posted 10-16-2008 08:17 AM            

I am grateful to ebvan for both the positive review of our journal article and for pointing out at least one of the likely numerous mistakes I let slip through. I take full responsibility for the mistake.

I do, however, share responsibility for any potential positive contribution with my co-authors and the many friends and colleagues who helped us by reviewing and commenting. We worked with a number of people on this, not the least of whom is Paul Menges of DACA. I consider Paul to be the most knowledgeable examiner or expert on Directed Lie Comparison (DLC) questions. Please note none of what we published necessarily reflects the beliefs or personal/professional policies of Paul Menges (or any of the other reviewers of the paper) or any government agency.

Since writing this we have heard from a few people who have tried this approach. I have had correspondence with Laura de Wells Perry of the Arizona Polygraph School and Ben Blalock in Florida regarding the application of the DLST to their work. Both have told me they experienced favorable results as compared to alternative, multiple issue techniques they employed in the past. Ben told me he has used the technique in over 100 law-enforcement pre-employment exams and Laura has used it in the PCSOT milieu. I hope others will consider this and report back to us success or failure.

I started a new post regarding the DLST in hope of additional comments. It is through thought provoking dialogue and civil discussion we are able to envision and achieve improvements. I believe much of the work has been done before us and all that is needed is an amalgamation of that work and then we test our hypothesis.

Thanks again ebvan for your comments. Much work goes into writing these pieces and comments like yours make it worth doing.

posted 10-16-2008 08:38 AM            

------------------
Ben

posted 10-16-2008 12:40 PM            

I've been very interested for some time in implementing a DL test, but didn't feel I really had the grasp of it enough to do so.

------------------
Ralph Hilliard
PolygraphPlace Owner & Operator

Be sure to visit our new store for all things Polygraph Related
http://store.polygraphplace.com

posted 10-16-2008 05:07 PM            

I spoke to Paul about these tests at some length during APA, and we really need to get him out in public explaining how he sets up the DLCQs in these tests. He's very good. (I'd just like to watch a video of him giving the intro over and over until I've got it down!)

posted 10-16-2008 09:25 PM            

All should be warned!! If I have questions I may bug the heck out of you in my effort to grasp the concepts.

------------------
Ex scientia veritas

posted 10-17-2008 08:13 AM            

***********************************

I am now going to discuss the second type of question, on this phase of the examination.

Everyone, at some point in their lifetime, has said something in anger, that they later regretted. You have, haven’t you?

Everyone at one time or another, has told a lie, to keep themselves out of trouble. You have, haven’t you?

When I ask you these two questions during this test, I want you to think of a specific incident for each question, and then lie, and say, no. Take a few seconds, and think of a time, when you have said something in anger, that you later regretted. Do not tell me about it.

Now, think of a specific time in your life, when you have lied to keep yourself out of trouble. Do not tell me about it.

Let’s practice. Make sure you lie, and say, no.

Did you ever lie, to keep yourself out of trouble?

Did you ever say anything in anger, that you later regretted?

Do you now completely understand these two questions?

Do you now completely understand the instructions on how to answer each of these two questions?

As I explained earlier, when you lie, your body responds, and I will be able to see the response, just as I did during the practice test. If however, you were given a test, and I saw no responses to any of the questions, it would look like you were telling the truth. For various reasons such as illness and fatigue, some people lose their capability to respond. Consequently, some questions must be asked that demonstrate that you continue to have the capability to respond when you are lying, and that you do not respond when you are telling the truth.

When I ask you the questions, "Did you ever lie, to keep yourself out of trouble?" and "Did you ever say anything in anger, that you later regretted?" on the exam, what will you do?

Exactly. Think of the particular incident you are lying about, and make sure to lie and say "no" on the exam.

------------------
Ben

posted 10-24-2008 12:43 AM            

What are the adjustments that must be made on scoring rules for the pneumographs?

Thanks.....

posted 10-24-2008 10:17 AM            

Read more carefully, and look at the question sequence, and you'll notice that the RQs are presented three times during a single chart, which is then scored as if it were three pseudo-charts.

The basic principles of measurement are still in play: take several measurements to get a more stable estimate of the value of whatever you are measuring.

