- Welcome to AntiPolygraph.org Message Board.
A forum for free and open discussion of polygraph-related topics
Posted by J.B. McCloughan
- Oct 05, 2002, 12:30 AMJust an idea to maybe spark some meaningful discussion. In a past discourse, George asserted the following:
Let us return to this discussion but with the GKT in place of the CQT. A sort of mirrored discourse, probably more of a discussion.
Quote
Again, as Lykken notes at p. 135, none of these studies are definitive, and reliance on polygraph-induced confessions as criteria of ground truth results in overestimation of CQT accuracy, especially in detecting guilty subjects, to an unknown extent. In addition, chance is not necessarily 50-50.
Moreover, as Dr. Richardson eloquently explained, CQT polygraphy is completely lacking in any scientific control whatsoever, and as Professor Furedy has explained, it is also unspecifiable and is not a genuine "test." Lacking both standardization and control, CQT polygraphy can have no meaningful accuracy rate and no predictive validity.
Let us return to this discussion but with the GKT in place of the CQT. A sort of mirrored discourse, probably more of a discussion.
Posted by Marty (Guest)
- Sep 19, 2002, 05:35 AMQuote from: J.B. McCloughan on Sep 19, 2002, 02:17 AM
However, the first amendment does not stipulate the necessity for police to convey specific details about a crime and its investigation to the media.
Of course it doesn't place such requirement on police and I didn't mean to suggest that it did. Rather, other countries often have laws that prohibit media exposure of criminal information that may be used at trial. Both Canada and the UK have strict rules in this regard since they have no first amendment serving as a media shield. The media here is subject to few limitations and even gag orders are applied only to the principals. The first amendment is an American treasure in spite of being an occasional pain in the ass.
As for CIT or GKT, I just prefer the former term though I can't say I really care much one way or the other. You are correct that the ultimate meanings are essentially equivalent, at least in a forensic setting.
-Marty
Posted by J.B. McCloughan
- Sep 19, 2002, 02:17 AMMarty,
Although guilty construes a person committed a crime and is a verdict in criminal trial, guilty is used in correlation with knowledge for the purposed assertions. Taken in this correlated context, guilty knowledge definably suggests that a person has or does not have criminal knowledge or knowledge of the crime. Concealed information suggests that the person has or does not have information of the crime. When the two are compared, there is little, if any, difference in the prejudice, presumptive, or accusatory inferences of the two avowals. It is my opinion that the nomenclature was changed to eliminate the presumed need to explore the aforementioned definition in a criminal proceeding , albeit quite simple, and possible preemptive challenges.
You are correct in saying that the Japanese and Americans convey information to the media on criminal investigations in a different manner. However, the first amendment does not stipulate the necessity for police to convey specific details about a crime and its investigation to the media. This has become common practice more frequent and to a greater degrees within recent years for quite different reasons.
The necessary training to apply the GKT is not so significant. The prior preparation needed to conduct a good GKT is more significant than that which is found in a CQT.
Your last statement;
I adamantly agree with.
J.B.
Although guilty construes a person committed a crime and is a verdict in criminal trial, guilty is used in correlation with knowledge for the purposed assertions. Taken in this correlated context, guilty knowledge definably suggests that a person has or does not have criminal knowledge or knowledge of the crime. Concealed information suggests that the person has or does not have information of the crime. When the two are compared, there is little, if any, difference in the prejudice, presumptive, or accusatory inferences of the two avowals. It is my opinion that the nomenclature was changed to eliminate the presumed need to explore the aforementioned definition in a criminal proceeding , albeit quite simple, and possible preemptive challenges.
You are correct in saying that the Japanese and Americans convey information to the media on criminal investigations in a different manner. However, the first amendment does not stipulate the necessity for police to convey specific details about a crime and its investigation to the media. This has become common practice more frequent and to a greater degrees within recent years for quite different reasons.
The necessary training to apply the GKT is not so significant. The prior preparation needed to conduct a good GKT is more significant than that which is found in a CQT.
Your last statement;
Quote
I would like to see work done (such as Drew's) that produces dependable mechanisms for detecting deception and don't require fooling the subject. There are valid uses for this.
I adamantly agree with.
J.B.
Posted by Marty (Guest)
- Sep 17, 2002, 05:20 AMAgreed. The CIT acronym is actually less biased and presumptive. OTOH, GKT seems to be the more common term here so I chose that. I too prefer the term CIT since it is less accusatory.
