posted 11-22-2004 03:56 AM            

i"m really happy to have a place to changes ideas and experience.

i just want to know which test format (MGQT, Quadri-Track zone, Utah, Direct-Lie, etc.) are you using for diferents situacions (single issue, multiple issue, sex ofender, criminal , pre-employ, etc.).

Cheers,

posted 11-22-2004 09:53 AM            

I use the Bi-Zone when the issue is direct and simple. I use the ZCT when the possibility exist that the subject may have knowledge of the issue if not directly responsible. I use an MGQT when there are multiple issues (ie. accusations of oral sex, digital penetration, and forced masturbation between the subject and the same victim). I try to keep it simple.

Also, the type of MGQT I use bracketts each RQ with a CQ. This may not be the standard, but my preceptor was using this during my training period and I found that I like having comparisons bracketting each relevant. Also, it is easier to keep up with the evaluation without having to "jump over" IR and remembering which CQ goes with which RQ. The format is IR,SR,C,R,C,R,C,R,C.

I hope this does not set off a fire-storm with other, more experienced examiners, but I do not see how having each "R" bracketted by a "C" would not be considered a good thing.

I was not trained in the "Utah" format so I am unfamiliar with it.

Anyway, those are the formats I use most often. I do try to stay away from the ZCT as much as possible because I do not like the symptomatic question.

P.S., I am familiar with Directed Lie tests, but I have as yet not had the opportunity to use one. I have used a Guilty Knowledge Test once and I loved it, unfortunately the opportunities to use them are very few due to the need for enough information about the crime being kept secret to allow for enough items for use in the test. I can direct you to more information about this type of test if you are interested. You can email me direct at dmorgan@spartanburgcounty.org.

[This message has been edited by D. Morgan (edited 11-22-2004).]

posted 11-22-2004 11:56 AM            

I enjoy this site too. People always give me things to think about, and there is a lot of experience here. I've got a lot of theory down, but you can't beat the years of experience so many of the people who contribute to these discussions have.

I use an R/I for screening exams, and I follow-up, if necessary, with a comparison question test, generally a DoDPI Bi-zone. I like starting with an R/I because, done correctly, it's a tough test to get through if you're lying. The problem is it has a high false positive rate, but those can be cleared up with a good ZCT. (After all, if I'm going to make a mistake with a screening test, I'd rather make an error in that direction.) (I do use CQ screening test if I suspect a person won't be able to get through an R/I because of sensitivity to a particular question(s), which could have several causes.)

The test mentioned above is an Air Force MGQT, so it is a standard test. You can run it with four, three, or two RQs. Many end them with an RQ, but a lot of people like to end with a comparison so they have one on each side of any RQ. See the AAPP "Examiner's Handbook" for details.

As for a "ZCT," any test format that has spots in which an RQ is compared to a CQ is essentially a ZCT. The term "zone" was developed by Cleve Backster. His tests have three zones: a red (the RQs), a green (the CQs), and a black (the symptomatic / outside issue questions), so even an MGQT has at least two zones in it. Generally speaking, any test format that was modified from Backster's version is a Modified ZCT, so that term doesn't mean much. We call the DoDPI two RQ test a Bi-zone, but it's really a Bi-spot because it has two spots (but three zones). (I still use the name Bi-Zone, but it drives many of the Backster trained examiners crazy, which I can understand.)

As far a Utah test goes, they're pretty flexible. The people behind the Utah test (Raskin, Honts, et al) believe more in principles than rigid techniques. Essentially any test format (within reason) that includes a couple well-developed CQs and RQs would fit the Utah standards. Some Utah tests use symptomatics, some don't. The "Honts version" doesn't, and that's the one I generally use when I run Utah tests. The other "versions" generally put the symptomatic question in position one (where everybody responds to them). The reason is not to detect an outside issue (which there is no evidence they do), but rather to assure the examinee there will be no surprises on the test. I'd stick to the published formats, but you won't get any argument from the researchers if you stick to the basic principles of polygraph. By "basic," I mean the scientifically supported principles now being taught. For an excellent review of the UTAH and other methods, see the Honts' and Raskin's chapter in Murray Kleiner's book, "The Handbook of Polygraph Testing." (I have the chapter in pdf format. I got in from Charles Honts with his permission to share it, so let me know if you'd like a copy.) I'm working on a brief (about five or six pages) Utah "How to" guide for those who won't read the longer, more detailed (but, I'll admit, drier – to most anyhow) literature geared for the academic community.

