Get documents about "Lee1 206 23 20Iacono
Document Sample
1 SECOND JUDICIAL DISTRICT COURT
COUNTY OF BERNALILLO
2 STATE OF NEW MEXICO
3 CS-2003-0026
4 KEVIN LEE, et al,
5 Petitioners,
6 vs.
7 HONORABLE LOURDES MARTINEZ, et al,
8 Respondents.
9 TRANSCRIPT OF PROCEEDINGS
10 On the 23d day of June 2003, at approximately 9:00 a.m,
11 this matter came on for hearing before the HONORABLE RICHARD
12 KNOWLES, DIVISION XV, judge of the Second Judicial District,
13 State of New Mexico.
14 The Petitioners, KEVIN LEE, et al., appeared by Counsel
15 of Record, CHIEF PUBLIC DEFENDER JOHN BIGELOW, by JEFF REIN,
16 Attorney at Law, 505 Central Avenue, N.W, Albuquerque, New
17 Mexico 87102;
18 And FREEDMAN, BOYD, DANIELS, HOLLANDER, GOLDBERG &
19 CLINE, P.A, by CHARLES W. DANIELS, Attorney at Law, 20 First
20 Plaza, Suite 700, Albuquerque, New Mexico 87102.
21 The Respondents, HON. LOURDES MARTINEZ, et al, appeared
22 by Counsel of Record, OFFICE OF THE ATTORNEY GENERAL by
23 MICHAEL COX, CHERYL JOHNSTON and ART WEIDEMANN, Attorneys at
24 Law, 111 Lomas, N.W, Albuquerque, New Mexico.
25 At which time the following proceedings were had:
1
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 JUNE 23, 2003
2 (Note: Court in session at 9:00).
3 THE COURT: Lee et al. V. Martinez et al. This has
4 a local cause number of CS-2003-26. If I could have
5 appearances, please.
6 MR. COX: Michael Cox, Cheryl Johnston, Art
7 Weidemann for the State of New Mexico.
8 MR. DANIELS: Charles Daniels and Jeff Rein for
9 the petitioners.
10 THE COURT: The other petitioners are aware of the
11 proceeding?
12 MR. DANIELS: Yes.
13 THE COURT: And my recollection is we discussed
14 this before, but just so it's clear on the record, it's
15 everybody's understanding that it's acceptable for the
16 defendants not to be present during the course of the
17 proceeding.
18 MR. DANIELS: Right. This is basically a legal
19 proceeding that's generally applicable to a question of law
20 in the State of New Mexico. It's not even case-specific.
21 THE COURT: I agree to the concept and the only
22 concern I had is since whatever decisions are made here may
23 influence their lives, and I am going to be taking testimony
24 from various experts in the field, I just want to make sure
25 it's clear.
2
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 MR. DANIELS: Yes, Your Honor. This is in the
2 nature of a law-making proceeding, even though there are
3 findings and conclusions. It will be reviewed by the
4 Supreme Court in its capacity as law-makers for the State of
5 New Mexico.
6 THE COURT: Okay. Are both sides ready to
7 proceed?
8 MR. DANIELS: We are ready, Your Honor.
9 MR. COX: We are ready.
10 THE COURT: Any preliminary matters before we
11 begin with the first witness?
12 MR. DANIELS: Maybe only with respect to marking
13 exhibits. Mr. Cox handed us exhibits that he is going to
14 use today and has marked them. We have supplied on both
15 sides an enormous quantity of material to the Court with
16 followings. I think it was important to make a full record
17 to have the materials available. I am wondering how you
18 want to get those marked for the record. I think all of
19 them probably ought to be considered by the Court because at
20 least they show the extent to which there has been
21 publication, public debate and so on, and they will be used
22 at the hearing either as demonstrative or illustrative, such
23 as some of the ones Mr. Cox presented here today, and I
24 think substantive as some of his may be used. Perhaps if we
25 just mark all of the things that have been submitted and
3
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 then be able to refer to them by Exhibit A, Document H-3, it
2 might be easier for the Court instead of marking every
3 single piece of paper separately.
4 THE COURT: That concept is fine but let me give
5 you information. I think you are aware the volume of
6 materials I received. I have got close to two full banker's
7 boxes full of materials. Many of them are in binders such
8 as the one that I have on the Bench. In all candor, I did
9 not bring them all back. In anticipation of the proceeding
10 I brought what I can carry. I have them in a discrete
11 location in the office and I can have them brought down.
12 The other thing is I have read a good portion of the
13 materials you provided to me for both sides. I haven't read
14 100 percent of it but I have read a lot of it. Some of the
15 stuff I have read I have underlined, and some of it I have
16 not. The only concern I have is as I was reading, if I
17 underlined or put stars next to it and stuff like that --
18 I'm sure the Supreme Court can read things on their own, but
19 I wanted both sides to be aware that if you have concerns of
20 something I noted or made a comment next to it, you know,
21 perhaps I should have anticipated this but I have marked
22 your exhibits.
23 MR. DANIELS: Number one, I personally don't have
24 a problem with that, with any of the markings in this case.
25 THE COURT: Rather than "he is a liar" or "he is
4
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 an idiot."
2 MR. DANIELS: But if there's any concern, we will
3 be glad to use the clean copies that we have here in
4 displaying them in court and have others made.
5 THE COURT: It may be best for the Supreme Court
6 to have a clean copies available. I can easily have brought
7 down the set, and I am hearing you say that may not be a bad
8 plan, so I will get somebody to bring down all of the
9 materials. Frankly, I was leaving them up there because I
10 anticipated going up and referring to them if I needed to
11 either during the evening or something like that. But are
12 you anticipating I need to be looking at these while the
13 testimony is going on?
14 MR. DANIELS: I don't think so. I think we can
15 project them. I see that there's an Elmo here. We can
16 project them up on the screen, hand them up if necessary.
17 THE COURT: Okay.
18 MR. DANIELS: In the same way we would a single
19 piece of paper. Both sides have different sets.
20 THE COURT: Okay.
21 MR. COX: I think we have enough extra copies that
22 you don't need to worry about writing on ours. We have
23 killed a lot of trees and have whole sets.
24 THE COURT: I am having ecological angst over the
25 whole process.
5
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 MR. COX: I am thinking we should see how this
2 goes. We have a pile of paper this high now. You may not
3 decide that all of it is relevant ultimately and you may
4 focus and we may be admitting certain items. We may want to
5 argue out what the actual record should be, because I
6 suspect there is a lot of stuff that will not become -- not
7 even being referred to out of the materials.
8 THE COURT: If that happens, it's fine with me.
9 To be really blunt, my perception of my role in this entire
10 proceeding, and either way is fine with me, but my
11 perception is I am in an unusual position, because as I
12 understand the Supreme Court order on this, I am to prepare
13 findings of fact and conclusions of law and indicate my view
14 of this under Alberico and the State cases on this.
15 But I also see, and my perception is I have the
16 opportunity to have both sides lay everything out for the
17 Supreme Court to revisit everything I do. Because they will
18 look at it on these individual cases and my perception or my
19 suspicion is they want to look over the policy on this as
20 well, so this is an opportunity for them -- kind of a
21 conduit, so it's kind of a weird thing.
22 Usually you know you will get appealed or there's a
23 reasonable chance in any case you have or do, but I am not
24 sure I have had one where an appeal is so guaranteed. I
25 feel I am more of a special master than a trial judge on
6
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 this.
2 MR. COX: I agree with you, Your Honor. In fact,
3 it will be a de novo review. You are almost like a
4 magistrate.
5 MR. DANIELS: We will mark them, Your Honor. If
6 there's something counsel persuades the Court ought to be
7 ripped out at the end, fine.
8 THE COURT: Basically if you think this is nuts,
9 let me know, but I will error on the side of giving them as
10 much material as you all think is really going to be
11 helpful. If there's stuff that neither side is referring to
12 during the course of the proceedings and the smoke clears
13 and you think well, maybe they can look it up if they really
14 want, we can handle it that way. Whatever. I'm going to
15 probably error on the side of giving them more information
16 than not. If either side wants something in, it's hard for
17 me to say no. Even if I do, I suspect you could send it to
18 the Supreme Court anyway.
19 MR. COX: Your Honor, we don't intend to object
20 about much of that. Two other things. We would like a
21 moment to set this up. Some judges are very particular.
22 Does it matter to you? Will this work?
23 THE COURT: This is fine, as long as I can see it.
24 As long as both sides are happy with their ability to see
25 it, I am happy too.
7
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 MR. DANIELS: I think we worked cooperatively on
2 this, Your Honor. I have a screen much larger than that
3 which might be easier for the Court to see which we brought
4 over on Friday. The security guard told us he locked it in
5 this courtroom's Judge's withdrawing room, whatever that is.
6 I haven't been able to find it this morning. There's a
7 screen and projector. There's a much larger screen and it
8 might make it better for Mr. Cox's witnesses as well as
9 ours.
10 THE COURT: Take five. I will join you. We will
11 take a short recess.
12 (Note: Court stood in recess at 9:13 to 9:20).
13 MR. COX: I will invoke the rule.
14 THE COURT: Okay. The rule is invoked.
15 MR. COX: Professor Iacono.
16 WILLIAM IACONO
17 (Being duly sworn, testified as follows)
18 DIRECT EXAMINATION BY MR. COX
19 Q. Would you please tell us your full name.
20 A. William Iacono.
21 Q. Spell your last name.
22 A. I-A-C-O-N-O.
23 Q. What is your educational background?
24 A. I have a bachelor of science degree in psychology from
25 Carnegie-Mellon University.
8
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 Q. Could you repeat that answer?
2 A. I have a bachelor of science degree in psychology from
3 Carnegie-Mellon University and a doctor of philosophy and
4 psychology from the University of Minnesota which I obtained
5 in 1978.