Japan has more secrecy surrounding criminal matters that is constrained here by first amendment considerations and it seems there may be some argument that CIT questions are more available to Japanese polygraphers as a result. Still, I think the major factor here is simple inertia. It's hard to change people who perhaps would have to look at significant retraining were the norms to be altered.
Please understand, I am a EE by profession and have never been subject to a polygraph but am curious about many things. I am drawn especially to areas where ignorance is actually seen as a desireable thing. There is something quite disgusting about a PLT test where the examiner is expected to lie about the "stim" test as well as the significance of the CQ's. I remember when I first saw the 60 Min program I thought of course, they are trying to get you to lie to calibrate your response! This site just documents what was pretty clear supposition from seeing the show. The site's documentation is consistent with what I have read from authoritative sources as well. However, I disagree that the polygraph is likely a 50-50 proposition. I do agree though that it is difficult or impossible to reliably calibrate and so it may well be impossible to prove that it is not 50-50. What I most object to is that in administering it the examiner exagerates the reliabilty - in order to enhance the effectiveness. Time and time again, and not just at this site, I have seen instructions to administer the STIM test then exclaim how it clearly indicates when the subject lied - regardless of the actual charts. Just rubs me the wrong way though I understand why it's done.
I don't know what polygraphy's future is but the idea that naive subjects are the most common seems unsustainable, amongst the more educated people being screened. The Internet has changed everything.
I would like to see work done (such as Drew's) that produces dependable mechanisms for detecting deception and don't require fooling the subject. There are valid uses for this.
-Marty
Japan has more secrecy surrounding criminal matters that is constrained here by first amendment considerations and it seems there may be some argument that CIT questions are more available to Japanese polygraphers as a result. Still, I think the major factor here is simple inertia. It's hard to change people who perhaps would have to look at significant retraining were the norms to be altered.
Please understand, I am a EE by profession and have never been subject to a polygraph but am curious about many things. I am drawn especially to areas where ignorance is actually seen as a desireable thing. There is something quite disgusting about a PLT test where the examiner is expected to lie about the "stim" test as well as the significance of the CQ's. I remember when I first saw the 60 Min program I thought of course, they are trying to get you to lie to calibrate your response! This site just documents what was pretty clear supposition from seeing the show. The site's documentation is consistent with what I have read from authoritative sources as well. However, I disagree that the polygraph is likely a 50-50 proposition. I do agree though that it is difficult or impossible to reliably calibrate and so it may well be impossible to prove that it is not 50-50. What I most object to is that in administering it the examiner exagerates the reliabilty - in order to enhance the effectiveness. Time and time again, and not just at this site, I have seen instructions to administer the STIM test then exclaim how it clearly indicates when the subject lied - regardless of the actual charts. Just rubs me the wrong way though I understand why it's done.
I don't know what polygraphy's future is but the idea that naive subjects are the most common seems unsustainable, amongst the more educated people being screened. The Internet has changed everything.
I would like to see work done (such as Drew's) that produces dependable mechanisms for detecting deception and don't require fooling the subject. There are valid uses for this.
-Marty
Posted by J.B. McCloughan
- Sep 17, 2002, 04:14 AMMarty,
I was just finishing a response to some of your last few messages, when I saw you had posted some new thoughts. First I will discuss those past assertions.
You brought up the problem of examiner bias. Bias has the potential for presence in most situations that involve human interaction or decisions, not limited to polygraph. If an examiner has knowledge of the correct answers to the questions, it does not present a per se bias. Although a naïve examiner could administer a valid GKT, they could in no way construct one. Drew's suggestion of pre-recorded questions and there alternative answers is one way to minimize bias. Another is to check the original examiner score/rank against blind examiners and algorithms for consistency. The subjecting of the test to known samples can also presumably reduce bias in a GKT. If there had been key information leaked, the known sample test would quite probably discover it.
More recently you wrote;
You are quite correct in this and I would just add that Japan uses polygraph only for criminal issues in a forensic setting. Japan refers to their procedure by a different nomenclature, Concealed Information Test (CIT). Also, polygraph evidence is excepted and used in the Japanese judicial system.
I was just finishing a response to some of your last few messages, when I saw you had posted some new thoughts. First I will discuss those past assertions.
