posted 05-26-2009 09:04 AM            

Very thorough material can be downloaded from here:
http://www.internationalpolygraph.com/blalock.htm

for specific questions, just call me directly at 239.357.2219 and I will walk you through setting it all up.

posted 05-26-2009 09:08 AM            

Last updated on March 23, 2009

1. SCOPE
This guide establishes essential elements for the conduct of the DLST question format for use in screening.

2. BACKGROUND
The DLST was designed following the format established for the Test for Espionage and Sabotage (TES). The TES was designed and researched by the Department of Defense Polygraph Institute Research Division Staff.

3. PRETEST PHASE
3.1. Question Review. During the pretest interview, only the questions for the sub-test about to be tested are reviewed with the examinee prior to the collection of charts. (See figure Fl.)

Figure Fl. Question Review Sequence for the DLST
Relevant (Rl & R2, or R3 & R4, or R5 & R6 if used)
Sacrifice Relevant (SR)
Directed-Lie Comparison (Cl & C2)
Neutral (N1 & N2)

3.2. Directed Lie Screening Test Questions

3.2.1. Relevant. This question pertains directly to the issue(s) being tested. The relevant question tests the possible direct involvement of the examinee.

3.2.2. Directed-Lie Comparison. The DLC question is a specialized comparison question. A properly constructed DLC question involves a minor transgression which should have some personal significance to the examinee. Upon acknowledging having committed such a transgression, the examinee is directed to lie when asked that particular question on the test. The question is separated from the relevant issue by category. (see figure F 2.)

Figure F2. Sample Directed-Lie Comparison Questions for the DLST
Did you ever commit a minor traffic violation?
Did you ever say something that you later regretted?
Did you ever lie to keep yourself out of trouble?

3.2.3. Sacrifice Relevant. This is the first question of the DLST format that refers to the relevant issues, and it prepares the examinee for the introduction of the relevant questions. (See figure F3.)

Figure F3. Samples of a Sacrifice Relevant Question for the DLST
Do you intend to answer the job history questions truthfully?
Regarding the job history questions, do you intend to answer truthfully?

3.2.4. Neutral Question. A neutral question is the first question asked during the data collection phase. It may also be asked in other positions on the chart. It is designed to allow the orienting response to habituate before a scoreable question is asked. It can be used to allow the tracing to stabilize prior to presenting the next stimulus. It should be unrelated to the issue being tested. Neutral questions are not scored. Several neutral questions may be reviewed and used as needed. (See figure F4.)

Figure F4. Samples of Neutral Questions
Are you now in Florida?
Are you now sitting down?

4. DATA COLLECTION

4.1. Sub-Tests. The DLST (based upon the TES format) was originally designed to test four relevant questions. The format can be expanded to test six relevant questions. The DLST consists of two sub-tests (A and B) when four relevant questions are tested, and three sub-tests (A, B, and C) when six relevant questions are tested.

4.2. Question Sequence. Sub-Tests A, B, and C are sequenced the same. Relevant questions #1 and #2 are asked on Sub-Test A. Relevant questions #3 and #4 are asked on Sub- Test B, and relevant questions #5 and #6 are asked on Sub-Test C. If desired, the DLC and neutral questions may be different for each sub-test. (See figure F5.)

Figure F5. Question Sequence for Sub-Test A for the DLST

Neutral #1
Neutral #2
Sacrifice relevant
First presentation of DLC #1
First presentation of relevant #1
First presentation of relevant #2
First presentation of DLC #2
Second presentation of relevant #1
Second presentation of relevant #2
Second presentation of DLC #1
Third presentation of relevant #1
Third presentation of relevant #2
Second presentation of DLC #2

4.3. Question Format Procedures. Test procedures require a minimum of three presentations of each relevant question. The physiological tracings should be free of artifacts in at least two of the three component tracings at each presentation of the relevant questions and able to be evaluated against at least one of the bracketing comparison questions. If after three presentations of the relevant questions artifacts prevent scoring of one or both of the relevant questions, an additional presentation of both relevant questions for that sub-test will be conducted. One of the following procedures will be used to accomplish this:

4.3.1. If the examiner realizes that a fourth presentation is necessary prior to the completion of the chart, a fourth presentation may be collected on the same chart. Three questions are added after 2C2. (See figure F6.)