The Concealed Knowledge Test is also in Kleiner's book, and Don Krapohl wrote an article on how to do it, which I think I have in pdf format too. It is a good test, but you've got to train people not to give away the information prior to polygraph. It's also a good test for a person who is otherwise unsuitable for testing, e.g., a person who has been interrogated (recently) or is too emotional for a CQT (like just after a murder, etc.).

posted 11-24-2004 12:55 PM         

I use the GKT a lot. I am on the major crimes team (as a polygrapher) and get called out on major crimes. My job is to see that some information is held back from the suspects and press.

I have also used the searching peak during investigations with good success.

I wear two hats. I am with Quality Assurance and investigations. I can't say enough about being on the investigative team. It has made polygraph a much more viable and accepted police tool.

Pre-employment: Federal LEPET format. It's defendable as DODPI will back it up if you ever get sued. Also, since it is now used by all the federal agencies, it's nice to say in court that you use the same format as FBI, DEA, CIA, Etc..

posted 04-18-2005 05:38 PM            

1. CHART1:I,SR,I,C1,R1,C2,R2,I,SY
2. STIM
3. CHART2:I,SR,I,C1,R2,C2,R1,I,SY
4. CHART3:I,SR,I,C2,R1,C1,R2,I,SY,C2,R2,C1,R1

I have 3 questions:
1. What is your prefer format ?
2. How many of you use SY questions? And in which format.
3. what do you think about this format ?

cheers,

posted 04-20-2005 08:06 PM            

How many You-phase tests do you have that are confirmed? I'm looking for those types of tests for a research project I'm doing, so are you willing to share? Right now, I need anything, but I suspect confirmed NDIs will be the toughest to obtain.

posted 04-21-2005 07:50 AM         

What do you think about Dayok's idea of having an archive of polygraph files we can shaer? I think it is a good idea, assuming it would be password protected. I once suggested a counter-countermeasure reading room to complement the countermeasure readingroom, but that idea never took off. Anyhow, if we were to have such an archive, I don't see why it couldn't be for anything relevant to polygraph.

------------------
John 8:32

posted 04-21-2005 01:27 PM            

I will be thinking about a way to give you that ability in the near future. I could upgrade this bulletin board to a newer version that allows uploads. That wouldn't be central but you could share files with each other that way. Unfortunately, I've worked a bit with the newer bulletin board and its actually not that user friendly because there are so many features. Will keep looking for a balance.

------------------
Ralph Hilliard
PolygraphPlace Owner & Operator
http://www.polygraphplace.com

posted 04-21-2005 01:46 PM            

There is actually very little research on the two-relevant question formats. As I mentioned, I am looking for confirmed tests of that type in order to produce suggested cut-off scores for Don Krapohl's OSS, which, at the same time, would be a validiation study of the format.

Don has provided me with the OSS spreadsheet template, which I've modified to do the calculations for two RQ formats, and Lou Rovner is willing to help with the number crunching or whatever else I need.

The last APA magazine had my data needs in it, which essentially says I want confirmed field tests - and not just the good ones.

If all goes well, I hope to verify or come up with hand-score cut-offs for examiners using a particular scoring system, but that's a ways off. Right now, I'm having trouble getting data. Everybody wants the results, but few want to give up the data. (Axciton and Lafayette will delete all ID info when copying exams, which is fine.)

posted 04-21-2005 06:41 PM         

I have no problem sending you confirmed charts, but unless you have a Stoelting CPSII system, they will have to be paper (I know, but I like to be different than everybody else).

I'm a little busy right now, but will work on getting it together in my spare time. I am currently testing for a federal intellegence agency (trying to leave this stinking desert!)and will be travelling a lot back east for the next couple of months. But, will be gathering in my spare moments.

When I was at a DODPI class last fall they were teaching Krapohl's line length method for scoring the pneumos, with a modification. DODPI uses any reaction in the pneumos, not just the first 7 seconds (?) as Krapohl scores. I have the OSS scoring system built into the CPSII, but find the pneumo scoring way off base, because of the time cutoff.

I was taught line length scoring in poly school, so it was nothing new to me, but they (DODPI), who had fought this method, were sheepisly admitting that Krapohl had been right all along.

posted 04-21-2005 07:45 PM            

Can you post the Federal LEPET format used for pre-employment exams please.