You brought up the problem of examiner bias. Bias has the potential for presence in most situations that involve human interaction or decisions, not limited to polygraph. If an examiner has knowledge of the correct answers to the questions, it does not present a per se bias. Although a naïve examiner could administer a valid GKT, they could in no way construct one. Drew's suggestion of pre-recorded questions and there alternative answers is one way to minimize bias. Another is to check the original examiner score/rank against blind examiners and algorithms for consistency. The subjecting of the test to known samples can also presumably reduce bias in a GKT. If there had been key information leaked, the known sample test would quite probably discover it.
More recently you wrote;
Quote
I have been wondering why in Japan the GKT is so common that many polgraphers have not even given a CQT! My conjecture is that since GKT is more recent, and Japan tends to use them for forensic apps rather than screening, that they naturally gravitated to what has long been widely agreed is more reliable than the CQT.
You are quite correct in this and I would just add that Japan uses polygraph only for criminal issues in a forensic setting. Japan refers to their procedure by a different nomenclature, Concealed Information Test (CIT). Also, polygraph evidence is excepted and used in the Japanese judicial system.
Posted by Marty (Guest)
- Sep 17, 2002, 03:43 AMJ.B.,
I have been wondering why in Japan the GKT is so common that many polgraphers have not even given a CQT! My conjecture is that since GKT is more recent, and Japan tends to use them for forensic apps rather than screening, that they naturally gravitated to what has long been widely agreed is more reliable than the CQT. In the States, CQT polygraphy has been around a long time and a certain inertia and comfort level is acquired. Thus, like a potential well where there is an adjacent, but lower minima, it is easier to just keep on doing things the old way than try something new even if better since one has to climb a hill first. For that matter Drew may find the same factor occuring in his efforts to improve the technology.
-Marty
I have been wondering why in Japan the GKT is so common that many polgraphers have not even given a CQT! My conjecture is that since GKT is more recent, and Japan tends to use them for forensic apps rather than screening, that they naturally gravitated to what has long been widely agreed is more reliable than the CQT. In the States, CQT polygraphy has been around a long time and a certain inertia and comfort level is acquired. Thus, like a potential well where there is an adjacent, but lower minima, it is easier to just keep on doing things the old way than try something new even if better since one has to climb a hill first. For that matter Drew may find the same factor occuring in his efforts to improve the technology.
-Marty
Posted by J.B. McCloughan
- Sep 17, 2002, 03:27 AMDrew,
I agree that the use of an information based polygraph, such as the GKT, is as available as one might make it. One of the hurdles involved in implementing this type of a test, as I have discovered through personal endeavors, is not in its availability but in refocusing the investigator's thought process of how and when to use polygraph.
Inasmuch as the GKT is used at the investigation onset, it can quite readily offer investigators many advantageous uses of polygraph as a forensic examination that are not available to them with a CQT.
Too list a few of the GKT's strengths, for those whom are new to it;
1. It offers investigators a non-evasive test that shields the innocent suspect of an investigation. A properly conducted GKT is non-accusatory. A GKT constructed with 16 good
questions has an assumed false positive error rate of.1% and a false negative error rate of .3%
2. It provides investigators with a focus early in an investigation, which can most certainly save the investigators and their agency countless investigative hours and resources.
3. It's is a test that can be mathematically figured to an assumed probability of error rate. (As Drew explained in his last post, a GKT's assumed probability of error may be reduced even further by subjecting it to a test against known samples.)
I know that there are many more advantages in using this type of a test. I just offered a few to further discussion on this topic.
I agree that the use of an information based polygraph, such as the GKT, is as available as one might make it. One of the hurdles involved in implementing this type of a test, as I have discovered through personal endeavors, is not in its availability but in refocusing the investigator's thought process of how and when to use polygraph.
Inasmuch as the GKT is used at the investigation onset, it can quite readily offer investigators many advantageous uses of polygraph as a forensic examination that are not available to them with a CQT.
Too list a few of the GKT's strengths, for those whom are new to it;
1. It offers investigators a non-evasive test that shields the innocent suspect of an investigation. A properly conducted GKT is non-accusatory. A GKT constructed with 16 good
questions has an assumed false positive error rate of.1% and a false negative error rate of .3%
2. It provides investigators with a focus early in an investigation, which can most certainly save the investigators and their agency countless investigative hours and resources.
3. It's is a test that can be mathematically figured to an assumed probability of error rate. (As Drew explained in his last post, a GKT's assumed probability of error may be reduced even further by subjecting it to a test against known samples.)
I know that there are many more advantages in using this type of a test. I just offered a few to further discussion on this topic.