Figure F6. Extended Question Sequence for Sub-Test A for the DLST
Neutral #1
Neutral #2
Sacrifice relevant
First presentation of DLC #1
First presentation of relevant #1
First presentation of relevant #2
First presentation of DLC #2
Second presentation of relevant #1
Second presentation of relevant #2
Second presentation of DLC #1
Third presentation of relevant #1
Third presentation of relevant #2
Second presentation of DLC #2
Fourth presentation of relevant #1
Fourth presentation of relevant #2
Third presentation of DLC #1

4.3.2. If the examiner does not realize that a fourth presentation is necessary until the chart has been collected, a short chart may be collected. (See figure F7.)

Figure F7. DLST Short Chart
I1 Neutral #1
12 Neutral #2
Sac Sacrifice relevant
3C1 Third presentation of DLC #1
4R1 Fourth presentation of relevant #1
4R2 Fourth presentation of relevant #2
3C2 Third presentation of DLC #2

4.3.3. Any of the sub-tests may be administered first.

4.3.4. When a sub-test (i.e., DLSTA) results in a decision of SR, a post-test interview will be conducted, but not until all sub-tests have been administered, to determine if there is a reasonable explanation for the responses. A repetition of the sub-test may be conducted whether or not the examinee provides a reasonable explanation for the responses.

4.3.5. When a sub-test (i.e., DLSTA) results in a decision of NO, explain that the subject failed to produce sufficient responses, review the relevant and directed lie comparison questions, and administer the entire sub-test again.

4.3.6. When a sub-test (i.e., DLSTA) results in a decision of NSR, move on to the next sub-test, if applicable.

4.3.7. After the completion of a sub-test, the questions for the next sub-test are introduced and tested.

4.4. Question Interval. Question intervals from stimulus onset to stimulus onset should not be less than 20 seconds.

4.5. Conduct of an Acquaintance Test. The ACQT will be collected as the first chart of the DLST examination. It is conducted in part to acquaint the examinee with PDD procedures. The known-solution ACQT is the only ACQT used by this agency.

5. TEST DATA ANALYSIS

5.1. Numerical Analysis. DLST charts are analyzed using either the 3-position or the 7-position scale.

5.2. No Split Opinions. An examinee is SR, NSR, or NO to the sub-test, not to a question.

5.3. Administrative Opinions. These opinions reflect the results of a series or an examination that are not based upon physiological responses to the applied stimuli, e.g., when the examinee terminates an examination or is practicing countermeasures. In these instances, administrative opinions such as inconclusive, purposeful non-cooperation, etc., are appropriate.

5.4. Spot Analysis. The DLST scoring system analyzes the physiological responses recorded on the chart by comparing the relevant questions to the largest response by component tracing of the DLC questions. (See figure F8.)

Figure F8. Spot Analysis for the DLST
1R1 and 1R2 compared to 1C1 or 1C2
2R1 and 2R2 compared to 1C2 or 2C1
3R1 and 3R2 compared to 2C1 or 2C2

When a short test is required:

4R1 and 4R2 compared to 3C1 or 3C2

When a fourth presentation is required:

4R1 and 4R2 compared to 2C2 or 3C1

5. 5. Test Data Analysis of a Fourth Presentation. When a fourth presentation of the relevant questions is necessary, only the relevant question(s) that could not be evaluated due to artifacts will be scored.

6. OPINION RENDERING CRITERIA

6.1. Significant Reactions. A minus three (-3) or less in any overall vertical spot, or an overall horizontal total score of minus four (-4).