6 Q. What have you done in terms of work since you
7 graduated?
8 A. I had a post-doctorate from the University of Minnesota
9 for a year and faculty appointment there in the department
10 of psychiatry, and for six years on the faculty at British
11 Columbia, Vancouver Canada. 1985 I returned to the
12 university of Minnesota as a professor where I have been
13 since.
14 Q. What positions do you hold at the university?
15 A. Currently I am a professor in the psychology department
16 and I have joint appointments in other departments,
17 including psychiatry, neuroscience, law and child
18 development. I am also a distinguished McKight University
19 professor and a Co-director of the Minnesota Center for Twin
20 and Family Research.
21 Q. What is a distinguished McKnight?
22 A. This is an endowed position that was contributed to the
23 university by the McKnight Foundation to aid the university
24 to retain top faculty, so people -- it's a competitive award
25 given by the university to faculty that they identify as
9
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 particularly important to the future of the institution. I
2 get an endowment for five years to spend plus this title.
3 Q. You mentioned the Minnesota Center for Twin and Family
4 Research?
5 A. Yes.
6 Q. What is that?
7 A. That's a project that I run that's a longitudinal study
8 of 2000 families that contain twin and adoptive adolescent
9 children. We study them longitudinally from typically
10 around age 11 to around age 30 in an effort to understand
11 how genetic and environmental factors contribute to their
12 mental health as they become adults.
13 Q. You said that you are the director?
14 A. I am the co-director with another faculty member.
15 Together we obtain grants that we use to fund this research.
16 Our grants presently total about ten million dollars so we
17 spend a little over two million dollars a year to run the
18 projects. We have approximately 50 employees that it takes
19 to execute this work.
20 Q. The grant process, is that competitive?
21 A. It is. Our funding is from the National Institute of
22 Health. We have to write proposals that are reviewed by a
23 grant committee, NIH, typically composed of ten or twelve
24 scientists who are judged to have expertise relevant to the
25 type of work that we do. They meet and review our work.
10
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 All of the members on the committee vote, assigning a score
2 that indicates the overall merit of the work that we do, and
3 then the scores are converted into percentiles where our
4 grant in effect is rated against the other grants that all
5 the other investigators in the country would submit to NIH.
6 Q. How did you do the last go-around?
7 A. The last go-around we did very well. The grants -- in
8 fact, the last two grants scored in the 97 and 99th
9 percentiles in the review process.
10 Q. Is that from 100 percent?
11 A. 100 percent would be perfect. You can't get 100
12 percent in the way they score things, so for us it means
13 more money, which is good.
14 Q. Now, the other that you are working in, is this a
15 competitive academic area? If you could describe that in
16 more detail, please?
17 A. It is competitive. It's -- in the first place, the
18 psychology department at the University of Minnesota is one
19 of the top ranked psychology departments in the country, so
20 a lot of people there are doing cutting edge research. It's
21 a department that's heavily supported by the university. We
22 are expected to public papers in top journals to get the
23 best graduate students in the country and to get funds to do
24 the type of research that we do. Our work is nationally and
25 internationally recognized. For instance, the University of
11
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 Minnesota is well known for work it has done in the twin
2 studies, and the clinical psychology department there, which
3 is what I am also a part of and was the director of for
4 seven years, is also internationally known as one of the top
5 clinical programs in the world.
6 Q. Can you tell us what is your specialty within the field
7 of psychology?
8 A. My specialty is psychophysiology.
9 Q. What is that?
10 A. Psychophysiology involves taking physiological
11 measurements from humans, like measuring brain waves or
12 cardiovascular activity, palmar sweating, and make
13 inferences about psychological process or things going on in
14 a person's mind. So, for instance, if a psychophysiologist
15 was interested in how stressful it is to testify, I might
16 put sensors on a witness' body and monitor the physiological
17 signals while the person is testifying and see if they are
18 more aroused and physiologically activated when sitting in a
19 chair doing something else.
20 Q. Is there a relationship between psychophysiology and
21 polygraphy?
22 A. There is. Psychophysiology is the parent science of
23 polygraphy. Polygraph testing involves making the same
24 sorts of physiologic report, goes to make inferences about
25 whether a person is being truthful or deceptive.
12
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 Q. Does it use similar measurements?
2 A. Well, psychophysiology includes measuring pretty much
3 every physiological signal that you can pick up from a
4 human, whereas in polygraph tests, typically respiration,
5 cardiovascular activity and palmar activity or the GSR, the
6 sweat glands on the fingertips are the three things
7 monitored.
8 Q. Have you received special recognition for your
9 accomplishments from a scientific organization?
10 A. I have. I received distinguished scientific
11 contribution awards from the American Psychological
12 Association, the Associate for Psychological Research. I
13 have also --
14 Q. Slow down just a little.
15 A. Sorry. I have also been elected to the status of
16 fellow in various divisions of the American Psychological
17 Association. I held various offices for professional
18 organizations, most notably serving as president for the
19 Society for Physiological Research, which is an elected
20 position.
21 Q. Do you publish scientific papers?
22 A. Yes.
23 Q. Any idea how many?
24 A. Over 200.
25 Q. You also presented papers or addressed scientific
13
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 meetings?
2 A. Yes.
3 Q. Any idea how many of those?
4 A. Probably close to or over 200.
5 Q. Any of these deal with polygraph testing?
6 A. Yes, they do.
7 Q. Any idea how many?
8 A. Probably the professional presentations at meetings and
9 the publications, probably about 50, and probably about half
10 of those would be papers and half would be presentations.
11 Q. Have you ever received government grants to do
12 polygraph research?
13 A. I have. I have received grants from the Research
14 Council in Canada, which functions the same way that the NIH
15 does here in the United States, so it's a competitive
16 process for obtaining grants. I have also obtained funds
17 from the Department of Defense Polygraph Institute.
18 Q. Are you currently doing any research in polygraph?
19 A. I am not doing any research on polygraph testing per
20 se, but a student of mine is doing a study on juries.
21 Q. Have you been a consultant to the governmental
22 agencies?
23 A. Yes, I have.
24 Q. What agencies do you recall?
25 A. I have consulted with the Office of Technology
14
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 Assessment, which is a scientific office that provided
2 information for congress on polygraph testing as it relates
3 to the employment context. I have consulted with the CIA,
4 President Clinton had a Joint Security Commission and I
5 consulted with those individuals who were interested in
6 trying to consolidate the use of polygraph tests across the
7 many different government agencies. I have also been a
8 consultant to the Department of Defense and served on an
9 advisory panel to the Department of Defense Polygraph
10 Institute for several years to consult with them about their
11 educational curriculum and research program.
12 Q. Have you been qualified as an expert witness in the
13 past for the purposes of testimony either in court or in
14 administrative hearings in polygraph?
15 A. Yes.
16 Q. Do you have any idea how many times?
17 A. Over 30.
18 Q. Is that also including New Mexico?
19 A. Yes.
20 Q. Do you know if that was state or federal or both?
21 A. Well, I am pretty sure I testified in state court in
22 New Mexico on one occasion, maybe two. I can't recall. But
23 I testified twice here, once in federal court.
24 Q. You are familiar with the Supreme Court Scheffer case?
25 A. Yes.
15
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 Q. Were you cited in there?
2 A. Yes, we were cited there.
3 Q. You testified in front of legislative bodies?
4 A. Yes, I testified before a legislative committee for the
5 Kansas State Legislature and also for the United States
6 Senate.
7 Q. On polygraph?
8 A. Yes, for the United States Senate was the Judiciary
9 Committee.
10 Q. Are you familiar with the National Academies of
11 Sciences report on polygraph?
12 A. Yes, I am.
13 Q. What is the National Academies of Sciences?
14 A. It's an organization that is composed of probably the
15 most distinguished scientists in the country who are elected
16 by people who are already members, and they take on various
17 tasks related to helping to formulate science policy and
18 integrating information about science to advance the
19 national indemnity for the country as relates to various
20 scientific topics.
21 Q. Did you participate in any way in the polygraph study?
22 A. The National Academies panel that investigated
23 polygraph held hearings and also invited people to speak
24 before the panel and I was invited to speak before the panel
25 so I gave them a presentation a couple years ago.
16
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 MR. COX: Your Honor, we offer Professor Iacono as
2 an expert in psychophysiology.
3 MR. DANIELS: No objection.
4 THE COURT: He will be accepted.
5 Q. I wanted to ask you about the National Academies study.
6 Do you know the scientific -- are you familiar with any of
7 the scientists that actually worked on the study?
8 A. Well, I met them as a consequence of giving the
9 presentation, but several of them I knew in advance because
10 I worked with them or had met them at scientific meetings or
11 just colleagues. For example, two members of the panel are
12 also past presidents of the Society for Psychophysiological
13 Research, so I worked with them in that context. One member
14 of the panel, David Faigman, he is the law professor in
15 California who put together the modern scientific evidence
16 volume that we have made a contribution to. There are
17 several other people on the committee that I have also known
18 through reputation or interactively.
19 Q. We will come back to that in a little bit. I would
20 like to start by asking you how a polygraph machine works?
21 A. A polygraph is sensitive recording device that measures
22 those three channels of physiological activity that I
23 mentioned, the cardiovascular channel that monitors the
24 pulse and blood pressure; respiratory channels that measure
25 the expansion and contraction of your chest when you breathe
17
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 in and out, and palmar sweating channel that measures the
2 electrical conductivity of your skin which varies when you
3 sweat.
4 This information is recorded on a moving chart paper or
5 with more modern applications the chart paper has been
6 replaced by a computer where the physiological information
7 is digitalized and might be displayed directly on the
8 computer screen in a way that would resemble what you would
9 see on chart paper.
10 Q. Is there any kind of specific physiological response
11 that could be identified as a lie?