In fact, the requirements for the DLST are more stringent than for other techniques. Three interpretable (artifact free) presentations of each RQ are required in order to render an opinion. DACA/2006 handbook indicates only two interpretable presentations are required for other methods. Error rates are estimated using the addition-rule for dependent probability events. Meaning that the likelihood of an error for R1 is added to the likelihood of an error for R2. Two RQ exams are less likely to result in error than three or four RQ tests. Simple. R3 and R4 are then handled in a separate series. This is exactly what is wrong with our current Model Policy on Screening exams. When considering a test whitbh published and credible accuracy arguably around 90%, one could argue the error rates to be as high as 50% for an exam with 5 distinct issues.

So, if we are wondering why the TES performs so well as a screening test, then there are obvious places to look. More data = a more stable estimate of the value of interest. Fewer issues = fewer errors. The challenge, then, is to pick the issues that really matter. I suggest we draw target issues from the empirical studies on risk assessment and risk management.

Is it possible that other screening tests might perform better, (more accurate, fewer INCs) if three artifact free presentations were required for each RQ in order to render an opinion?

.012

r

------------------
"Gentlemen, you can't fight in here. This is the war room."
--(Stanley Kubrick/Peter Sellers - Dr. Strangelove, 1964)

posted 10-24-2008 06:34 PM            

What are the adjustments that must be made on the scoring rules for the pneumos?

By "artifact free" do you mean (only) a RQ that is without distortion? The DLST requires three RQ's that are free of distortion and capable of being interpreted?

posted 10-24-2008 08:21 PM            

What are the adjustments that must be made on the scoring rules for the pneumos?

By "artifact free" is it meant (only) a RQ that is without distortion? The DLST requires RQ's that are free of (any) distortion and capable of being interpreted?

Thanks.....

posted 10-24-2008 09:37 PM            

I understand "artifact-free" to mean interpretable or useable. It is admittedly vague, and could be strictly interpreted as requiring that a stimulus presentation would be useable/interpretable only when none of the data from any of the components are artifacted. This would be a punishing, unproductive, and perhaps unatainable requirement. If two of three compoenent segments for any stimulus presentation are artifacted, then that presentation is unusable. Three interpretable presentations are required for each stimulus.

Now take a moment and wonder about why the TES seems to perform so well as a screening test. It's structure and design does not allow for the interpretation of excessively marginal data.

If we want to improve the decision accuracy of the polygraph, then one thing to do is to require field examiners to gather a sufficient volume of good/interpretable data. Now think about the three-to-five chart rules with the Utah technique, and the fact that this feature seems to be responsible for the greatest portion of increased accuracy and reduced inconclusives, compared with the other features of the Utah.

Another feature of the TES, is that RQs are limited to two for each series. There is a mathematical advantage to this. Consider that there is an error rate for every polygraph test and every technique, whether that error rate is known or not. There is also an error rate for every individual stimulus question. If we imagine the error rate to be about 10% (given an estimated accuracy of the polygraph at 90%), and if we assume that 10% error rate might also apply to individual questions, then we would have to add 10% + 10% for each qeustion (according to the addition-rule for non-independent probability events). You can see that with more RQs the error rate climbs uncomfortably high. Of course, the error rate might not be as high as 10%, or it might be higher. The statistical principles remain the same. More distinct RQs, mean a greater overal rate of error. Separating the targets into separate series may be amoung the better solutions to this phenomena (thank Sheila Reed and the other researchers who seem to have already known about things like multiple comparisons and the addition rule).

Another distinct feature of the TES is that targets are defined by the technique, and it appears there is little need to monkey around with them. This way, targets are defined by agency policy, and risk-prediction/risk-management priorities, and not by any individual professional's value system or psychologized assumptions about what matters. It's a scientific test.

.012

r

------------------
"Gentlemen, you can't fight in here. This is the war room."
--(Stanley Kubrick/Peter Sellers - Dr. Strangelove, 1964)

posted 10-25-2008 01:16 AM            

I don't believe a strict interpretation of "artifact free" is reasonable. As we know, there can be an artifact (distortion) in one component that may NOT distort the tracings in remaining components of the same RQ, i.e., an examinee can move his finger(s) that distort the EDA and have no effect on remaining components.

I found the following information scanning Google & hitting on the Science Direct website titled:
"New measures improve the accuracy of the directed-lie test when detecting deception using a mock crime"
Brian G. Bell, a, , John C. Kircher and Paul C. Bernhardt

Department of Educational Psychology, University of Utah, Salt Lake City, UT 84112 USA

Received 23 June 2007; revised 18 January 2008; accepted 28 January 2008. Available online 8 February 2008.