I also used two relevants (single issue) in my criminal exams ( C,R,C,R,C) but use the strongest control to either side for both
relevants, not just the proceeding control for the second.

posted 04-21-2005 08:07 PM            

I don't want to hand-score paper charts as that takes about 45 minutes per exam, and it does add a little variance to the measurements. I have software for every company out there, so I can view any type of charts. I don't need to view them at all though as I use Kircher's Extract program to just go in and suck out all the measurements automatically. Then I plug those measurements into a spreadsheet and let it do the math and produce a score. (Even that takes me a few minutes, but much less than 45.)

With OSS (hand-scoring) usually you measure line-length somewhere from 10 to 15 seconds from question onset. (I think the computer version of OSS uses 10 seconds since that's what the Extract program is set at.) As long as you are consistent, the numbers will work out. (I usually use about 12 or 13 seconds just to get more data.)

The only thing line length will not capture is a rise in baseline, so it's very accurate. (I should qualify that and indicate that it captures all the criteria associated with deception except a baseline rise. There are all kinds of breathing criteria that people score that have no correlation with decption.) Polyscore and Identifi (and OSS) both measure line length to compute their scores. (Your computer's scoring algorithm does the same thing.)

I'd appreciate anything you can provide.

AD,

posted 04-22-2005 07:21 AM         

I think the current message board is fine. It's easy to use and does what it's supposed to. As my grandfather used to say, "If it ain't broke don't fix it."

More important, I think it would be a mistake to allow people to upload files, because some FOG (or maybe even Georgie himself) will no doubt upload some virus or trojan. Also, having files scattered across messages wouldn't be as good as having a central, password protected area like the one you created for countermeasures information. Come to think of it, why not just expand the countermeasure reading room to include other relevant materials?

------------------
John 8:32

posted 04-22-2005 11:03 AM         

Well, that will make sending you charts a lot easier. I'll work on deleting biographical data (Stoelting doesn't do it automatically for some reason) and see what I can get together.

As far as the pneumos go, I agree with line length, I just think the computerized scoring isn't flexible enough yet and so far, I have found DODPI's modification (which they claim is the result of studies) to work better. As an example:

1. The tracings on the comparison question are consistant and virtually unchanged during the duration of the question, being of say moderate height (looks like normal breathing).

2. The tracings 2 or 3 seconds before the start of the relevant question are somewhat larger and stay that way 3-4 seconds into the asking of the relevant question, then just after the answer the subject throws a couple of great apneas followed by an almost straight line (a little exageration here for color since I can't draw on this post).

The computerized OSS will give the strongest weight to the comparison question since it only measures the first 10 seconds and ignores most of the reaction on the relevant. The OSS scoring system and I agree a lot more than we disagree, but about 2 out of 10 times I look at it and go "No way".

DODPI teaches taking the best reaction (line length) in one question and compairing it to the best reaction in the other (of equal durations), not starting a time frame at the start of the question. The reaction of course has to be timely. DOPI claims their studies show that having a set start point and duration misses a lot of data, and while I'm not a researcher by any means, my experience after attending the class is they appear to be right.

I am hoping to pick Dr. Krapohls mind some day about DODPI's modifications. If you ever talk to him about it, let me know what he says.

posted 04-22-2005 11:35 AM            

expanding on the countermeasures reading room is fine, but that is all manually done...ie time I don't have a whole lot of. My plan is for a place where I can stay out of the middle of things except where really needed. You guys log in securely and share all the files you need. Also, I would make sure there were no infiltrators or viruses. It will be coming...just hang tight.

------------------
Ralph Hilliard
PolygraphPlace Owner & Operator
http://www.polygraphplace.com

posted 04-22-2005 11:40 AM            

It's interesting to note you disagree with OSS about 20% of the time. When Don compared OSS to hand-scorers, OSS outperformed them.

All OSS looks at is line length, EDA and cardio amplitude.

It compares the strongest reaction (the shortest line in the pneumos and the biggest rise in the other channels) in the adjacent comparisons to the RQ. It weighs the EDA the most heavily.

You can run OSS by hand (time-consuming) and go out to 15 seconds to capture more of the reactions, if it applies to your examinee.