12 A. No, there is no lie response, no equivalent of
13 Pinocchio's nose growing too long, so you can't just measure
14 physiological activity and know that a person is lying. In
15 fact, there's no physiological response that's unique to any
16 emotion.
17 Q. So how does a machine differentiate between truth and
18 lie if there's no specific response that can be identified
19 as a lie?
20 A. The machine doesn't really make the differentiation.
21 The examiner has to make the differentiation in the context
22 of conducting a procedure that involves asking people
23 different types of questions and determining which questions
24 they respond more to than others, and then making an
25 inference based on that, whether the person is lying or
18
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 telling the truth.
2 Q. Is there more than one polygraph technique?
3 A. There are many polygraph techniques.
4 Q. The one used the most, what would be that called?
5 A. Control question test.
6 Q. Is that -- can you describe how that works?
7 A. That would be used the most in criminal investigations,
8 but there's another test that would be called the employee
9 screening test that has various things. That one actually
10 might be used the most because it's given to employees.
11 It's given literally to thousands of people. But in the
12 criminal context, the control question test is the test
13 that's given the most.
14 It has three components. They are referred to as the
15 pre-test interview, and then the actual test where the
16 physiological data are recorded, and then a post-test
17 interview, which might include an effort to obtain a
18 confession, if the individual was deemed to have failed the
19 physiological test.
20 In the pre-test interview the examiner meets with the
21 examinee. They review the examinee's health and sleep and
22 diet and whether the person is taking medication, those
23 sorts of things that could affect the test. The examiner
24 hears the examinee's account of the charges and the
25 accusation against him, and the examiner works with the
19
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 examinee to formulate questions, and there's two types of
2 questions that are critical to the outcome of the test. One
3 is called the relevant question, and this is the
4 did-you-do-it question, so if it's a bank holdup, for
5 instance, it would be, "Did you rob the bank on the night of
6 March 16th," or something to that effect.
7 It's important for the examiner to make sure that the
8 language in this question is clear and that the person
9 taking the test understands it so there's a bit of give and
10 take during the pre-test interview to establish that the
11 examinee can actually answer the question no and feel
12 comfortable that he is clear and understands the question.
13 Then the other important question that's developed in
14 the pre-test interview is called the control question. This
15 is a question that's designed to get at integrity of the
16 person who is taking the test and might be involved in a
17 crime like this. So, for instance, for a bank robbery, the
18 control question might take the form of, "Have you ever
19 stolen something from somebody who trusted you?" And this
20 question is presented to the examinee in a way that's
21 designed to get the examinee to answer no, even if they have
22 and can recall having taken something from someone who
23 trusted them.
24 So the control question is presented in the context of
25 a presentation that goes something like this. "I want to
20
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 know if you are the type of person who would be likely to
2 commit a bank robbery, and if you were, maybe there would be
3 other aspects of your character that would be flawed. For
4 instance, it wouldn't be the case, would it, that you would
5 have ever stolen something from somebody who trusted you,
6 would you?"
7 That type of psychological manipulation is used to
8 essentially coax the person answering the control question
9 no, even if they felt that they wanted to answer yes. If
10 the person answers yes, the examiner would say, "Well, tell
11 me about whatever you have done," and would basically say
12 something to the effect of, "Certainly you haven't done
13 these sorts of things repeatedly, have you? So if I were to
14 modify the question to say something like, 'Other than what
15 you told me about, have you ever stolen something from
16 someone who trusted you, you could answer that no, couldn't
17 you?'" Eventually you reach a point where the person says
18 they can answer the questions no.
19 But the assumption is when the person is answering the
20 question no, they are lying or at least very concerned about
21 their answer. So the intent of the control question, which
22 is also sometimes referred to as a probable lie question, is
23 to elicit a response that would be indicative of lying.
24 Q. This is compared with the relevant question?
25 A. This is compared with the relevant question. So the
21
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 way the test ends up being scored or administered further,
2 is now that the questions are formulated, the person is
3 attached to the polygraph equipment and now these questions
4 are read to the person who then answers them no, and the
5 physiological data are recorded, and the test is typically
6 repeated several times maybe with the questions in a
7 different order, and sometimes there is a test that's
8 included, another kind of test included with this procedure
9 that goes by various names. It's called the card test, the
10 stimulation test, the acquaintance test.
11 This is a procedure that's designed to convince the
12 person taking the test that it actually works, and there are
13 many variations of it, but an example would be you would
14 tell the person to pick a number between one and five, and
15 sometimes that number would be written down on a piece of
16 paper and placed face-up in front of the examinee. Other
17 times the examinee would be told to take a marked card from
18 a deck so the examiner knows for sure what number card they
19 have taken and they conceal it from the examiner.
20 But the point is that this card test or stimulation
21 test is then followed by a series of questions that might be
22 "Was the number 3, was the number 5, was the number 1," and
23 the physiological measurements are made. At the end of the
24 test the examinee is told that the test is working well, a
25 distinct response is evidenced by the number that the person
22
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 has chosen, and then the other parts of the control question
2 test that is the next succession of questions that would be
3 repeated three times.
4 So the next two tests, for instance, would be
5 administered after the stimulation test.
6 Q. Let me ask you before we move on, why did the
7 polygraphers want to convince the examinee that the machine
8 is working using the card test?
9 A. Well, it's important to convince the examinees that
10 it's working so they will have their -- as the concept goes,
11 as it's articulated by people who present these tests, it's
12 important to convince them it's working so they will be
13 properly focused on the questions that are likely to pose
14 problems for them. So the idea is if you believe the test
15 works and you are guilty, then you will worried about your
16 response to the relevant question. If you are innocent and
17 you believe the test works, and you know you are lying to
18 the control questions, then you will be worried that you are
19 going to give a large response to those questions.
20 The way the test is presented, it's presented in such a
21 way that the examinee is led to believe that the answer to
22 all of the questions is important to the outcome the test.
23 So there's no distinction made between relevant and control
24 questions or that you need to worry about how you respond to
25 one question but not so much about how you respond to
23
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 another. They are basically told that these are all the
2 questions and you want to be completely truthful when you
3 answer them, and the test really works, and if you are
4 truthful with this test which is almost infallible we will
5 be able to tell whether you are lying or telling the truth
6 today. That would be how it's presented.
7 Q. After they do the skin test and the series of charts,
8 what happens then?
9 A. Then the data are examined, and there are different
10 ways that the data are examined. Some polygraphers have
11 been trained in a method referred to as global chart or
12 global analysis in the outcome. These examiners take into
13 account everything that's presented before them in their
14 determination of guilt or innocence.
15 For instance, they would use the physiological tracings
16 but also the case facts, their analysis of the examinee's
17 account of the story relating to the accusation, whether or
18 not the person makes eye contact, whether the person seems
19 excessively nervous, inconsistencies in the person's
20 stories, these sorts of things, along with the physiological
21 data would all be used to reach a verdict.
22 Many examiners today use what's called a numerical
23 scoring of the charts. This involves trying to do a
24 quantitative appraisal of which type of question elicits the
25 larger response. So the way this works, for instance, each
24
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 control question and each relevant question are examined in
2 pairs for each channel of physiological activity. So if you
3 take the first control question, the first relevant
4 question, you would examine, for instance, the galvanic skin
5 response, the GSR response, to term determine if the
6 relevant question gave the larger response or the control
7 question gave the larger response.
8 If the relevant question -- a numerical score is
9 assigned. If it's a lot larger it would be minus one; if
10 not, it's a minus three. If the control question produced a
11 larger response it would get a score of plus one, plus two
12 or plus three.
13 Then this is repeated for the respiration channel and
14 the blood pressure channel and repeated for each pairing of
15 the control and relevant question that show up on the test,
16 which typically has three such pairs. Then the test is
17 typically repeated three times, so the information is
18 assembled across these many repetitions of the asking of the
19 questions and their pairing, and the three physiological
20 indicators of response.
21 Eventually all these numbers are summed, and if the sum
22 is plus six or greater, the person is being truthful. Minus
23 six or smaller, the person is deemed deceptive.
24 Another approach to the chart scoring that has become
25 more commonplace nowadays is based on computer analysis of
25
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 the charts. So polygraph examiners who have physiological
2 data digitalized and entered directly into a computer can
3 use commercially available software that will do the
4 equivalent of numerical scoring of the charts. And also
5 produce a probability statement regarding the likelihood
6 that the person is being truthful or deceptive.
7 Q. You described this as control question test, probable
8 lie?
9 A. Yes.
10 Q. Are there other types of tests besides this?
11 A. There are. There's a directed lie test. The directed
12 lie test is very similar to the control question test,
13 except that the control questions are replaced with what are
14 called directed lies. So in the control question test, the
15 control question is something that the examinee is more or
16 less tricked into answering deceptively or at least presumed
17 that the person is answering deceptively. This involves a
18 skilled examiner to be able to pull off this manipulation in
19 a way that's convincing.
20 So the directed lie is enough to get around the problem
21 by just identifying one of the questions as one that the
22 person is deliberately lying to. So an example would be,
23 "Have you ever broken even one regulation or rule?" So the
24 person would say, "Yes, I have done that." The examiner
25 would say, "Good. It's important because I'm going to ask
26
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 you this question so I get an idea of what your response
2 looks like when you are lying. In fact, I want you to think
3 of a time when you did that and answer the question no. I
4 want you to deliberately lie when you answer the question."
5 This test, other than that one difference, is pretty much
6 like a regular control question test in the way it's
7 administered. The questions are paired and it's scored.
8 Q. What about relevant and irrelevant test?
9 A. The relevant/irrelevant test is the first polygraph
10 procedure that was introduced back in the 1920s and 30s. It
11 is unlike the other two procedures in that there is no real
12 control question or comparison question that gives you an
13 idea to get an example of what a person's response looks
14 like when they are telling something that would resemble a
15 lie.