A portion of the site abstract stated "There were no significant differences in decision accuracy between probable-lie and directed-lie tests, but respiration measures were more diagnostic for the probable-lie test." It was also reported " - a new respiratory measure improved the accuracy of the directed-lie test."

END.....

posted 10-25-2008 11:19 AM         

give me your email address, have some private questions for you if you don't mind.

posted 10-25-2008 04:33 PM            

I am posting this as I don't think John would mind. This will hopefully shed some light on the conclusions they drew from the DLCs.

Respiration Responses during DL tests.

We conducted exploratory analyses of second-by-second changes in respiration excursion associated with neutral, DL, and relevant questions for participants in the effective feedback conditions. Responses to neutral questions served as a baseline of normal respiratory activity. As compared to this baseline, innocent participants breathed more rapidly and/or deeply in response to DL questions. Innocent participants appeared deceptive, not because their respiratory activity in response to relevant questions was suppressed, but because their respiratory activity in response to DL questions increased. Ordinarily, when a person orients to a perceived threat, their breathing slows and becomes shallow (Lynn, 1966). The observed increase in respiratory activity shown by innocent participants in response to DL questions was not the type of orientation response that is typically observed during polygraph examinations (Kircher & Raskin, 1988; Timm, 1982).

During PL tests, innocent subjects are likely to be deceptive to probable-lie questions, and they want to avoid detection on those questions. Innocent subjects do what guilty subjects do when guilty subjects are asked relevant questions; innocent subjects attempt to avoid detection by inhibiting their physiological responses to the probable-lie questions.

In contrast, during DL tests, subjects are told that it is important that they appear deceptive on DL questions. The subject’s task is different. Rather than trying to avoid detection, subjects try to appear deceptive. Rather than trying to inhibit their reactions to comparison questions, subjects attempt to produce reactions to those questions. To accomplish this task, knowingly or not, innocent subjects alter the one physiological measure over which they have the greatest control, their respiration.

PL and DL questions place different demands on subjects, and respiration is sensitive to those differences. When subjects attempt to inhibit responses, their respiration is suppressed, and when they attempt to produce responses, their respiratory activity increases.

Interestingly, innocent subjects appear to use respiration to produce reactions to DL questions, but guilty subjects do not. If respiration suppression is a measure of inhibition, then guilty subjects use the same strategy whether they are given a PL or DL test. In both the PL and DL test, the responses by guilty subjects to relevant questions are more suppressed than their responses to comparison questions, and their responses to comparison questions are more suppressed than their responses to neutral questions. Guilty subjects have but one goal; they want to avoid detection. They attempt to do this by inhibiting their responses to both comparison and relevant questions. Thus, unlike innocent subjects, it does not appear that guilty subjects try to produce reactions to DL questions; they try to inhibit them.

Research by Gross and Levenson (1993) is consistent with the idea that respiration suppression is an indication of deliberate attempts by subjects to suppress the expression of emotional responses. They measured general somatic activity, respiration cycle-time, and respiration amplitude in participants who attempted to inhibit outward displays of negative emotion. Although they found no significant reduction in respiration rate or amplitude, they did find that attempts to suppress emotional displays were associated with reduced somatic activity. The latter finding is consistent with the idea that attempts to inhibit emotional responses are associated with general motor quieting, and this includes a reduction in respiratory activity. Their failure to observe effects on respiration directly may be due to their methods of measuring respiratory activity.

Prior research indicates that although measures of respiration rate and amplitude individually have low reliability and validity for the detection of deception, a composite measure of line length from the same data is highly reliable and diagnostic (Kircher & Raskin, 1988; Podlesny & Kircher, 1999).

The findings from the present study and those from Horowitz et al. (1997) argue that the procedures for evaluating respiration responses recorded during DL tests should be modified to avoid false positive errors.

Overall, numerical evaluations of respiration responses during DL tests were not diagnostic, and in the standard effective-feedback condition, numerical scores for respiration were almost significantly more negative for innocent subjects than for guilty subjects. Obviously, it would be better to drop the evaluations of respiration responses altogether than to include them if they work against a valid decision.