6.2. No Significant Reactions. At least a plus one (+1) at each overall vertical spot and an overall horizontal total score of at least plus four (+4).

6.3. No Opinion. If it is not SR or NSR, it is NO with the exception of administrative opinions.

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

posted 05-27-2009 09:04 AM            

Excellent job. I teach part time at an international program and I've been trying to get them to introduce this format into their training schedule. This info will help my cause. Thanks

posted 06-01-2009 07:32 PM            

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

posted 06-02-2009 08:37 AM            

question:

posted 06-02-2009 08:04 PM            

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

posted 06-27-2009 06:57 AM            

p.s. Sorry for not so good English, yet.

posted 06-29-2009 08:37 AM            

quote:

---

TES format only uses two RQs, and that is probably one of the reasons for its high accuracy. The beauty is that you run them both three times in only one chart. That means you can run six RQs over three charts, and your accuracy is going to be much better than the coin-flip accuracies you'd get running six RQs in a single test (over three charts)

---

posted 06-29-2009 12:00 PM            

It makes no sense to ignore the improvements of the TES/DLST and handicap it with the known problems of traditional decision policies, when the TES decisions rules have been shown to work better.

Start changing things from a tightly controlled and well structured technique and you cannot be quite sure what you've got.

If your gonna invent a new screening technique, you ought to take the time to do it thoughtfully and thoroughly. Validate it with multiple studies. Publish it. You get the picture. Then, your new technique, to provide an real advantage, should perform as well or better than the best technques that are presently available - including the TES/DLST.

Remember - If it ain't broke, why fix it.

Not sayting the TES/DLST is perfect. But if we are going to remediate some deficiency, we have an obligation to quantify the deficiency and quantify the improvement. Otherwise we are doing our profession and our communities a disservice. Major modifications of a good technique should be based on a lot more than our expertisizing from the field.(I just made up a new word - expertizing: imparting great value to our expert opinions and anecdotable experience, when we haven't a shread of scientific data, while knowing that many hypothesis and fancy-ideas don't seem to stand up to empirical scrutiny).

For example, DLCs have been shown to work in investigative/diagnostic exams such as the Utah Zone. However, I believe there was a study that showed the TES (which is a screening technque) performed weakly in a diagnostic setting. Ben, do you have this?

It is important to be a little careful with changes like the ones you are suggesting.

.02

r

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

posted 06-29-2009 12:01 PM            

#1. There have been no published studies (that I am aware of) using probable lie comparisons with the DLST (TES). While it may sound like a good idea, I would not recommend running it with PLCs, since we don’t have those studies to demonstrate that it does work. I personally believe that we should be able to use either PLCs or DLCs interchangeably, regardless of formats, but that is only an opinion. Since there are no studies using PLCs in this format, it would be easier to defend if we used DLCs, since that was what was used in the studies. An 88% accuracy rate in a multiple-issue technique is pretty impressive, especially compared to the multiple-issue technique alternatives.

#2 Whether or not we lose some accuracy without discussion of DLCs between charts on a DLC type exam is an interesting question, but has not been studied with the DLST (TES) format. Nevertheless, considering the high accuracy rate of this format, the lack of between chart discussion does not appear to have a significantly negative effect in this format.

posted 06-29-2009 01:48 PM            

Donald J.Krapohl said
"Valid principles, it can be agreed, are the building blocks of valid techniques.
Therefore, if one knows which principles are valid and which are invalid, development of valid techniques is a straightforward process."

PLCs have more validation studies than DLCs, therefore a test is similar to TES with PLCs seems to be more valid than with DLCs. Of course, only validation studies can finally resolve this issue.

How do you think, do we need validation studies of the TES with PLC questions or the DLST is nothing better?

posted 06-29-2009 03:26 PM            

We examiners have been tweaking our stuff for a long time.