16 In a relevant/irrelevant test, people are asked
17 questions like, "Are you sitting down, did you rob the
18 bank," and the idea would be if you robbed the bank you
19 would respond more to the relevant question, but if you
20 didn't, there would really be no difference in the response
21 of the two questions. The relevant/irrelevant test is still
22 used. It's not used very much in criminal investigations
23 but it is used by some examiners including employment
24 context.
25 Q. How about something called the guilty knowledge test?
27
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 A. Guilty knowledge test, the procedure is different from
2 these types of procedures because it's not based on lying
3 per se. It's based on recognition memory. And it's based
4 on the idea that people give an orienting response or they
5 give a recognition response, would be another way you can
6 say this, to a stimulus that they are familiar with.
7 In this case the stimulus is a detail about a crime
8 that they may or may not be involved with. So the guilty
9 knowledge test is presented as a series of multiple choice
10 questions, and again, if you use the bank robbery as an
11 example, for instance, of what the question might be like
12 would be, "If you held up this bank, then you would know
13 what the holdup note that the teller was handed had written
14 on it. Did it have written on it," and then you would have
15 a sentence. "Give me all your money, did it have written on
16 it this is a stickup?" You might have five alternatives.
17 If you held up the bank you would know what you wrote on the
18 holdup note and you would give a recognition response to
19 that alternative but no recognition response to any of the
20 other alternatives because they mean nothing to you.
21 If you are innocent, you give no recognition response
22 to anything whatsoever, because you have no idea what's
23 written on the holdup note. That would be the first
24 multiple choice question. There would be a series that
25 would relate to the main robbery. For instance, the next
28
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 question might be, "If you robbed this bank then you will
2 remember what the teller that you robbed looks like." And
3 you might show pictures of five people. One of them is the
4 teller and the other four pictures would be people who were
5 not the teller. If you robbed the bank, you might recognize
6 the teller; if you didn't, you wouldn't have reason to
7 recognize any of the people.
8 The next question would be, "If you robbed the bank you
9 might know how much money you got. Did you get $10,000,
10 etc." If you didn't, you wouldn't have a recognition
11 response. So by having a series of questions and seeing if
12 the person has a recognition response to each of the guilty
13 knowledge alternatives, eventually you would have an
14 indication of whether or not the person is being truthful
15 when they deny involvement in the crime.
16 Q. Now, in the area of what you called specific incident
17 polygraph, is that also the kind that we use in criminal
18 cases?
19 A. Yes.
20 Q. Is the control question or the probable lie the most
21 common?
22 A. Yes.
23 Q. Let's talk about that for a while. Set aside the other
24 two. Are the probable lie or the comparison lie tests
25 standardized in the way that they are either administered or
29
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 sworn?
2 A. They aren't very standardized.
3 Q. What do you mean?
4 A. Well, the administration of the test clearly is not
5 very standardized. The examiner actually uses the pre-test
6 interview and the estimation of the character of the
7 examinee to figure out, for instance, what vocabulary to use
8 in the questions, what questions would be appropriate, what
9 ones wouldn't, what terms to review with the person, what
10 terms not to review, how to formulate the actual questions,
11 what would be appropriate control questions under the
12 circumstances, how far to go in pushing the person to reveal
13 information about the content embedded in the control
14 questions that's designed to elicit a probable lie. All of
15 these things are just related to the formulation of the
16 questions.
17 Two examiners trying to develop the same question list
18 from the same person are actually going to have different
19 question lists. It's very unlikely they would have the same
20 question list. Then the questions need to be ordered. So,
21 for instance, you need to decide which control question
22 comes first, which one comes second. The first time the
23 test is presented should the control question A be presented
24 with relevant question A or relevant question B. You need
25 to decide whether to do a stim test at all.
30
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 Some polygraph examiners don't do the stimulation test.
2 Some do before the first chart, some after the second chart,
3 and some only if they feel the first chart doesn't indicate
4 strong responses.
5 Then the questions are presented in a different order.
6 Again, what order would that likely be. It's up to the
7 examiner. Between the administration of the first chart and
8 the second chart, the questions are reviewed again with the
9 examinee to make sure that the examinee is still comfortable
10 with the questions.
11 The questions might be changed then, might not. That's
12 the decision that the examiner would need to make, so there
13 are many points along the way in the administration of the
14 test where the test is clearly not standardized. Even the
15 physiological channels that are measured might vary somewhat
16 from examiner to examiner, although they are all likely to
17 involve the three components that I mentioned.
18 The scoring, there are many scoring approaches
19 available to score polygraph tests. There are different
20 schools of thought, for instance, about how to best do
21 numerical scoring. If you get a computer program and that
22 computer scores your charts because that's an invariant
23 procedure, at least that choice of that procedure would be
24 standardized in the sense that every time you put the same
25 information into the computer you should get the same
31
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 result, but choosing which computer program to use, for
2 example, would be an example of another way in which the
3 procedure is not standardized.
4 So throughout the procedure basically it's -- there are
5 many cases or many places where examiner judgment is
6 important in determining how to formulate questions, how to
7 present them, how to go over the material with the examinee,
8 etc. and all of these make the procedure nonstandardized.
9 Q. Is the polygraph test an objective test?
10 A. An objective test is one in which the results of the
11 test are independent of the behavior of the examiner, and
12 that's not the case here. In fact, most people in the
13 polygraph profession will tell you that the most important
14 part of the test is the quality of the examiner, him or
15 herself, and that that person is crucial to getting the test
16 to work properly.
17 So any procedure that's put together like that is
18 clearly not objective, it's subjective. The outcome depends
19 on many decisions that the examiner makes along the way and
20 different examiners could get different outcomes because of
21 the decisions they make.
22 Q. What about the scoring? People always agree on
23 scoring?
24 A. They don't always agree on scoring either. People who
25 are trained in the same scoring procedure, so that they
32
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 follow the same rules, typically obtain fairly similar
2 results. There have been many research studies, including
3 studies that I have done, that show that, for instance,
4 individuals who are trained to do numerical scoring in
5 exactly the same way are likely to get scores that are very
6 similar. So one examiner might score something plus ten and
7 the other might get plus 13 or something. They wouldn't
8 necessarily be identical but very similar.
9 So the scoring end of things, there's a little more
10 objectivity and standardization, but the input in there's
11 much more subjectivity.
12 Q. Have you also looked at the difference between the
13 original score, the original polygrapher, how he scores, and
14 a blind scoring rates the test?
15 A. Yes. Typically, the original examiner is -- well, this
16 is kind of a complicated question, because it gets into how
17 you decide who is telling the truth and who is being
18 deceptive.
19 But typically blind rescores of original examiner's
20 charts are likely at least to get something close to the
21 same numerical score if they use the same procedure. In
22 research that we have done, one of the things that we have
23 noticed is that blind rescoring of charts usually leads to a
24 more conservative verdict, so, for instance, if we did a
25 study with the Royal Canadian Mounted Police where the
33
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 original examiner scored 275 charts, and the scores were
2 very spread out, and when we had the blind independent
3 examiners rescore the charts, the scores were much more
4 pressed towards the middle. So the independent evaluators
5 that rescored the charts were less likely to assign high
6 deceptive scores or high truthful scores and more likely to
7 assign conclusive scores.
8 But in general if they follow the same procedure, they
9 got scores that were close to those of the original
10 examiner.
11 Q. What accounts for that?
12 A. I think in part what accounts for it is the fact that
13 the original examiners are influenced by the extra
14 information that they have when they score the test. In
15 fact, our study with the Royal Canadian Mounted Police
16 showed that. The original examiners actually at times would
17 overrule their numerical scoring. In other words, they
18 would have a numerical score that might indicate deceptive,
19 and they would write a report that indicated an inconclusive
20 or even a truthful outcome.
21 So in other words, they were paying close attention to
22 the extra polygraph cues, the cues that come from
23 information not contained in the polygraph charts, to help
24 them decide how to make a decision.
25 The blind examiners don't have this information. All
34
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 they have is the physiological tracings alone and they just
2 -- without that information they know what's a control
3 question, what's a relevant question. They don't know what
4 the charge is, don't even know what the wording of the
5 questions are. If it's a bank robbery, murders, they don't
6 know. They are looking at the charts to see if the response
7 to the relevant questions is larger than the response to the
8 control questions. They are unaffected by the extra
9 information, and that's probably a major reason why they
10 show the shift.
11 Q. What is the Rosenthal effect?
12 A. The Rosenthal effect is a phenomenon identified about
13 30 years ago in psychology. It relates to the fact that
14 psychologists and other scientists and people, for that
15 matter, who are only human, who have an investment in a
16 theory or idea or hypothesis that they then proceed to test
17 are likely to arrange the test in such a way that they get
18 favorable results.
19 So, for example, if I am doing a psychological study to
20 test the theory, I have to design an experiment that many do
21 that. If I am heavily invested in the theory like it's the
22 Iacono theory, for instance, then I am likely to conduct my
23 experiments in such a way that it generates data that are
24 favorable to my theory.
25 This is one of the reasons why psychological research
35
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 and social science research needs to be replicated. Because
2 oftentimes people who do the first study are doing subtle
3 things. These are not deliberate distortions but just
4 making decisions as they go along that are more likely to
5 give the outcome they want than not. Then when an
6 independent scientist attempts to replicate the study,
7 making decisions where there's no investment in getting a
8 particular result, oftentimes they don't replicate the
9 study. That's the nature of the Rosenthal effect.
10 Q. You are talking about an inadvertent -- not a
11 deliberate but inadvertent influence?
12 A. That's right. People are human. They might try to be
13 objective and control their biases but, of course, we can't
14 do that perfectly. The fact that we can't do it perfectly
15 then affects the decisions we make and the steps we take,
16 for instance, when conducting psychological research.