There were no significant differences between PL and DL tests in the accuracy of polygraph outcomes. However, the respiration responses of innocent subjects to directed-lie questions were unlike those of innocent subjects to probable-lie questions. When innocent subjects were presented with directed-lie questions, their respiratory activity increased. These findings suggest that numerical scoring rules and computer algorithms for probable-lie tests are not optimal for DL tests. Specifically, subjects should not be considered deceptive if they show greater suppression in response to relevant questions than to directed-lie questions. In light of these findings, we recommend that respiration responses not be numerically evaluated or computer scored until it can be established with verified field polygraph examinations that those methods are appropriate for DL tests.

Hope this helps,

posted 10-25-2008 07:36 PM            

That was perfect.

For these reasons. DLC exam types in the Limestone system will set the results for all standardized logged R/C ratios to zero for pneumo measurements. It is still possible to mark artifacted or CM data at the pneumos where necessary, and the Test of Proportions will still capture that data.

Mark and I have recently spoken with the programmers at Lafayette who are working on a DLST solution within the Lafayette OSS-3 tool.

There is more to look forward to.

At some point we'll have to develop a canonlical discriminate function sans-pneumos, and optimize the component weighting for that model.

r

------------------
"Gentlemen, you can't fight in here. This is the war room."
--(Stanley Kubrick/Peter Sellers - Dr. Strangelove, 1964)

posted 10-26-2008 02:35 PM            

If pneumo scores are "significantly more negative for innocent subjects than for guilty" during DL tests, why is the DLST being used at this time if it is suspected they work against a valid decision? If the studies "suggest that numerical scoring rules" are "not optimal for DL tests," especially as it relates to respiration, shouldn't they be discontinued until further research confirms validity relative to respiration? How do we "score" or evaluate pneumos to render an opinion in DSLT's if it is (recommended) they "not be numerically evaluated or computer scored."

As indicated in the Journal article and the '97 Horowitz report, what are the "adjustments" or "modifications" that must be made on the scoring rules for the pneumos? Is it only that we don't consider deception if there is greater respiratory response to a RQ than to a DLQ?

posted 10-26-2008 03:40 PM            

When hand-scoring, some people only assign scores of ones to all channels, which means they have weighted all channels equally, and none of our algorithms do that (or they wouldn't work that well). The study suggested pneumos not get weighted that heavily, which is the advantage of seven-position scoring (in some sense, anyhow).

I don't think this study said whether removing the pneumo scores increased or decreased (or did nothing to) correct decisions, and that would be interesting to know. They do suggest not using them at all, which would imply they don't help enough to bother. (Keep in mind this is one data set. I've seen some that show cardio is a waste of chart paper.)

We do know DLCQs look different, and in the few studies we have, they are still scored like PLCQs (for the most part). Together, those studies show such tests work, and in screening situations, work well.

I still haven't had a chance to dig any out to show you all though.

posted 10-27-2008 03:35 PM            

A study is referenced that suggests not using the pneumos at all; and as you state, this implies "they don't help enough to bother." How do we meet the minimum three component requirement to administer an exam if we eliminate the pneumos or choose not to use the pneumo tracings during an evaluation? I don't know specifically what you mean as it relates to polygraph when you refer to PO2.

My concern is using a procedure that appears to have too many red flags right now. As a consequence I would not be comfortable and confident using the DLST at this time. I look forward to following the research and remain interested in learning what adjustments or modifications have been suggested when scoring the pneumo tracings produced during a DLST.

END.....

posted 10-27-2008 04:56 PM            

No-one is saying you have to use the DLST. If you don't like it, don't use it. It doesn't cost you anything.

That said, the studies on the TES have shown that it does outperform other screening techniques.

Even if you don't use it, it might still make sense to learn about it. There are evidently some good things about it. When we sift through the fog of strong opinions and fancy ideas, we can eventually start taking note of the common features of the better techniques. Somewhere in there, is an opportunity to wittle away at some more of the procedural chaff and improve the polygraph test by focusing on what actually matters. That is, unless we think that polygraph techniques rae good enough at present, or are as good as they can ever be. There are some not-so-dumb people who would like us to do nothing, and continue to engage in arcane procedures that are validited only the fact that someone said so (the pied-piper system). That will be a vulnerable long-term posture to maintain.

.012

r

------------------
"Gentlemen, you can't fight in here. This is the war room."
--(Stanley Kubrick/Peter Sellers - Dr. Strangelove, 1964)

posted 10-27-2008 07:42 PM            

Since it's reported in the current (Polygraph) article that a pneumo for any DLC technique can't be analyzed using the same criteria as are used for PLC testing - what are the scoring rules referred to that must be adjusted (or modified)? I've asked this question a couple times now and I don't believe it has been answered.