I might take a look at the DoDPI confirmed case database and see if any close calls (INCs) with a hand score can be brought into decision range by measuring line length - to see if the eyeballs missed anything. I've had tests with a lot of zeros in the pneumos, and I'm wondering if line-lengths would change that. After all, sometimes it's hard to see any difference, but measuring cleraly shows one reaction to be stronger. (Just how much stronger to make the point is something that needs to be figured out.)

What makes OSS nice is the research behind it. You can testify to the probability of truth or deception and establish error rates, etc., which is why it was designed.

I'm confused with what you said. Is DoDPI teaching the use of different time periods for each zone you score. For example, compare C4 to R5 for 17 seconds, and score C9 to R10 for 12 seconds so that you catch the entire reaction in the strongest one?

posted 04-22-2005 04:25 PM            

He is certainly intelligent enough and has done enough for this profession to seem like he has one but unless he got one since I asked him last it's non existent.

By the way, if your going to the AAPP come early and talk to Don at the ASTM meetings. Always has time for anyone and makes sense of things that some might find confusing.

Jack

posted 04-23-2005 12:44 AM         

I won't be at AAPP as it fell at the end of the fiscal year and the department preferred I go to APA, which is at the start of the new fiscal year.

What DODPI was teaching is not having a set reaction length or starting point. R5 is bracketed by C4 and C6. The reaction in R5 is 10 seconds long and starts 3 seconds into the question. C4's strongest reaction starts at the beginning of the question. C6 shows little or no reaction. You would compare a 10-second time span for each question, but the starting point for each is different. It is comparing reaction to reaction (or lack of if that is the case), not a set duration with a designated starting and ending point. The comparison time span will always be an equal amount of reaction time for each question, but it is not a set length. The next set of questions might be a 12-second length.

I'm not sure that it would make a difference if there was a standard set length for scoring (10 seconds?), but the hard fast rule of only scoring the length from question onset loses data. Scoring from the start of reaction (as long as its timely) for 10 seconds (or whatever) seems to me to make more sense than only scoring from the question onset.

As I recall, DODPI based this on the fact the physiological changes in respiration caused by the fight or flight response is often delayed in comparison to the EDR and Cardio. When blood volume increases and more O2 is flowing to the body, but the body is not expending an increase in energy, the amount of O2 intake is decreased as the body doesn't need as much, hence the change in respiration. Since the action potential that causes the change in respiration does not occur until the blood volume is increased and the body decides the blood is oxygenated enough, there is often a several second delay in a respiration change compared to the EDR and Cardio. DODPI argues that it only makes sense to base the length to be scored at the start of the reaction, not at the start of the question. I'm an old cop, not a researcher, but when it was explained like this it made sense to me, especially since I have disagree with the OSS scoring of the pneumos for the very same reason-ignoring scorable data.

My statement of only agreeing with OSS 80% of the time was only on the pneumo scoring. As far as the complete charts go, we are a lot closer, 95%? Usually it is only the difference of one calling it inconclusive and the other DI/NDI. Even then the inconclusive is leaning in the same direction. The 5% of the time we do disagree, it is usually because of artifacts that I was too lazy to do editing on, so the computer thinks they are reactions.

I like the OSS, but I guess years of having Stan Abrams beat me with a stick if I touched the computer scoring key is still with me

posted 04-25-2005 08:51 AM            

Search engines keep pointing me to companies that sell OSS software and I would like to read background information and see how it is hand scored.

I have a basic idea of the process, but without a source document of some type I don't really feel comfortable plunging ahead or investing in the software.

posted 04-28-2005 05:06 PM            

They do work, but I think you need to have a reason to justify the loss (though minimal) in accuracy.

posted 04-29-2005 06:52 AM         

The primary benefit of the DLC is in non-suspect, applicant or routine screening type testing. In fact, several studies leading to the TES demonstrated higher accuracy of the DLC technique over the PLC type testing in the screening format used. If memory serves me right, these studies were compiled under a 1995 report by the DoDPI Research Division, and referenced in the DLC article recently published in Polygraph.

Dr. John Kircher (and Dr. Honts, et al.) have done several studies utilizing DLCs and reported accurate results. One point that has always been overlooked by most using DLCs is that the DLCs cannot be evaluated in the same manner as PLCs, particularly the pneumo component. This has been argued by the "old" MI examiners for years and has been argued by research (Horowitz, et al., and more recently John Kircher, et al.). Both of these studies were also referenced in the article.