17 Q. Now, is this Rosenthal effect, any reason to believe it
18 would not also influence the polygraph examiner?
19 A. It's the same situation. But for the polygraph
20 examiner the hypothesis involves the examiner's sense of
21 whether the person is guilty or innocent based on all of the
22 information presented at the time of the polygraph test. As
23 you mentioned, the examiner has access to the case facts,
24 interviews the examinee. If the examinee's account of the
25 crime, for instance, is full of inconsistencies and
36
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 unconvincing, the examiner might believe that the person
2 must be guilty. You don't even have to do the test. This
3 then might affect how the test is conducted.
4 Likewise if the examinee was particularly impressed
5 that the person taking the test seems to be getting a raw
6 deal and really seems to have been truthful, this can also
7 affect the way the test is conducted.
8 Now, again, the point is not that the examiner is
9 deliberately trying to alter the way the test is going.
10 It's that the examiner, being only human, and picking up on
11 these kinds of cues, is likely to make decisions along the
12 way that favor the expected outcome.
13 Q. Do you know if the National Research Council, the book
14 we have been talking about, Polygraph and Lie Detection out
15 of the National Academies, do you know if they looked at
16 this issue of subtle influence?
17 A. I believe they did.
18 Q. I would like to show you -- Page 42, Counsel. I have
19 highlighted a portion there. Would you read that to
20 yourself and tell us what they think of that possibility.
21 Do you know which chapter that's in? Why don't I put it on
22 the screen. Does this indicate that they have found or at
23 least in the course of the study others have indicated to
24 them that polygraph examiners do alter test results subtly?
25 A. Yes, that's basically what it says there. Depending on
37
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 the context, it changes the way the test is conducted in
2 making adjustments based on their expectations.
3 Q. And is it fair to characterize polygraph tests as
4 psychological if they are not standardized or objective?
5 A. It's not, because that's at least part of the
6 definition of a psychological test. It's a standardized
7 procedure that provides an objective index of a person's
8 behavior. A polygraph test such as the control question
9 test would fail that definition.
10 Q. How long have they been around, polygraph tests?
11 A. They have been around for about eight years, and the
12 control question test itself has been around for over 50
13 years.
14 Q. Is 50 years enough time, in your opinion, to come up
15 with a sound theory for a polygraph?
16 A. One would think you would have a sound theory by this
17 point, because they say a -- in psychology, typically
18 procedures that are important and that people pay attention
19 to develop a theoretical basis over the course of a few
20 years or a decade, so 50 years would be an outliar if you
21 didn't have a theory around such a procedure.
22 Q. What is the current state of knowledge regarding the
23 theoretical basis of why a polygraph might work?
24 A. It's pretty dismal. There isn't really any prevailing
25 theory or consensus as to why polygraph tests might work.
38
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 There actually, in addition to there not being any real
2 clear idea of why the test works that people embrace,
3 there's been very little research that's attempted to get at
4 the theoretical underpinnings of why the tests work. So the
5 net result is we don't really have a theoretical
6 understanding of why these tests work. The test I am
7 talking about here would be the control question test.
8 Q. Why is a theory important?
9 A. The theory is important -- it's critical to advancing
10 science, and as it relates to a technique, it's critical
11 because it helps us understand why a procedure is likely to
12 work or not likely to work. With that type of understanding
13 we can then generate hypotheses or ideas that we want to
14 test that help us refine the theory. So, for instance, if I
15 had a theory that the control question test works based on
16 differential basis of arousal mechanisms, I could generate a
17 hypothesis about arousal and generate experiments that
18 manipulate people's arousal when they take the test to see
19 if it influences the outcome. But if I don't do that type
20 of study or have a theory that even allows me to do that
21 type of study, I am never going to get at the underpinnings
22 of the test, never figure out how it works, never be able to
23 enhance the theory and improve the test. I would have a
24 difficult time understanding what effects the test and what
25 doesn't.
39
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 I guess my analogy, you could think of it as something
2 like this. If I had -- say I had a procedure for analyzing
3 organic chemicals, and I thought, you know, I bet I can help
4 solve crime scenes by matching tissue samples at the crime
5 scene to those of human beings by running the human being's
6 tissue sample and the one from the crime scene through my
7 organic analyzer. So I do this and sometimes I get a match
8 and sometimes I don't, but most of the time I get a match.
9 Now I say the error rate of the procedure is 10
10 percent, so I have this machine, looks at organic chemicals
11 and we take samples and we studied 100 people and see an
12 error rate of 10 percent, so it looks like I have a forensic
13 tool. How happy would we be with that? To me, that's kind
14 of the state of polygraph in the sense that we have people
15 doing studies trying to get at the error rate without
16 understanding what's going on.
17 The problem here for my organic analyzer is I don't
18 know what it is exactly I am analyzing. So, for instance,
19 if a subject has diabetes, is diabetes going to produce a
20 different output in my organic analyzer? Well, the tissue
21 of a person with diabetes is different from the tissue of
22 someone who doesn't have it. What about drugs? Since I am
23 measuring organic chemicals and they can be altered by
24 drugs, could it be the case that someone could take a
25 chemical and make it so their tissue sample doesn't match
40
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 the one at the crime scenes? Or perhaps the person is using
2 drugs at the time they committed the crime and the sample is
3 contaminated by that. What about a neurological illness,
4 could this affect the tissue sample? Could the food you eat
5 that day affect the tissue sample? We don't know. So in
6 order to continue using the test I have to do hundreds of
7 empirical studies and I am never certain for sure when I am
8 safe with my test and that there isn't something that could
9 throw it off.
10 Of course, in real forensic science now, we have a test
11 that works like that. It involves looking at DNA, and we
12 might have started the procedure 100 years ago with some
13 kind of a chemical analyzer like what I am talking about,
14 but we learned about DNA and learned how it relates to genes
15 and learned that people show individual differences in their
16 genetic constitution, and we learned that you can measure
17 the things from something like an organic analyzer in such a
18 way that allows you to generate a DNA fingerprint, and we
19 know that DNA fingerprints are constant over time and are
20 not affected by diabetes and don't change with diet and are
21 not affected by the drugs they take, etc.
22 Now with the new theory that we refine and understand
23 and apply in a forensic context we can feel comfortable when
24 we go to a new setting or test a person that maybe we
25 haven't tested before who is different in some way. For
41
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 instance, let's say the person had a disease that we had
2 never heard of. Because we know how DNA works and knows the
3 disease doesn't involve DNA we know we can use the DNA
4 results.
5 So having a theory allows us to understand the
6 procedure and refine it to the point where we can understand
7 how to apply it in a way that keeps errors low and makes us
8 confident in the result. That's what's missing in the
9 polygraph test.
10 Q. The National Academies looked at this issue?
11 A. Yes. They concluded that polygraph testing is
12 essentially devoid of theory. They used terms like
13 "atheoretical". Despite decades of work there's no theory
14 and they are clearly bemoaning the fact that there's no
15 theory.
16 Q. Was there also a study done on polygraph by an
17 organization in Germany?
18 A. Yes, there's a man named Fiedler at the University of
19 Heidelberg in Germany. The German legal system and the
20 German courts were considering the possible use of control
21 question polygraph tests as admissible evidence in Germany,
22 and this psychologist who has no connection either pro or
23 con with the polygraph field, my impression is he knew very
24 little about it but had the basic psychological training
25 that one would need to have to take on a task of reviewing
42
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 this technique, he and his colleagues reviewed the technique
2 and analyzed it, focusing in particular on the validity of
3 the test.
4 They concluded that it has no validity, that there's no
5 real theory and that it has many shortcomings, including not
6 being standardized or effective, for instance.
7 Q. I would like to show you State's Exhibit 12. Is this
8 the study that you are talking about?
9 A. Yes.
10 Q. Did you just bring that with you this weekend when you
11 came?
12 A. I did.
13 MR. COX: I have given a copy to Mr. Daniels. I
14 move the admission of Exhibit 12.
15 MR. DANIELS: I was just handed a copy this
16 morning and haven't had a chance to look at it so I won't be
17 able to cross-examine but we can address it later when we
18 present our case.
19 THE COURT: Any objection to the admission?
20 MR. DANIELS: I have no objection to your
21 considering any of this.
22 THE COURT: Twelve is admitted. Do I have a copy
23 of this?
24 MR. COX: No, just got here.
25 (Note: State's Exhibit 12
43
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 admitted into evidence).
2 Q. We talked about the problems with the theory and the
3 National Academies agrees with you. This problem with the
4 theory is this something new or has this been brought up in
5 the past?
6 A. It's been brought up in the past. If you look at the
7 research literature on polygraph, it's almost entirely on
8 does it work or doesn't it. And there's not very much out
9 there about theory. Other reviews of the literature that
10 have been published in scholarly journals, for instance,
11 have pointed to the fact that there is no theory and it's
12 difficult to advance the field in the absence of a theory.
13 Q. That's the polygraph theory or advocates reacted by
14 doing theoretical work?
15 A. No, I think that's another thing brought up in the
16 National Academies report, is it was a surprise to them. In
17 fact, the research agenda, for instance, that the federal
18 government response source doesn't involve getting at the
19 theoretical underpinnings that would help us understand why
20 the test may or may not work.
21 Q. As opposed to theory, are there assumptions involving
22 the underlying control probable lie test?
23 A. There are three simple ones that need to be met for the
24 test to work.
25 Q. Did you prepare a chart that kind of gives us an idea
44
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 of what these assumptions are?
2 A. Yes.
3 Q. Tell us about the first one.
4 A. The first one is fundamental to the test because since
5 there is no lie response, the examiner has to be able to
6 formulate questions that the subject will answer
7 deceptively. This is an easier assumption to see satisfied
8 when you are thinking about the relevant questions, because
9 the relevant questions get at the "did you commit the crime
10 or not" issue. But because some crimes might involve
11 complex concepts like what constitutes inappropriate
12 behavior in a sex crime, this sort of thing, it's possible
13 that the relevant questions won't be properly formulated.