Have my questions been out of line or unreasonable for anyone interested in learning about the DLST?

posted 10-27-2008 08:44 PM            

As I have already stated, researchers have known for some time that in any given study one channel may not tell you much of anything, so you're getting hung up on something that's not all that justified. (After all, when the pneumos were scored, the test still worked well.) I have a data set that works best if you remove both the pneumos and the cardio, but I'm not ready to toss them.

Do you know if your pneumo scores are helpful to your decisions in any test? Unless you score tests in which ground truth is known, and then you analyze all of your scores (statistically), then even though we know the pneumos to be diagnostic (in some tests), they may tell you nothing - or they may actually result in scores in the wrong direction.

Here's the bottom line: DLCQTs have the highest accuracy for screening exams. Why? We're still figuring that one out, but it's probably not the discriminatory power of the pneumos with traditional scoring criteria. (In other words, the test will outperform, on average, the PLCQTs (screening) we have subjected to analysis. You can get hung up on the pneumos, but you'll have a hard time justifying not using them when you look at the overall accuracy rates.

posted 10-27-2008 09:11 PM            

Sorry. Your question is not unreasonable. We are all still learning about the DLST. I thought someone answered your questions, but it was probably not that clear. None of us really use this test yet, but some information is available, and the studies suggest it is quite defensible. Maybe more defensible than other techniques over time.

It's not a perfect solution, but you'll probably be fine if you zero all the pneumos.

Ben has been using the technique. Perhaps he can provide more information.

Bottom line is, following the science is good idea. Scientific studies indicated DLC work just fine, and the TES is a strong performer in its intended use. It outperforms other screening techniques. Several studies tell us that.

Following the science may feel uncertain at times, especially if we are used to being sold confidence (not science) from existing techniques. There is an important difference here. Its tempting to just do what we've always done, and pretent it's great or perfect, or good enough. As long as we don't look at the data, and how uncertain those techniques actually are, then it will be tempting to fiegn some form of false confidence.

Science will always have some portion of uncertainty to it. If we are really desperate to reduce the anxiety that results from uncertainty, then someone will come alone selling confidence - but then we have to surrender our independent and intelligent judgement along with the anxiety of uncertainty. The more desperate we are for confidence, the more vulnerable we are to behaving like lemmings. The only and best way to actually reduce our uncertainty is to measure carefully, do the math, and follow the numbers.

It's the same idea as saying - "trust your charts." Except that with the DLST and other DLC techniques we are faced with having to learn a new way to look at the data.

Sure the polygraph is a good test, but it could be better. Do we think its going to improve by not following the science. Do we think we'll come up with viable solutions by guessing according to our own fancy ideas - which are usually just an example of our personal belief system or value system? Keep in mind that most fancy ideas (hypotheses) - even the good sounding ones - turn out to be nothing more than fancy ideas.

Lets follow the science.

.012

r

------------------
"Gentlemen, you can't fight in here. This is the war room."
--(Stanley Kubrick/Peter Sellers - Dr. Strangelove, 1964)

posted 10-29-2008 08:40 AM            

"It's not a perfect solution, but you'll probably be fine if you zero all the pneumos."

Please tell me I am not understanding your comment as intended. How can any examiner "probably be fine if you zero all the pneumos?" Are you suggesting an examiner "zero all the pneumos" (before) any scoring of the charts? Isn't this the same as an examiner using only an EDA and cardio component to administer an examination? I suspect this is not what you mean when you state "we are faced with having to learn a new way to look at the data?"

I have no problem "following the science." My uncertainty is not simply about dealing with a new technique but my lack of confidence in the DLST for the reasons I have identified since the beginning of this topic. Uncertainty not because it is new, but due to the fact (research) has identified problems affecting the pneumos as they relate to this test.

The article you co-authored in the current issue of "Polygraph" states a pneumo for a DLC technique can't be analyzed using the same criteria as are used for PLC testing; as a consequence "scoring rules must be adjusted." What are the scoring rules referred to in this article that must be adjusted?