Ben is correct about the value of emphasizing validated principles, and not becoming overly rigid about silly things.

On the other hand, if you have a pristine '76 BMW 2002 Tii, why ruin it with aftermarket wheels and a loud pipe? Its a nice ride, but not gonna win any races against a modern station wagon or economy car. Enjoy it for what it is - a great piece of engineering. (OK, bad example.)

So here is my idea for a polygraph reality TV show...

Pimp-my-Polygraph-Technique.

Just start with any basic Zone or MGQT, then can add special questions, exotic or retro decision rules, and all kinds of fancy schtick - and have a good time doing it.

Anyone know a producer and some commercial sponsors?

.02

r

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

posted 06-29-2009 05:04 PM            

The TES/DLST is the current race winner.

The BMW Tii was a bad example - I just happen to like those old machines.

I've now fallen out of love with my iX, after a thouroughly miserable experience removing the transmission and transfer case to replace a clutch. Rear differential had to be removed just to get the drive shafts out. The clutch was fine, but the release bearing had suddenly fallen apart at about just under 300K miles. Along the way, I broke an exhaust manifold stud, and a front CV axle boot. So, decided to replace the axle.

Nick pretending to do something important.

.

Depending on which neighborhood, an altitude, in which you live, traction can sometimes be more important than speed. That front axle has come in really hand in the snow. On the other hand, the spirited little iX doesn't serve my need to haul my catapult to the hills for practice at tossing things. For that a nice old ugly truck, is much better.

For me to s'pose anything about the use of PLCs in the TES/DLST would be to engage in a bit of expertizing (not that I've never done that), because we don't really know, because it hasn't been studied that way. What we do know is that it seems to work very well when we don't pimp it out to our individual liking. Why take a perfectly good technique, and administer in in an unproven way?

The DLC is part of the TES. Part of the DLC is that is does not require the same degree of manipulating and confusing the examinee. Change that and you change the overall test experience for the examinee.

If you want a PLC screening technique then chose any of the good modern versions of the venerable MGQT. Of course, you can't expect TES performance from an MGQT, any more than you'd enjoy hauling a catapult with an old BMW. You could do it, but it might be Jed Clampett moment.

Which gets me wondering, if you were stuck on a desert island with only one or two polygraph techniques, which ones would they be?

r

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

posted 06-29-2009 09:13 PM            

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

posted 06-30-2009 03:54 AM            

quote:

---

The PLCQT screening exam studies show accuracies about the same as you'd get with coin flipping

---

I suppose it is more correct to compare comparable things, for example, tests with two RQs.
Are there screening exam studies show accuracy, for example, of the Air Force MQGT test with two RQs (I SR C R R C), PLCs and three charts?

posted 06-30-2009 08:06 PM            

quote:

---

I suppose it is more correct to compare comparable things, for example, tests with two RQs.

---

Maybe, maybe not.

posted 07-01-2009 02:06 PM            

Let's face it, we're already above the median accuracy for the single issue test as discussed in the NAS report.

posted 07-12-2009 07:38 AM            

"Probable DLST" is an oxymoron.

In addition, by using the format you propose, I would be concerned that by repeating the CQ's and not repeating the RQ's that any CQ reaction would be more attributed to a format induced artifact than psychological set.

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

posted 07-12-2009 08:21 AM            

The format has 3 presentations of 1 relevant issue resolved in a single chart.
Sub-test A
Neutral #1
Sacrifice relevant
First presentation of DLC #1
First presentation of relevant #1
First presentation of DLC #2
Second presentation of relevant #1
Second presentation of DLC #1
Third presentation of relevant #1
Second presentation of DLC #2

Sub-test B
Neutral #1
Sacrifice relevant
First presentation of DLC #1 (3)
First presentation of relevant #2
First presentation of DLC #2 (4)
Second presentation of relevant #2
Second presentation of DLC #1 (3)
Third presentation of relevant #2
Second presentation of DLC #2 (4)

For 4 issues it must be 4 sub-tests.