14 But at least that's the one question that you think would be
15 formulated properly. It's much more difficult to formulate
16 control questions that the suspect will answer deceptively.
17 In fact, one can only hope that the person has the
18 proper concern about the control question when they take the
19 test, because there's actually no way to know for sure when
20 the person takes the test that they are in fact answering
21 the control question deceptively or properly concerned about
22 it.
23 Q. Well, if, for example, the person does not lie in the
24 control question and they go through the process of
25 manipulation and there is nothing to hide when they say no,
45
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 they are telling the truth, how would that affect the test?
2 A. It depends whether they are guilty or innocent. But if
3 they are innocent and they don't care at all about the
4 control question, like they say "this question doesn't
5 bother me, I can answer this honestly," what they are
6 probably going to do is fail the test because the relevant
7 question contains an accusation, and the accusation is going
8 to be disturbing to anyone. It could even be more
9 disturbing, for instance, if the accusation involves someone
10 who is close to you, like a relative who was killed or a
11 relative who was abused. It's obviously going to be
12 disturbing to you because if you are accused of the crime,
13 you could be convicted, lose your job, go to prison, these
14 sorts of things.
15 The accusation that's contained in the relevant
16 question is a threat to you, and it could be very
17 disturbing, in other words, just to be asked the question
18 independent of whether you are telling the truth or not. So
19 if, you know, I walk into a room, for instance, and I
20 publicly accuse you in front of other people of sexually
21 harassing me, let's say, just the fact that I accuse you
22 would probably produce a strong physiological reaction, and
23 false accusations, in other words, produce the same
24 physiological reactions oftentimes that a true reaction
25 would.
46
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 The polygraph picks this up. So if you are not
2 concerned about the control question and are being falsely
3 accused of people in the relevant question, you will fail
4 the test.
5 Q. Is there any way to know, given a particular
6 individual, how much of a threat the control question would
7 be in connection to the relevant?
8 A. No, there isn't. I think actually that quote you
9 projected from the National Academies of Sciences on Page 42
10 of the book actually points that out. We don't know what
11 deceptive threshold to determine how to calibrate, in other
12 words, the control question, so it produces just the right
13 amount of lie concern that produced the desired effect for
14 the innocent versus the innocent person. In fact, the
15 German report, Exhibit 12, makes the same point.
16 Q. Is it going to change depending on who the person is?
17 Some people might be less threatened by a relevant question
18 even if they are guilty?
19 A. That's something we would need to study, but that would
20 be a reasonable expectation, depending on what the mechanism
21 is that produces the result.
22 But an example is a person who is grappling with an
23 accusation that's relatively new is probably going to
24 respond to it more strongly than a person who is grappling
25 with an accusation that's old. This is a phenomenon known
47
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 as habituation, and involves the fact that the first time
2 you are exposed to something that's novel, you respond to it
3 strongly, and over time you habituate to it.
4 So this would be true of someone accused of a crime in
5 the sense that if I am confronted with people repeatedly
6 over and over again for days, months, even years, about my
7 involvement in a crime, if I am saying, "No, I didn't do it;
8 no, I didn't do it; no, I didn't do it," eventually I will
9 become increasingly comfortable with my denial of the
10 involvement in the crime even if I was involved. That's
11 habituation and that lessens my response to the relevant
12 question. So to the extent that occurred, that would
13 increase the likelihood that I pass the test.
14 Q. Would it depend on the background of the person being
15 accused?
16 A. It could as well, because some criminals are probably
17 routinely invested in denying their involvement in criminal
18 acts in a way that would make them comfortable with it.
19 However, there is research that has examined at least one
20 type of criminal that is believed to routinely lie about
21 their involvement in crimes, and that's the person called
22 the psychopath or sociopath. These people have been
23 examined in two laboratory studies, and in those studies
24 they were found to have no particular advantage in this
25 crime scenario used in the studies.
48
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 The mock crime scenario used in the studies that was
2 done is still not the same type of scenario that one finds
3 in real life where people are dealing with accusations that
4 extend out over periods of time and that they might be
5 heavily invested in and this sort of thing.
6 We really don't really know, in answer to your
7 question.
8 THE COURT: Mr. Cox, this may be a good time for
9 the morning recess. Let's go ahead and break for about
10 fifteen minutes and what I will do, since I am working
11 remotely, Ms. Harper will make sure everybody is here and I
12 will be following in.
13 (Note: Court stood in recess at 10:30 to 10:50).
14 Q. Before we broke I think we were talking about
15 Theoretical Assumption No. 1. What is your professional
16 opinion about whether that is a reasonable assumption?
17 A. No. 3 is one of the ones that's more reasonable, but
18 it's still implausible that it would hold all of the time.
19 Q. I would like to talk about directed lie but let's not
20 do that at this point. We are talking about the probable
21 lie. What about Assumption No. 2?
22 A. That is people react more strongly to the control than
23 the relevant questions. This assumption, I think, is
24 problematic for the reason that I mentioned before the
25 break, which is that the control questions might not mean
49
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 much to the innocent person, and the relevant questions mean
2 a great deal to everyone that asked them. To that extent
3 that you have that type of imbalance, innocent people will
4 look deceptive on the test.
5 Q. Is that because they react to the relevant question
6 because it's an accusatory question?
7 A. That's right. The relevant question is an accusation.
8 It has a certain emotional impact, just being asked it. The
9 control questions, which seem fairly innocuous and which
10 most people would understand aren't going to land them in
11 jail, for instance, or get them prosecuted, are innocuous in
12 contrast although they might involve something that looks a
13 little bit like an accusation. So the imbalance and the
14 emotional impact of being asked the two types of questions
15 for an innocent person is going to skew the test to they are
16 deceptive.
17 Q. What about Assumption No. 3? Why isn't that the same
18 as Assumption No. 2?
19 A. I'm not sure what you mean.
20 Q. Why is there a difference between the guilty person
21 responding more strongly than the converse? Why aren't they
22 mirror images? That's a bad question. Why is it a
23 different assumption?
24 A. Well, it's a different assumption because this
25 assumption is how the test has to work for a guilty person
50
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 or it won't work. The other assumption is how the test has
2 to work for an innocent person or it won't work. So, in
3 fact, in the literature there's an effort to sort of
4 separate how well the test works for innocent and guilty
5 people. Because different psychological mechanisms are
6 probably involved for innocent and guilty people, it
7 wouldn't be correct to look at the error rate without
8 looking at it independently for the two conditions. So you
9 can't really combine them because the psychological process
10 that's involved for an innocent person taking the test is
11 quite different from the process for a guilty person.
12 Q. What is the strength or weakness of this assumption?
13 A. Well, there's a couple weaknesses. One I already
14 mentioned is the habituation problem. That is oftentimes
15 when guilty people take a test, they have been dealing with
16 the relevant question in the form of an accusation for an
17 extended period of time. It doesn't matter whether they are
18 hooked up to a polygraph or not during that period of time.
19 Their body is responding even though it's not picking up
20 anything because there are no sensors attached. What's new
21 to the guilty person when they take a test is the control
22 question. These are questions they haven't seen before, and
23 assuming they were naive would be quite unexpected. So the
24 salience of the control question is heightened in the sense
25 that they are new and the accusatory relevant questions are
51
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 old.
2 So to the extent that's the case, that would explain
3 why a guilty person will actually respond more to the
4 novelty of the control question, and to the habituating
5 significance or nonsignificance, if you will, of the
6 relevant question.
7 Another problem here is that if you understand how the
8 test works, or even if you don't, to the extent that the
9 control question deals with something that you are very
10 worried about, you are going to pass the test even if you
11 are guilty. So an example would be for the bank robber, the
12 possibility when he says, "Did you ever steal from someone
13 who has trusted you," that he has committed some other
14 significant undetected crime that would be covered by that
15 question. And he might feel, "Well, the hot water I am in
16 now is pretty bad, but if they find out about my involvement
17 in the other problems I am really sunk," in which case the
18 person would respond more strongly to the control question
19 than the relevant question even in that context.
20 But probably the most serious threat to the validity of
21 the third assumption is the fact that people, in fact, can
22 -- if they can understand how the procedure works -- use
23 counter-measures to defeat the test by augmenting their
24 response to the control question.
25 Q. How does that work?
52
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 A. There are different ways to do it. In general,
2 attempts to suppress the response to the relevant question
3 do not work. It's hard to turn the body off when it's
4 responding automatically. So most of the research in the
5 area is focused on attempts to augment the response to the
6 control question. Turns out this is relatively easy to do.
7 Anything you do to stimulate your body will create an
8 increase in the galvanic skin response and a change in the
9 respiration and blood pressure that can look like the type
10 of response you give when you are answering the question
11 alone.
12 For instance, if I pinch myself, bite my tongue
13 lightly, tighten my anal sphincter, push my toes on the
14 floor, tense the muscles in my body, all of these acts would
15 produce these types of physiological reactions.
16 So the key to a counter-measure is to come up with a
17 physiological self-stimulation that can't be detected by the
18 sensors, and the sensors are attached to your arms and your
19 chest and your fingertips.
20 So if you do something like wiggle your toes or push
21 your toes on the floor or lightly bite your tongue, these
22 are examples of self-stimulations that would go undetected.
23 Q. Do they work?
24 A. In research that's been done, these counter-measures --
25 physical counter-measure maneuvers have been quite
53
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 effective. In fact, studies have shown that these physical
2 counter-measures and mental counter-measures which involve
3 doing stressful mental arithmetic when the control question
4 is asked in laboratory simulations produce 50 percent or
5 more instances where guilty people look truthful on a
6 polygraph. Those are called false negatives.