It's no replacement for the PLC technique but another highly valid and accurate techique that presents a unique value to those employing routine screening. However, one must be more conservative with its implementation (instructions to the examinee) and with its evaluation, than has been taught by most schools since its instroduction to the community in 1982 by Lou Fuse ("old" MI examiner).

By the way, there will be a two-hour lecture about DLC use at the APA this year. Check out the article written in Polygraph, I believe earlier this year.

posted 04-30-2005 08:02 AM         

You're right on. Since Fuse's introduction of DLC use to the world, there has been much debate about the value of discussing DLCs between charts. Some will argue that discussing DLCs (pushing?) between charts and evaluating the spots using the same criteria as used for PLCs will definitely result in higher false negatives. I'm amazed that results by Utah and others have been as good as they have been given these methods. Maybe the DLC is even more robust than some think!

Fuse and the original MI folks would never have done that (and yes, I've spoken with him about this). Their methods were much more conservative and the complaint in their arena was that DLCs were prone to FPs! But since it was used primarily as an initial screening test, they didn't mind. The examinee was not emotionally tweaked (unless he was lying and got himself wrapped around the axle) and therefore, no problem conducting more specific testing. This thinking follows the Krapohl suggestion of "successive hurdles." And these folks were doing it in the 60s and 70s! Don and I have talked extensively about this.

Another problem with the DLCs and one that (in my opinion) might make them prone to FNs is that most examiners believe the DLC needs to be significant and or emotion evoking. Quite the opposite. The examiner is challeged to balance the test through professionalism, competence and demeanor (good pretest)...if the examinee is truthful, he doesn't respond to RQs. If he lies, he does. Most of the scientists and examiners using the technique do not adhere to this belief...another possible explanation for the FNs, in my opinion. The tradeoff of course is that you may increase FPs. But if you adhere to Krapohl's suggestions, that might not be a problem (in that arena). (I accept I may be wrong for as you pointed out, there haven't been lots of field studies!)

One last word, bottom line with screening...you have to look at relevant responses. The original DLC use was scored more like an RI but one with some distractor for the examinee and one with a little more objective measurement. That was used in the Barland, Honts, Barger study (1989). Although criticised for analysis and protocol concerns, and an overall dreadful accuracy rate for all techniques uses, PLCs were FAR more prone to FNs than were the RI and DLC modules. The DLCs were far more accurate in that study. You can download the study from Honts' site.

It's not the anser for all testing, for sure, but no one technique is. I personally think that it is one which can be easily mis-used because of the unique aspects which are sometimes misunderstood ('pushing' DLCs...evaluating like PLCs, etc.).

posted 05-02-2005 12:01 PM         

Excerpt

US V Gilliard January 21, 1998 (Findlaw.com)
http://caselaw.lp.findlaw.com/scripts/getcase.pl?court=11th&navby=case&no=969459MAN

The Government, however, presented evidence suggesting that the addition of directed lie control questions would skew the results of a polygraph examination in favor of a guilty examinee. One of the Government's experts, Dr. Stanley Abrams, testified that he performed a study in which he compared the probable lie control question technique to the directed lie control question technique. Dr. Abrams found that people appeared more truthful under the directed lie control question technique than under the probable lie control question technique, and has concluded that the directed lie control question technique is substantially less than 90% accurate. Dr. Abrams testified that although using the directed lie control question technique reduced the number of false positives, i.e., innocent people appearing to be lying, it increased the number of false negatives, i.e., guilty people appearing to be innocent.

With respect to whether the hybrid technique has attained general acceptance within the relevant scientific community, Dr. Honts testified that the hybrid method is used by himself, Professor Raskin, the Arizona School of Polygraph, the Arizona State Police, and four other individual polygraphers. The Government, on the other hand, presented evidence that the hybrid technique is disfavored not only by the Government's experts, but also by federal government agencies. Dr. Abrams testified that the Internal Revenue Service and the Drug Enforcement Agency do not use the hybrid technique, and that according to the Department of Defense Polygraph Institute, none of the federal agencies are able to use it.

Considering the paucity of tests and published studies addressing the validity of the hybrid technique, and Gilliard's failure to show that the hybrid technique has gained general acceptance within the relevant scientific community, the district court did not abuse its discretion in finding the Honts Polygraph evidence to be inadmissible under Fed.R.Evid. 702. Furthermore, due to the lack of corroboration by other studies and polygraphers, the district court did not err in declining to assign much weight to the rate of accuracy that Dr. Honts attributed to the hybrid technique.