The possible advantages:
1. More accuracy, power of the single-issue test (?)
2. We exactly know what the issue is SR, it is useful for post-test, follow-up breakdown, specific-issue tests (?)
3. ?
...

The possible disadvantages:
1. Need more sub-tests than in the DLST
2. Need more DLCs for several sub-tests (?)
3. ?
...

What else?

Are there another advantages of usage of two RQs (not one, for example) in the DLST, besides a performance?

Ebvan, the probable "DLST" is not the "Probable DLST", it means the probable, possible screening test with DLCs, because we have not it with one RQ now.

posted 07-12-2009 05:06 PM            

What you propose would be 4 specific issue tests rather than a screening test.

I have no reason to believe that it wouldn't work, but if you run the DLST exam as designed and your examinee does not react to any of the relevant questions then the test is over after 2 charts.

You propose a minimum of 4 charts regardless of the outcome of any 1 chart.

For pre-employment, using the DLST, as designed, simply takes less time to test a truthful examinee.

For the most part, in the United States Pre-employment testing is generally reserved for Law Enforcement, Intelligence, and Security Clearance applicants.

These applicants usually have been through several elimination processes and background investigations before being scheduled for polygraph. While there are exceptions that require special attention, most applicants have little in their backgrounds that might disqualify them for employment.

The DLST allows quick and efficient testing of candidates.

All that being said, the greatest potential benefit for using the DLST is standardizing a screening technique for use thoughout the profession that is applicable for pre-employment screening, post-conviction sex offender testing, and continued employment security screening. While it may not be the best possible test for every situation, it is easy to learn, easy to interpret, and applicable to a majority of screening situations as "The first test"

In closing, what you propose might work as well or even better than the DLST. My only real issue with what you propose is that it moves us farther away from a standarized technique which I see as one of it's primary benefits.

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

posted 07-13-2009 03:26 AM            

There is only one difference between the DLST and what I propose. My test has only the one RQ, because of that it needs more sub-tests.
There is one issue per one sub-test. If we have two issues for the screening examination it will be two sub-tests, if we have four issues it will be four sub-tests and so on.
Any of the sub-tests may be administered first.
For example:
Sub-test A: I-1C1-1R1-1C2-2R1-2C1-3R1-2C2
Sub-test B: I-1C3-1R2-1C4-2R2-2C3-3R2-2C4
Sub-test C: I-1C1-1R3-1C2-2R3-2C1-3R3-2C2
Sub-test D: I-1C3-1R4-1C4-2R4-2C3-3R4-2C4
...

The difference between the "single-issue test" (for example Utah ZCT) and what I propose is that my test requires the only one chart when a sub-test (i.e., Sub-test A) results in a decision of NSR. When a sub-test (i.e., Sub-test A) results in a decision of SR a repetition of the sub-test may be conducted. When a sub-test (i.e., Sub-test A) results in a decision of NO, entire sub-test must be conducted again.
Test procedures require a minimum of three presentations of one RQ. If after three presentations of the RQ artifacts prevent scoring of the RQ, an additional presentation of the RQ for that test will be conducted.
If the examiner realizes that a fourth presentation is necessary prior to the completion of the chart, a fourth presentation may be collected on the same chart:
I-1C1-1R1-1C2-2R1-2C1-3R1-2C2-4R1-2C1
If the examiner does not realize that a fourth presentation is necessary until the chart has been collected, a short chart may be collected:
I-3C1-4R1-3C2.

If we have SR in the DLST we do not know exactly the question to which an examinee is lying. In my test we do know this question and it is good for post-test, follow-up tests (breakdown tests).

The DLST is a standarized technique, I definitely think it is a good advantage.