7 In other words, the laboratory studies at least suggest
8 that people can self-stimulate and escape detection, even
9 though they are guilty. Those studies also show that even
10 though the examiners and the research were -- knew ahead of
11 time that people were going to use counter-measures, they
12 couldn't tell who used them and who didn't. So it's not
13 only the case that people can use them. They can use them
14 effectively without getting detected.
15 Q. Do the examiners think they can spot those?
16 A. Virtually every examiner I worked with has always said
17 that they believe they can catch people who use
18 counter-measures, and that's based on the fact that in their
19 own experience oftentimes they run into people who clumsily
20 try to do this. So an example would be they have a blood
21 pressure cuff on their arm and the person tenses the muscle
22 in the arm, which then causes the polygraph to record that.
23 Or they deliberately engage in a controlled breathing effort
24 so that they pace their breathing so it doesn't vary. An
25 examiner can measure those things, and when they see that
54
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 they can guess pretty well that the person is using
2 counter-measures.
3 What the examiners don't see is all of the people that
4 use counter-measures that go undetected because by
5 definition they use them successfully and they went
6 undetected. They are not going to come back and say, "Hey,
7 I beat the test because I used counter-measures."
8 Q. How hard is it to make the counter-measures, first?
9 Start with the physical ones. Is it something that requires
10 practice?
11 A. The research suggests that people on their own don't do
12 a good job of spontaneously figuring out how to use
13 counter-measures to do this. So the research that's been
14 published has indicated that if people have no more than a
15 half hour of exposure to the theory of question development,
16 which is presented here on the slide -- there's two types of
17 questions and in order to pass the test you want to get a
18 big response for the control question. So the person needs
19 to be told that. They are taught what the control questions
20 actually look like so they can identify them and contrast
21 them to the relevant questions. Then they are told what to
22 do, like bite your tongue or subtract 7 from 239 serially as
23 fast as you can each time you do a control question.
24 When they are given less than a half hour instruction
25 that includes those elements, then they seem to be quite
55
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 successful. There are also people who have made available
2 on the internet and in other means similar types of
3 instructions about how to beat the polygraph. So, for
4 example, there's an extensive treatment of the subject on a
5 webpage called anti-polygraph.org which was prepared by
6 people who have taken polygraph tests and have the
7 scientific background to review the literature and learn a
8 great deal about it.
9 They have might not a manual that explains how to do
10 this but that includes some of the same elements in the
11 research. It's also the case that there is another webpage
12 called polygraph.com that you can go to. I'm not sure how
13 it works now, but up until recently if you give a Mastercard
14 number, that person will call you on the phone and coach you
15 on the phone regarding how to beat the polygraph test using
16 the same procedures.
17 Q. So the process by which the uninitiated learns the
18 difference between control and relevant questions, does that
19 involve basically figuring out whether it's a question about
20 the crime as opposed to one of those have-you-ever
21 questions?
22 A. Yes.
23 Q. So if they can tell that difference, is that enough to
24 spot the two kinds of questions?
25 A. Should be enough.
56
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 Q. Then they are told do nothing for the relevant, bite
2 your tongue on the control?
3 A. That's right. If they bother to read -- go to the
4 public library, get David Lykken's book, A Tremor in the
5 Blood, it has information that explains extensively how the
6 test is constructed and explains how counter-measures are
7 used. The work of Drs. Raskin and Honts, most of their work
8 tells us what we can actually say based on research
9 regarding how counter-measures work. In fact, if you read
10 their papers, you can understand what will and what won't
11 work in terms of the counter-measures, self-stimulation
12 exercises.
13 Q. So would it be fair to say people in the room may have
14 a pretty good start on the counter-measures?
15 A. One would think so -- hope so.
16 Q. Now, are there ways to detect counter-measures?
17 A. There are ways to detect counter-measures, specially
18 physical counter-measures. If you attach enough sensors to
19 a person, you can probably detect virtually every
20 counter-measure that they do. For instance, there are
21 counter-measures that involve manipulating muscle groups,
22 and you can record the activity of the muscles by putting
23 electrodes on them. So if you covered a person's body with
24 electrodes, you could do this -- even a person biting their
25 tongue, if you put electrodes on the muscles that control
57
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 tension in the jaw, you could detect if they did it when
2 they were asked the question. That would at least be
3 possible.
4 The mental counter-measures, of course, you can't
5 detect.
6 Q. Are you familiar with something called a movement
7 sensor?
8 A. I know that's something that some polygraph examiners
9 use, a sensor in a chair. I'm not quite sure how sensitive
10 that is, but doing something like pressing your toes on the
11 floor is an example of a counter-measure that probably
12 wouldn't work well if you have a chair sensor, but biting
13 your tongue shouldn't have any effect.
14 Q. You said that none of these are effective against the
15 mental counter-measures?
16 A. There's no way to detect the mental counter-measures
17 because we have no way of probing people's minds to see what
18 they are doing when they are answering the questions.
19 Q. Now, I think we have gone through all three of these
20 assumptions. In sum, do you think these are reasonable
21 assumptions on which to base the test?
22 A. I don't. I think there's so many ways it can go wrong,
23 that in the absence of the theory for why it should go
24 right, we should be suspicious.
25 Q. Let me ask you, you mentioned that the lack of theory
58
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 in the comparison or probable lie test has hampered research
2 and development. Is that true in the guilty knowledge test?
3 A. No, in fact it's the opposite in the guilty knowledge
4 test.
5 Q. Does the guilty knowledge test have a theory?
6 A. It does. The theory is based on recognition memory,
7 and the idea that when people are presented with a stimulus
8 that they recognize, it elicits a brain response that is
9 transmitted through the autonomic nerve system and can also
10 be recorded in the central nervous system.
11 One of the ways that the guilty knowledge test ends up
12 being developed is you can use what we know about how the
13 brain stores memory and how it's retrieved in recognition
14 memory tests to develop other tests that might work better
15 than autonomic measures, for instance, to detect a person's
16 choice to the multiple choice in the guilty knowledge test.
17 So in our work and other people's, we have been
18 mentioning EEG or brain wave activity when people are
19 exposed to the guilty knowledge items and what we can show
20 is people is people show a characteristic brain response
21 when they are shown an item, a word or a picture, flashed on
22 a computer screen that is something material. This is a
23 different frame response than what you see to the other
24 things presented to them on the computer screen at the same
25 time, which are things that are not stored in the memory.
59
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 In fact, one of the things that we have been able to do
2 by putting together what we know about brain mechanisms and
3 the theory of the guilty knowledge test and what we know
4 about the psychology of memory is to develop a built-in
5 control when we use the guilty knowledge test that makes it
6 difficult for people to defeat it. So, for instance, what
7 you would do is you would have a person see words from a
8 crime scene flashed on a computer screen and there would
9 also be words on the computer screen that have nothing to do
10 with anything that's in their memory.
11 Then there would be a third set of words, and these
12 would be words that we know are stored in the memory. What
13 you do is you tell the person to push a button every time
14 they see a word they recognize, and the words that we know
15 are stored in their memory, they will push a button "yes"
16 when they see that, otherwise push a different button "no"
17 in the other words, would elicit the same brain response
18 that we see to the guilty knowledge items flashed on the
19 screen.
20 So with this type of guilty knowledge test we have what
21 the brain's response to known memories looks like, what the
22 brain's response to stimuli that has no associated memory
23 with it, and then what does the response to the guilty
24 knowledge test look like, does it look more like the
25 response to the known memories or more like the response to
60
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 the words that have no memory meaning to the person. By
2 making that determination, you determine whether or not the
3 person has the memory or not, and, therefore, whether they
4 are involved in the crime or not.
5 Now, this procedure hasn't been progressed past
6 laboratory study. The point is because we understand we are
7 getting at information in people's memories and there is a
8 memory recognition response, we can use the theory coupled
9 with what we know about physiological recording to enhance
10 the procedure and make it more effective.
11 Q. The use of technology like EEG, was there a concern by
12 the NAS study that people using the control question have
13 not reached out to embrace some of the new technologies like
14 that?
15 A. Yes.
16 Q. Are you familiar with other people that are, for
17 example, using CAT scans, that sort of thing, in an effort
18 to explore the lie response that the brain physiologically
19 produces?
20 A. There are other people that are doing this, yes.
21 Q. Has any of that been linked up to the control question
22 test that we are talking about here in the research?
23 A. No.
24 Q. I wanted to ask you about evaluating polygraph testing.
25 Have you taken a look at the research on that? Have you
61
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 done some research on that?
2 A. Yes.
3 Q. Are there studies that have been published in the peer
4 scientific literature about polygraph accuracy?
5 A. Yes.
6 Q. Do you have any idea how many of those?
7 A. Well, in terms of studies that deal in particular with
8 accuracy, I'm not quite sure that I could categorize them
9 that way, but hundreds of studies published on polygraph,
10 and certainly many studies published on polygraph that
11 others have used to make claims about accuracy that probably
12 wasn't reflected in the original intent of why the study was
13 done or how it got published.
14 But there are some studies that are specifically
15 designed to tackle the accuracy question, and those are the
16 field studies where they use actual cases that have been
17 carried out to try to determine how well the polygraph
18 detects people who are guilty and innocent. For those
19 studies there's probably a dozen or so that have been done,
20 maybe 15 or something like that, not a lot.
21 Q. Do you think they are adequate -- is there adequate
22 peer review to establish an accuracy the rate?
23 A. I'm sorry.
24 Q. Is there adequate peer review to establish an accuracy
25 rate in real live testing?
62
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 A. I'm not quite sure --
2 Q. Do you feel the research is adequate to determine what
3 the accuracy of a field polygraph is?