Posted by Marty (Guest)
- Sep 13, 2002, 05:53 PMQuote from: George W. Maschke on Sep 13, 2002, 08:04 AM
Marty,
In polygraph circles, it is an article of faith that the examiner must be pre-provided with as much information as possible about the incident being investigated, the case facts, and the person being examined.
h practice.
Yes, but where the CQT is used as an interrogation tool this is valid seeing as the purpose of interrogation is obtaining a confession - so long as it is limited to that. It is invalid, even downright dangerous, when the CQT is used in screening tests. For example, in the famous 60 Minutes segment, examiner bias is the best explanation for the 100% false positive results of their 3 tests.
The GKT, OTOH, can be applied by a naive examiner since interrogation isn't necessarily part and parcel of the GKT that is given a suspect.
-Marty
Posted by Drew Richardson
- Sep 13, 2002, 08:09 AMJ.B.,
I don't believe the majority of your polling questions can be meaningfully answered based on currently existing evidence. Your first question is the clear exception. Those polygraph methods that are generally classified as concealed information tests (to include the Guilty Knowledge Test) are the only ones with a theoretical basis for practice and which therefore hold any potential as representing valid diagnostic instruments. None of the so-called "Lie Detection" formats have any scientifically justified basis for practice, contain any semblance of scientific control, have any degree of acceptance in the scientific community, or have any basis for routinely meeting established tests for trial court admissibility.
Because of the aforementioned shortcomings of those polygraph formats/applications currently most widely in use, I believe concealed information testing represents the only hope for meaningful and productive polygraph practice. Because these tests will require some effort and other expenditure of resources (most worthwhile endeavors do) to collect and preserve investigative information, it will be incumbent upon the polygraph community during periods of necessary field research and later operational practice (if justified) to do much more than offer the usual lip service paid to this type of examination, i.e., "It's nice to be able to use information-based testing when available..." The extent to which this testing is "available" will be directly related to the efforts put forth by investigators and polygraphers alike as they fufill their respective roles in advance of test administration.
Because others have commented upon the potential problem of examiner bias and its potential effect on in-test stimuli presentation, I suppose I will follow suit. Because concealed information tests are strictly information based, do not require the pre-test "setting" of control questions, etc., and in general do not require any of the exam-specific interaction between examinee and examiner that occurs in a lie detection test, concealed information tests can be constructed and administered so as to avoid the aforementioned problem. The guilty knowledge test (questions and alternative answers) can be pre-recorded by an individual completely unfamiliar with the matter being investigated. Although poor stimulus choice, etc still could exist and confound (as with any test), the potential bias reflected in the voice inflection of a case-knowledgeable examiner would be eliminated as well as any spurious affect/effect produced through the typical "lie detection" pre-test encounter. A separate benefit (other than eliminating examiner bias effects) of the pre-recording of the exam stimulus set (again, unavailable with lie tests) is that it can be used to validate the stimulus set (not just general validation) for any given investigation. This is accomplished by having some representation of those known to be familiar with the investigation (e.g., investigators) and those known to be unfamiliar with the investigation take the exam. By doing this one can look for potential sources of false positive and false negative error and modify the stimulus set appropriately before the administration of the actual examination(s) of interest.
Regards,
Drew Richardson
I don't believe the majority of your polling questions can be meaningfully answered based on currently existing evidence. Your first question is the clear exception. Those polygraph methods that are generally classified as concealed information tests (to include the Guilty Knowledge Test) are the only ones with a theoretical basis for practice and which therefore hold any potential as representing valid diagnostic instruments. None of the so-called "Lie Detection" formats have any scientifically justified basis for practice, contain any semblance of scientific control, have any degree of acceptance in the scientific community, or have any basis for routinely meeting established tests for trial court admissibility.
Because of the aforementioned shortcomings of those polygraph formats/applications currently most widely in use, I believe concealed information testing represents the only hope for meaningful and productive polygraph practice. Because these tests will require some effort and other expenditure of resources (most worthwhile endeavors do) to collect and preserve investigative information, it will be incumbent upon the polygraph community during periods of necessary field research and later operational practice (if justified) to do much more than offer the usual lip service paid to this type of examination, i.e., "It's nice to be able to use information-based testing when available..." The extent to which this testing is "available" will be directly related to the efforts put forth by investigators and polygraphers alike as they fufill their respective roles in advance of test administration.