Thanks.

posted 07-13-2009 04:10 AM            

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

posted 07-13-2009 05:00 AM            

quote:

---

4.3.4. When a sub-test (i.e., DLSTA) results in a decision of SR, a post-test interview will be conducted, but not until all sub-tests have been administered, to determine if there is a reasonable explanation for the responses. A repetition of the sub-test may be conducted whether or not the examinee provides a reasonable explanation for the responses.

---

Someone please explain to me, if we have SR in the Sub-test A of the DLST, is it recommended to do post-test and to repeat the sub-test A immediately after the first presentation of the sub-test A, or we must to conduct Sub-test B and/or Sub-test C before the post-test and the repetition of the sub-test A. If it is the first variant and we have SR in the Sub-test A after one or two repetitions of the sub-test A, is it recomended to conduct the next sub-test B or sub-test C or the examination must be terminated with SR decision in the Sub-test A? Are this instructions in the DLST like instructions in the TES or another?

Thanks.

posted 07-13-2009 11:10 AM            

quote:

---

What you propose would be 4 specific issue tests rather than a screening test.

---

Not quite.

It is not the number of issues that defines a screening test. It is the presence or absence of a known problem that differentiates screening from diagnostic tests.

When there is a known problem (allegation/incident/symptom), along with reason to suspect the person's involvement, it is a diagnostic test.

When there is no known problem (fishing trip), it is a screening test - regardless of the number of issues. Some screening tests are single-issue screening tests (rubber-glove prostate exam for example: there'd better be no other reason to do that...)

We have to be careful not to confuse our mental models with our language. Polygraph examiners have for years used the term "specific issue" meaning single issue, with a mistaken understanding that this means diagnostic accuracy. Some trainers will still make this mistake.

The NAS/NRC noticed this confusion and suggested the more descriptive term "event specific" because it enforces a clearer understanding of the circumstances of a diagnostic test (known event or allegation) with the circumstances of a screening test (no know event or allegation).

This is important to us in PCSOT and other areas that are becoming more interested in successive hurdles testing strategies, because we often intend to use a test with greater diagnostic accuracy (meaning: increased specificity) at the second stage of testing. We have to keep in mind that test results, in the absence of external evidence or symptoms, are not themselves a sufficient ethical or scientific basis for a diagnostic conclusion. Test results, coupled with evidence, are sufficient for a diagnosis.

skar:

Something to think about in a one RQ exam: Iatrogenic effects (unanticipated undesired effects) are not uncommon in the sciences. It is possible that a single RQ causes increased attentional loading (salience) onto that question, resulting in a possible increase in the potential for FPs. It is also possible that a single linguistic approach depletes the volume of stimulation and sensitivity to deception, causing a potential increase in FNs. It is also possible that the present strategy of providing multiple versions of the single issue (differing only in linguistic construction) is the optimal solution: dispersing attentional loading/salience for truthful persons, and providing optimal stimulation for deceptive persons. It is also possible that the screening practice of including multiple RQs is optimal for the same reasons.

We don't know, because we haven't studied it enough. If you are going to study it, then that would be great.

What is important to the long term success and credibility of the profession of scientific polygraph testing, is that we begin to refrain from just making up new things based on our anecdotal experience (which is always biased ouround our work environment).

With well established and validated techniques and principles becoming increasing visible in our own published scientific literature, it is not wise to replace known things with unknown unproven psychologizing and unproven expertizing (because we've been doing this for X number of years, and surely our expert opinions and expert experience is a suitable replacement for data from scientific study.)

So, if you are planning to study these questions, that's great. If you are planning to simply use some expertized solution, then it would be best to be very very cautious about that. Field practices are best based on the data. When we don't have data, we sometimes standardize things simply to limit the volume of chaos - but we cannot afford to mistake that kind of standardization for scientific validity.

.02

r

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

posted 07-13-2009 12:06 PM            

You have correctly exposed my error. I did state "Specific Issue" when I meant "Single Issue". I thank you for your diligent attention to detail and I beg 10,000 of your most gracious pardons, but I am not quite prepared to commit Seppuku over the deal.