END.....

posted 10-29-2008 09:33 AM            

You seem to be quite troubled by this DLST concept and how the respiration channel is to be interpreted during evaluation of directed lie comparison questions. Ray mentioned you would be well advised to essentially "zero" all of the pneumos. Barry has provided you will ample clarification. May I refer you to the Bell et al. (Bell, Raskin, Honts & Kircher 1999) study in which one of the things they reported was the respiration channel was scored a zero about 80% of the time in the Utah approach.

May I ask you to share with us what technique you are presently using in screening situations? Perhaps we can find some clarifying published research to assist you in understanding this concept of conservatism in the respiration channel.

posted 10-29-2008 10:39 AM            

I just finished pulling together all of my Polygraph articles dealing with DLCQs and the last was Paul's article which had the best description of dealing with the respiration channel.

posted 10-29-2008 11:31 AM            

Yes, I am “troubled” relative to the interpretation of the pneumos when using the DLST as it is being suggested we “zero” this component!

As you are a co-author of the “Polygraph” article, I would appreciate an answer to a few questions, a couple of which I asked during my 8:40am post to Ray.

1. Are you suggesting an examiner “zero all the pneumos” (before) any scoring of the charts? If you do this, why record (any) pneumo responses? Is it only due to the APA By-Laws (3.5.1.1) this recording is being done? As for scoring, as you know, specific types of scoring is only required for evidentiary/specific issue examinations (3.10.1).

2. (If) zeroing is what is being suggested, isn’t this the same as an examiner using only an EDA and cardio component to administer an examination?

3. What are the scoring rules referred to in this article that must be adjusted?

I do not conduct any screening examinations at this time. I liked the structure of the DLST right up until the time I encountered this pneumo issue.

posted 10-29-2008 01:26 PM            

1. Are you suggesting an examiner “zero all the pneumos” (before) any scoring of the charts?

No I do not suggest this.

In the Kircher et al. (2001) work they found In both the PL and DL test, the responses by guilty subjects to relevant questions are more suppressed than their responses to comparison questions, and their responses to comparison questions are more suppressed than their responses to neutral questions. Conversely they found when innocent subjects were presented with directed-lie questions, their respiratory activity increased which suggests that numerical scoring rules and computer algorithms for probable-lie tests are not optimal for DL tests. If I see strong suppression in response to a relevant question, I would score it. I would not expect to see strong suppression to a DLCQ. While the data are limited, this taken as a whole suggests if we were simply measuring the line length, we can anticipate a truthful subject will produce a response that is greater in length at the DLCQ than at the RQ.

If you do this, why record (any) pneumo responses? Is it only due to the APA By-Laws (3.5.1.1) this recording is being done? As for scoring, as you know, specific types of scoring is only required for evidentiary/specific issue examinations (3.10.1).

The respiration channel is arguably the one that can be most easily manipulated by the test subject. You may have heard the term "effector" along the way as a term used to describe the respiration channel because it can so easily "effect" the other waveforms. Even in psycho-phys labs where they are interested in collecting EDA, blood pressure and/or blood-pulse volume they will often monitor the respiration movement of the test subject to ensure any observations were not influenced by movement associated with respiratory activity.

2. (If) zeroing is what is being suggested, isn’t this the same as an examiner using only an EDA and cardio component to administer an examination?

I do not believe this is what Ray was suggesting you do as a practice. I think he was offering you a solution that would above all do no harm. Ray and I have studied a lot of data and have spread sheets created through brute force computer techniques that allow us to assume the data are relatively normal. We have looked at what happens when you simply remove the respiration component from the scoring and we found we lose sensitivity and specificity. I realize there exists a chance to lose these (sensitivity and specificity) by zeroing the respiration channel but if you were uncomfortable with scoring only strong suppression at the RQ then this is a possible solution.

If you were to “zero” the respiration channel, that would leave the EDA, relative blood pressure and blood-pulse volume channels (PLE). Don Krapohl and I addressed the relative contribution of the PLE in a paper we wrote last year. We reviewed the Kircher & Raskin 1988 data and found the PLE contributed as much to total score as the relative BP channel.

3. What are the scoring rules referred to in this article that must be adjusted?

Please see above.

I do not conduct any screening examinations at this time. I liked the structure of the DLST right up until the time I encountered this pneumo issue.

Then it would probably not behoove or benefit you to change from what you are doing. You are obviously well read and are likely using some technique for event-specific testing that is already supported by some body of research.

