test subject is unwittingly maneuvered into confessing to an act that will result in charges. (That example is not that complicated since the courts have already decided that such confessions can be used, just as pretext phone calls, and "apology" letters which investigators have some alleged offenders right to their victims.)

quote:

---

This statement is in no way is intended to demean, insult or cast a negative light on anyone that chooses to conduct a test of a person using more than one drug. We all evaluate the examinee and circumstances and give it our best.

---

I think its really important to be able to discuss these empirical issues in a professional forum.

quote:

---

I believe the test we are discussing can be used as an example of why we should think twice about testing a person using multiple drugs. I don't know if this was the only reason the charts turned out as they did but I suspect the drugs played a large part.

---

Agreed. We should absolutely think twice about testing people on multiple meds. We should also thoroughly consult the referring agent regarding the objectives of testing, and educate out customers and consumers about what they should really expect.

In all likelihood the anomalies in these (060327) test charts are the result of this subjects multiple medications, and the underlying medical and psychiatric issues that necessitate those medications (and the possibility that he attempted to over-utilize his meds on the date of the examination).

As I indicated in the report, sometimes people who function optimally on meds also produce test data of optimal interpretable quality while taking multiple prescription meds.

This test (posted a while ago) is on a severe PTSD patient with multiple medical and psychological diagnoses.
http://www.raymondnelson.us/qc/050204.html

He has one of the most severe cases of PTSD that we've seen

He's studied polygraph CM's, but the interesting thing here is that the physiological responsiveness of his data are largely indistinguishable from persons without serious medical and psychological diagnoses.

I'll find the med and diagnostic info if people are interested.

posted 03-28-2006 09:01 PM            

Starting a new topic whenever we change gears for a length of time would make it easier to follow those other side discussions.

There are studies that show that some medications do not affect the overall diagnostic ability of a polygraph examination and none that I know of that conclusively suggest otherwise (statistics at work). This is not the same as saying that medications do not affect the physiology of a subject, the physiological tracings, or the ability to come to a diagnosis. It simply means that the overall diagnostic ability of an examiner to form an opinion on physiological tracings received from subjects on those medications tested has not been shown to be reduced below the degree of acceptance (I know, more statistics).

My opinion is that Ray did the appropriate thing by conducting the examination, seeing what physiological data he received, and then diagnosing the physiological data he received.

Medications, however many, may have varying effects on each individual. Hence, the reoccurring listed warning of side effect symptoms “may include headaches, stomach pains, nausea, constipation, and diarrhea” on most medication commercial on television.

We need to be mindful of what, if any, medication(s) a subject is taking, the subject’s known or reported medical condition(s), if any, and their potential effect(s), if any. This is for obvious reasons that I doubt need to be addressed. If we are not comfortable with the medication(s) and/or medical condition(s), we can always seek a professional opinion of those qualified to clear the individual to take a polygraph examination. If we are still not satisfied, we can opt not to test the individual.

I am certain that everyone here has had at least one subject who produced different physiological tracings on different charts and the same chart. Sometimes these variances may be due to an intentional or unintentional act by the subject, an error on our part, etc. Sometimes we just do not know what has caused a change. In a diagnostic test we are simply forming an opinion on the usable data, if we can get an appropriate amount of usable data to form an opinion. If we cannot, we do not give an opinion when there is insufficient usable data. E.g., if a subject moves on the first relevant issue chart causing it to be unreadable and then stays still on the second and third, our opinion as to the diagnosis will be based on those readable charts (considering that this is within acceptable practices in your jurisdiction). If before, during, or at the end of the examination we as practitioners are not comfortable with either or both the subject and/or the physiological data we can always choose to forgo an opinion and suggest a re-examination.

posted 04-01-2006 07:38 AM            

A couple questions about information I'm not familiar with as I don't know anything about a computerized system. What is a "raw" and "centering" EDA (GSR)?

Also, in the hand score you presented you only gave a plus and minus value versus a numerical score. What is the reason a numerical value is not assigned?

Thanks again for all your information.

posted 04-01-2006 12:16 PM            

A couple questions about information I'm not familiar with as I don't know anything about a computerized system. What is a "raw" and "centering" EDA (GSR)?

Also, in the hand score you presented you only gave a plus and minus value versus a numerical score. What is the reason a numerical value is not assigned?

Thanks again for all your information.

posted 04-02-2006 11:14 AM            

Raymond must be elsewhere, so I'll answer your first question. The "auto center" is just like on an analog instrument. The instrument automatically centers the pen to keep a stable tracing. The problem is that you lose some of what is there as you've altered the tracing. "Raw" is just that: the tracing raw, in that it is free flowing, so to speak. It requires a lot of centering often times, but you don't lose anything (except what your instrument filters, which you can't control anyhow).

On the Limestone, you can see both at the same time. I prefer "raw," but since most examiners prefer auto, I use that so they are more willing to QC my charts. (If you're not used to raw, it'll drive you crazy at first.) Many say always use auto; others, never. This is the best of both worlds.

As for your second question, here's my best guess as I know another who does it. He's using a three-point scale: +, -, and 0. It's the same as +1, -1, and 0, and similar to the confusing d, DD, t, or whatever it is I learned (and forgot) when studying the Backster system.

Taylor,

posted 04-02-2006 01:55 PM            

Barry and Taylor,

I would be very careful trying to draw any conclusions about prevalence of meds based upon any of our anecdotal experience, as this would be heavily influenced by the “selection bias.” The same problem exists for DI and NDI rates and problem behaviors, which may vary across types of programs and working situations. I work closely with a couple of programs, some of which specialized with very disturbed and complusive persons. One is the “last house on the block” (before prison) when offenders are discharged unsuccessfully from other programs, and a couple of others do a very good job structuring and monitoring offender's lifestyles (which affects polygraph outcomes). Certain programs which specialized with higher risk, treatment failures, and multi-problem offenders will yield different polygraph “experience” compared with the average outpatient sex offender treatment program that serves community sentenced offenders who are most often estimated as low to moderate risk for recidivism.

Poly761

I thought you might be asking about the EDA/GSR confusion.

EDA is probably the more correct term for this. GSR is a bit arcane, though still used. Part of the confusion is that older instruments may have used the term to refer to resistance, while most newer (computerized) instruments are probably not limited to resistance as the measured response characteristic – though they tell us very little about these things.

Here is a chapter from a psychophysiology text – it indicates that GSR is only one of many electrodermal responses -
http://www.raymondnelson.us/qc/EDA.pdf

from

Jaakko Malmivuo & Robert Plonsey: Bioelectromagnetism - Principles and Applications of Bioelectric and Biomagnetic Fields, Oxford University Press, New York, 1995.

also an article on 3 and 7 position scoring (the K-man is beginning to assume mythical proportions here)

http://www.raymondnelson.us/qc/1998_3_and_7_position_scoring.pdf
(this should be a separate thread)

When scoring systems are somewhat distinct and none really outshines the other under empirical scrutiny, then scoring systems become a matter of policy, not science. This article seems to address mostly the numerical portion of scoring, and decision threshold issues, while other articles address the data (feature) analysis. There is a more recent article on feature analysis suggesting that the simpler Utah system offers some advantages (simpler systems are generally more reliable).

More later.

r

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