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

posted 07-13-2009 06:03 PM            

If it ain't broke, don't fix it.

----------

The TES/DLST seems to work well in it's intended usage.

The only reservations I have about it are: 1) the potential use of DLC techniques in the same programs for which examinees will also undergo PLC testing for the same purpose, and 2) how to adapt to PCSOT circumstances in which the examinee makes substantial admissions.

It would seem to me to be preferable to use one approach - DLC or PLC. I don't know this with any certainty, but it's worth paying attention to the potential complications.

The second one is interesting. In a TES program, substantial admissions are probably unlikely - they would amount to an arrest for treason. In a TES program, I imagine the testing task would be that of testing any involvement in the issue of concern.

In a DLST/LE program there might be admissions regarding some issues - and these alone may or may not disqualify the applicant. So, it may be necessary to test the limits of the examinee's admissions - if he is not disqualified already.

In a DLST/PCSOT environment it would be common for some examinees to make a number of admissions about some types of rule violations - pornography use, alcohol/drugs, child contacts, unreported sexual partners, etc. It is sometimes useful to test the limits of these admissions, and sometimes not. Sex offenders, like other humans, tend to minimize - there is always more - unless they are externally motivated to exaggerate. External motivation is usually in the form of a legal problem that someone wishes to mitigate by emphasizing a mental health problem, but can also occur in the form of lengthy, high-pressure, or highly-coercive interrogation (especially with younger persons or functionally impaired persons), and, of course (perhaps?) torture.

People mostly minimize and under-report, and our basic assumption should be that there is more information somewhere. The question is this: is it possible to know "everything," without prompting someone to over-report. What is the probability that someone reports perfectly - no under-reporting, and no over-reporting. Another question is this: what incremental validity is achieved, in our risk assessment and risk management decisions, by attempting to gain that additional information, when the person has already made substantial admissions.

Another question: what is the likelihood that someone will pass the test, having just made substantial admissions? Answer: low.

Yet another question: what is the value of testing a known issue, when we know we will most likely be unable to adequately resolve the unknown issues about which the examinee has made no admissions?

Sure, there may be times when it is clinically useful to let someone struggle with the test, and follow up in therapy with some form of come-to-Jesus conversation.

There are plenty of times when a therapist or PO will state clearly that they have become primarily interested in the accuracy of resolution to the other questions, for which the examinee made no admissions.

Despite the fact that secrecy and deception are normally occurring animal activities - often necessary for survival - a values-based approach will state that lying and deception are bad. We therefore nearly always say that we expect complete truthfulness about everything. To the extent that we allow ourselves to engage in concrete thinking, we will feel it necessary to test the limits of the examinee's admissions.

Offenders learn that they will have trouble on their polygraphs. To help ensure they will "pass", then sometimes choose to exaggerate. Later, they will answer the surveys of research-minded polygraph opponents regarding being compelled to make false admissions by the "unfair" polygraph process. To them, exaggeration = false-admission.

In short, when we test the limits of substantial admissions, we may not just be limiting our ability to effectively investigate the remaining unknown issues, for which the examinee has made no admissions - we might be loading our enemies' gun for 'em.

It is sometimes better to simply remove a question from a test when the examinee makes a bunch of admissions. Of course, we should assume there is more - and it would be pretentious to ever assume we know everything.

A present problem with the DLST is that it we do not have a procedural solution for removing a question from the test when it is wise or necessary to do so.

Someone should study your one RQ method for that purpose.

.02

r

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

posted 07-13-2009 10:42 PM            

I am concerned that if we assume the role of decision makers we encourage the people who are supposed to be making the decisions to abrogate their responsibility.

When you say "we do not have a procedural solution for removing a question from the test when it is wise or necessary to do so.", do you mean that the test structure doesn't allow it or that it simply lacks an easily definable threshold that allows us to standardize the justification to remove a question?