Honts & Alloway (2007) found a TES approach did not work as well as other approaches in an event-specific setting. (Again this is just one project.) I created the table below for you to assess their results when applying this technique to an event-specific mock crime.

Table 2. Decisions of truth and deception
____________________________________________________________________
Decision
Information

DI NDI
___________________________________________________________________
Informed guilt
Guilty 8 2
Innocent 3 7

Naive guilt Guilty 8 2
Innocent 4 6

posted 10-29-2008 01:28 PM            

Q: Are you suggesting an examiner “zero all the pneumos”
A: No. When MI examiners first began using the test, it was scored globally. Numerical scoring was instituted in the early 80s and then a 5 position scale (+2 -2) was taught for scoring the test. If pneumographic tracings at the DLCs looked "too good to be true" or extremely exaggerated, little weight was given to them unless the RC respiratory patterns showed absolutely no response.

Q: (If) zeroing is what is being suggested, isn’t this the same as an examiner using only an EDA and cardio component to administer an examination?
A: No, see above answer. They are evaluated but they are looked at differently than with PLCs.

Q: What are the scoring rules referred to in this article that must be adjusted?

A: As previously stated, a 7 position (=/-3) was not used. A two position (+/-2) was used. Additionally, they had a rule that if two of the three presentations of any relevant question spots did not garner a positive score out of 3 askings then the test was inconclusive regardless of spot totals.

The bottom line here is that the DLC screening test as conducted by MI back in the 80s is a totally different test than simply replacing the PLCs with DLCs both in scoring and the pre-test setup. With that said, the research done has shown that DLCs and PLCS treated the same when scored on specific issue tests work equally well.

posted 10-30-2008 05:01 PM            

------------------
Ben

posted 11-02-2008 12:29 PM            

1. DLST is derived from the TES format
2. The TES format appears to be highly accurate for a screening exam.
3. Blalock provided a workable DLQ introduction Script
4. The testing format is clearly outlined by Blalock.
5. While presented as a “Directed Lie” test there is no articulable reason to believe that using “Probable Lie” Comparison Questions would negatively impact its accuracy but that a study indicates that pneumo tracings appear to be more diagnostic in a PLC.
6. The testing format lends itself to screening issues as well as research because it allows for a more consistent or objective construction of relevant questions. Id est, we can ask everyone basically the same questions.
7. A seven position scoring system may prove to be better way to score this exam as a means to avoid over-reliance upon what appears to be a marginal diagnostic capability in the pneumo channel.
8. Skip is obviously enjoying whatever access he gets to the APA Library. I’m putting their back issues CD n my Christmas list. Does anyone know if the CD is indexed?
9. Conscientious Polygraph Examiners can and often do disagree on almost anything except that there is a need to continue to improve both the science and professionalism that encompasses our profession.

------------------
Ex scientia veritas

posted 11-02-2008 01:37 PM            

I wouldn't say the pneumos aren't diagnostic. They may be more diagnostic - we just haven't figured out how to get them to tell us what they mean (if they mean anything). Then again, they may not be diagnostic. I score DLCQTs when I run screening exams. (I usually use the Utah MGQT.) Sometimes I see the ugly pneumos we've discussed here; other times you wouldn't be able to tell the difference. I suspect RLL would correlate with the other channels, but it would be a weaker correlation than say the PO2 and Cardio, but we already knew that to be the case in PLCQTs, didn't we?

posted 11-03-2008 10:31 AM            

I just feel a little uneasy when we get too euphemistic about the real value (or lack thereof) of pneumographic scoring. One could say the pneumos are polygraph"s achilles heal----only the achilles heal is actually an essential part of moving forward. Perhaps the pneumos are more like the tonsils.

posted 11-03-2008 03:07 PM            

1. DLST is derived from the TES format

Actually, the DLC pre-dates the TES by many years. Since the early 60's shortly after Backster introduced PLCs, Military Intelligence examiners were using this technique for screening and periodic testing years before the TES (and DODPI/DACA) even existed.

6. The testing format lends itself to screening issues as well as research because it allows for a more consistent or objective construction of relevant questions. Id est, we can ask everyone basically the same questions.

The real value perceived by MI examiners at the time it was developed was the ability to repeatedly test examinees who were being given frequent polygraph tests without the need to continually "sell" them on the PLCs.

8. Skip is obviously enjoying whatever access he gets to the APA Library. I’m putting their back issues CD n my Christmas list. Does anyone know if the CD is indexed?