Because others have commented upon the potential problem of examiner bias and its potential effect on in-test stimuli presentation, I suppose I will follow suit. Because concealed information tests are strictly information based, do not require the pre-test "setting" of control questions, etc., and in general do not require any of the exam-specific interaction between examinee and examiner that occurs in a lie detection test, concealed information tests can be constructed and administered so as to avoid the aforementioned problem. The guilty knowledge test (questions and alternative answers) can be pre-recorded by an individual completely unfamiliar with the matter being investigated. Although poor stimulus choice, etc still could exist and confound (as with any test), the potential bias reflected in the voice inflection of a case-knowledgeable examiner would be eliminated as well as any spurious affect/effect produced through the typical "lie detection" pre-test encounter. A separate benefit (other than eliminating examiner bias effects) of the pre-recording of the exam stimulus set (again, unavailable with lie tests) is that it can be used to validate the stimulus set (not just general validation) for any given investigation. This is accomplished by having some representation of those known to be familiar with the investigation (e.g., investigators) and those known to be unfamiliar with the investigation take the exam. By doing this one can look for potential sources of false positive and false negative error and modify the stimulus set appropriately before the administration of the actual examination(s) of interest.
Regards,
Drew Richardson
Posted by George W. Maschke
- Sep 13, 2002, 08:04 AMMarty,
The problem of examiner bias you raise is quite relevant to CQT, R/I, and GKT polygraphy. In polygraph circles, it is an article of faith that the examiner must be pre-provided with as much information as possible about the incident being investigated, the case facts, and the person being examined.
With regard to "facilitated communication," see also The Skeptic's Dictionary and Gina Green's article, "Facilitated Communication: Mental Miracle or Sleight of Hand?" A number of parallels could be drawn between the phenomenon of "facilitated communication" and current polygraph practice.
The problem of examiner bias you raise is quite relevant to CQT, R/I, and GKT polygraphy. In polygraph circles, it is an article of faith that the examiner must be pre-provided with as much information as possible about the incident being investigated, the case facts, and the person being examined.
With regard to "facilitated communication," see also The Skeptic's Dictionary and Gina Green's article, "Facilitated Communication: Mental Miracle or Sleight of Hand?" A number of parallels could be drawn between the phenomenon of "facilitated communication" and current polygraph practice.
Posted by Marty (Guest)
- Sep 13, 2002, 05:57 AMThe GKT can be validated statistically assuming each question is independent and has high entropy (close to 50-50 probability of answer A or B) and that there are enough questions available. The most likely source of error though may be unintentional cues. It is best that the examiner not know the answers to the asked questions to avoid this. Just how difficult this is, even for honest, well intentioned examiners is well understood. Even the most educated and experienced professionals such as medical doctors are careful not to reach broad conclusions without prophylactic measures such as double blind technique.
Here is an example of a widely used practice, "Facilitated Communication" which has proven to be largely bogus yet has significant believers in parts of the academic community and demonstrates the risks inherent in any such process where the examiner's biases may be a factor.
http://soeweb.syr.edu/thefci/apafc.htm
I think that GQT would be amenable to scientific study though since experimental techniques to check the statistics are well known. This may be particularly applicable here since clinical trials sometimes are done where the underlying processes are not well understood.
-Marty
Here is an example of a widely used practice, "Facilitated Communication" which has proven to be largely bogus yet has significant believers in parts of the academic community and demonstrates the risks inherent in any such process where the examiner's biases may be a factor.
http://soeweb.syr.edu/thefci/apafc.htm
I think that GQT would be amenable to scientific study though since experimental techniques to check the statistics are well known. This may be particularly applicable here since clinical trials sometimes are done where the underlying processes are not well understood.
-Marty
Posted by J.B. McCloughan
- Sep 13, 2002, 01:38 AMThere have been many a past debate on the scientific validity of the Control Question Test (CQT). However, little debate has occurred with regards to the Guilty Knowledge Test (GKT). I would like your input on the scientific validity of polygraph when using the GKT.
I will begin this thread with a very simple public poll.
1. Is the GKT based on sound scientific theory?
2. How accurate is the GKT?
3. If asked, would you take GKT polygraph?
4. Should the GKT be accepted as evidence in court?
I will begin this thread with a very simple public poll.
1. Is the GKT based on sound scientific theory?
2. How accurate is the GKT?
3. If asked, would you take GKT polygraph?
4. Should the GKT be accepted as evidence in court?