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

posted 07-14-2009 01:00 AM            

Of course we, as polygraph examiners, and as individuals, don't make risk assessment or risk management decisions - unless we wear another hat in some other place - any more than we make hiring decisions, or decisions about guilt. These decisions are made by risk evaluators, risk managers, personnel administrators, and courts.

We do participate in the processes attached to those decisions, and it is our obligation to understand that - else we will fail to understand the responsibilities and limitations of our particular role in those processes.

Our test result, coupled with evidence, in the form of a confession or other evidence developed by other investigators may be a basis for someone's conclusion about something.

It does us no good to try to deny our role in these processes. On the other hand, it does us a lot of good to understand the role that polygraph information and polygraph result play in risk assessment and risk management tasks - even if the leg-work of those tasks is executed by other professionals.

We will do smarter and more useful work if we understand the context, just as sex offender polygraph examiners might want to know a little bit about sex offenders (including how things sometimes go wrong), security compliance examiners might want to know a little about the dynamics of security (including how things sometimes go wrong), and law enforcement examiners might want to know a little bit about things like police-work and criminals. Spend a little time in the trenches and you gain valuable experience that can be of assistance in helping a referring agent select effective targets or make an action plan for what to do next when the examinee makes a bunch of admissions.

Many of our referring agents want us to be subject matter experts. They want our help in using the polygraph effectively. They want us to be sufficiently cross-trained that we understand a little bit about their work and their needs. We want them to be sufficiently cross-trained that they can work expediently with the strengths and limitations of the polygraph test, including things like effective handling of the examinee before test and effective target selection. Otherwise, we're at risk for some silly expectation that we simply doing whatever silly thing they ask of us. Worse - we're all at risk for acting like vacuous stuff-shirt experts who babble unaccountably at each other while trying to mask some fear-driven insecurity that others will find out that we don't know what the f%^&* we're doing. It's better to get a little cross-training and know a little bit of enough about what the f*%^ we're doing that we can both understand it and explain it in simple enough terms that we aren't dependent on some need to just baffle others with bulls(&*%. A credible and accountable scientific profession is explainable to outsiders. Otherwise we're stuck in some unimpressive circular explanation that say's "you're not trained in this so you can't understand." Circular arguments like this are a red flag to people that something is amiss. When you encounter a circular argument it generally means that someone either doesn't really understand the material, or is simply faking and making it up as they go. As Einstein said: "if you can't explain it simply, then you don't understand it well enough." (I know, I don't always explain things simply - but I'm learning too.)

The TES documents that we have access to do not include any procedural instruction regarding the removal of a question. It is also clear that it is a tightly structured technique, and it appears that the tight structure is perhaps partly responsible for its effectiveness.

With the common varieties of MGQTs it seems to be more commonly acceptable to remove a question or two when necessary, and conduct the test with three or two RQs. We don't worry so much about changing the characteristics of the ole' MGQT, and subsequently changing the test's validity profile (our assumption that the validity data from research is representative of and applies to the test we conduct), because the MGQT isn't all that well studied and it's validity characteristics are largely unknown anyway.

With the TES we have a test with impressive validity characteristics, and it is desirable to be able to claim that supporting research data as representative of the test we conduct. Change the test too much and you loose that claim, and you're back to the unknown validity that we're trying to get away from.

I suspect we are mostly in agreement here, and simply interpreting, at times, each other's postings for their concrete face-value - which is a kind of hazard when not speaking face to face. Not at all surprised that you already know about the event-spec/single-ix/specif-ix confusion.

Always fun arguing with you. Hope to see you at APA this year.

Niters,

r

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

posted 09-14-2010 11:17 AM            

And what type of questions do you recommend for a specific issue examination, also DLC or PLC?

Thanks.

posted 05-31-2011 07:55 AM            

Thanks.