4 A. No, I don't feel the research is adequate at all.
5 Q. Why is that?
6 A. Well, in the first place, I think it's important to
7 understand that peer review is always carried out in the
8 context of a question that a scientist wants to answer. And
9 whether the experiment or the design of the study allows one
10 to investigate this question and come up with a reasonable
11 conclusion.
12 A lot of the papers that are published in the peer
13 review literature are actually publications that raise
14 serious questions about the accuracy of polygraph tests. So
15 even though they have undergone the peer review process, the
16 conclusion they have is polygraph testing is not very
17 accurate.
18 A number of the studies that have gone through the --
19 that are published haven't gone through a scientific peer
20 review process. For instance, they get published in trade
21 journals run by police organizations or by the polygraph
22 organizations.
23 Q. For example?
24 A. Well, for example the Journal of Police Science
25 Administration, which is a journal that was operated by, I
63
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 think, the International Association of Chiefs of Police.
2 Or the Journal of Polygraph, which is a journal that is the
3 official publication of the American Polygraph Association.
4 These journals don't provide the type of peer review that
5 you would get if you publish in the top professional
6 psychological science journals, for instance.
7 And these studies then, as one consequence, have many
8 many flaws that make it very difficult to understand what to
9 make of the way the study was done or whether there's any
10 conclusion that you could make that would have any validity
11 whatsoever.
12 So the literature itself is quite variable. The
13 quality of the studies are very variable. A number of the
14 studies that are published in the very best journals that
15 actually have gone through probably the most competitive
16 peer review that are published show that polygraph tests
17 don't work. So it would be very difficult to use this
18 literature to say that the peer review process has produced
19 some sort of a consensus that polygraph tests are valid.
20 Q. I would like to ask you about what should be in a valid
21 study. I will show you State's Exhibit No. 2. Is this a
22 chart that you prepared for us today?
23 A. Yes.
24 Q. What is this?
25 A. These are requirements that should be met if you are
64
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 going to have a study that allows you to determine the
2 accuracy of a polygraph test like the control question test.
3 Q. Let's start with No. 1. What do you mean when you say
4 field study using real life cases?
5 A. There are two types of experiments or two types of
6 study, I guess. One is not really an experiment -- that you
7 can do. One is what's called the laboratory study or mock
8 crime study, and this is a study where the experimenter
9 controls everything and can assign people to groups based on
10 whether they are guilty or innocent. The person assigned to
11 the guilty group might be told to go down the hallway and
12 take an envelope out of the desk that contains $20. The
13 innocent person is told nothing, and then a half hour later
14 the people take the polygraph test to find out who took the
15 $20 and who didn't.
16 These laboratory studies are typically carried out not
17 so much to answer questions about how accurate polygraph
18 test is, but to answer questions about things that you
19 control in a laboratory environment. These would be things
20 like whether or not people have certain personality
21 characteristics produce different results on a proposal
22 graph test or whether drugs have certain effects, whether
23 different questions produce different results on a test.
24 Because the studies themselves don't involve the
25 motivational and emotional investment that people have when
65
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 they take a real life polygraph test, these laboratory
2 studies are seen as so dissimilar from real life that they
3 are seen as not really being useful for estimating the
4 accuracy of the polygraph in real life settings.
5 That's not just my opinion. The National Academies of
6 Sciences report is very clear on this point, and the
7 analysis of polygraph by the German scientist also reaches
8 the same conclusion about laboratory studies. The
9 laboratory studies are useful, but they don't get at the
10 psychological factors that come up in real life cases where
11 people may go to prison, may be prosecuted, may be released,
12 these sorts of things that govern how they take the test.
13 So you need a field study using real life cases where
14 people have actually taken tests as part of a criminal
15 investigation.
16 Q. What about No. 2? What is ground truth?
17 A. The advantage of the laboratory study is that people,
18 because they are assigned to a condition where they are then
19 asked to do something like take $20, we know who did what,
20 because we can tell them to do this and check to see whether
21 they did it. But in real life we don't know who did what.
22 So a real trick for doing field research has been how to
23 figure out ground truth, and there have been various
24 procedures that have been used. And most of them have been
25 found wanting in different ways. For instance, you could
66
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 use a judicial outcome to establish ground truth, but
2 judicial outcomes are affected by legal technicalities that
3 lead to exclusion of evidence that might incriminate a
4 person that is actually guilty, finding people guilty beyond
5 a reasonable doubt, etc.
6 So typically these have not been used. In fact, the
7 only procedure that's really been used that most people
8 think has much credibility is to rely on the confessions,
9 the idea being if a person confesses they are guilty. If
10 there are multiple suspects in the case, that confession
11 will clear others, and verify their innocence. So
12 confessions have been the criteria for establishing ground
13 truth.
14 Q. We will come back to that in a minute but let's finish
15 the list. Why is blind chart scoring important?
16 A. Blind chart scoring is important because the key
17 question, I think, that is before the Court is what does the
18 reporting of the physiological data add to the determination
19 of the person's truthfulness. I presume the Court wouldn't
20 be interested in having as a witness someone who says, "I
21 have been trained in truth detection and I have interviewed
22 this person and I reviewed the case facts and in my opinion
23 this person is lying," or, "In my opinion this person is
24 telling the truth." That's what the whole point of the
25 procedure of going through a trial is trying to determine
67
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 guilt. Of course, the polygraph is rendering an opinion
2 that's extraneous of the other information. So the key
3 element in court is what do the physiological traces tell us
4 that we don't already know from the other evidence
5 available.
6 The only way we would know that is if they blindly
7 score the charts. That is, have people or computers score
8 the charts with no reference to the case facts to see how
9 they came out. That way we will know what the physiological
10 data itself adds to the determination of guilt versus
11 innocence as opposed to the physiological data coupled with
12 the examiner's opinion.
13 Q. Is this something called value added in terms of your
14 analysis of what the Court should be doing?
15 A. It's not a psychological term, value added. I think
16 maybe business uses a term like that, but I think the
17 concept is there. What does the physiological data add over
18 and above what you have already got? In psychology, this is
19 called incremental validity, involves the idea that you
20 already have a tremendous amount of information. Now you
21 add this physiological data. How does that boost the
22 accuracy of the overall determination, say, of a trial, as a
23 consequence of having those data.
24 Q. No. 4, independent criteria for establishing ground
25 truth?
68
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 A. This has been the criterion that's essentially ruined
2 our ability to get at the accuracy of polygraph tests
3 because it's the way most studies have been done, and the
4 field studies, that is, that have used confessions is they
5 obtain the confession right after the person has taken the
6 polygraph test and the reason the confession is obtained is
7 because the person fails.
8 The problem here is that we would not get the
9 confession if the person didn't fail the test. And once the
10 person has confessed -- because we already know the person
11 failed the test. If we look at the test to see if it was
12 failed, even if you have a blind judge do this, it will
13 almost always be failed. So we need independent
14 confessions. An example would be someone takes a polygraph,
15 they don't confess, they are arrested two years later and at
16 that time they confess to the earlier crime that they took
17 the polygraph on. That would be an example of an
18 independent confession.
19 Q. Have you prepared a chart to give us an idea of what
20 you are talking about?
21 A. Yes.
22 Q. This is State's Exhibit 3, Your Honor. I will give you
23 a pointer. I will zoom in -- first let's get the title of
24 this. It is Why Confession-verified Cases Provide an
25 Overestimate of Polygraph Accuracy or Why Confessions are
69
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 Not Independent of Test Outcomes. Is that what the flow
2 chart shows us?
3 A. Yes.
4 Q. What are we looking at here?
5 A. Well, it's the procedure that would be followed to
6 determine whether or not a case is ever selected for field
7 validity study. When you do a study like this, you have to
8 identify your subjects, and the subjects are people who you
9 have a ground truth criterion on. What this flow chart
10 shows us is how, when you do one of the studies, the actual
11 steps one goes through to determine who is included and who
12 is excluded.
13 Q. Start there.
14 A. The polygraph test is administered. That's the first
15 step to selecting the test. There are two possible
16 outcomes. One is scored deceptive and one is scored
17 truthful. If you follow the arrow to scored truthful this
18 will not be selected for the field study. The reason it
19 won't be selected is if the person takes the polygraph test
20 and scored truthful, they pass, the examiner is not going to
21 try to get a confession from the person. This could be a
22 guilty person who has taken the test and passed, but this
23 guilty person who would then be an example of an error would
24 never be included in the field study because there would
25 never be a confession.
70
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 When they get up here, the guilty person that scored
2 truthful is unlikely to look at the examiner and say, "The
3 joke is on you, I actually did it." So these people would
4 escape detection. These would be errored. They would not
5 be used. So the cases that get selected require the testing
6 scored deceptive and then we now want to determine whether
7 or not a confession was obtained. If the test was scored
8 deceptive and there's no confession obtained, we move up on
9 this loop or branch, I should say, and that case is not
10 going to be used in the study either.
11 So in other words, every time where an innocent person
12 is scored deceptive, that person is not going to confess,
13 and all those mistakes where innocent people scored
14 deceptive will be excluded from the field study.
15 It's possible that occasionally an innocent person is
16 scored deceptive and then will be included anyway. When
17 that would happen, all that would do is inflate the accuracy
18 of the polygraph test.
19 Q. What do you mean?
20 A. Well, I mean it would be an error but it would go
21 undetected. We would never know, because the innocent
22 person admitted guilt. It would be a problem with the
23 criterion.
24 Q. You are talking about the people who for one reason or
25 another falsely confess?
71
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 A. They would falsely confess, but because they falsely
2 confess after the test was erroneously scored deceptive, we
3 would rate the polygraph as accurate even though it wasn't.
4 It would increase the accuracy of the field study, whereas
5 the other two tests are excluded and we never know about
6 them. They also increase the accuracy of the field study.
7 So what cases are included in the field study? Well,
8 they are cases where the polygraph test is administered,
9 scored deceptive. We get a confession, and now we select
10 this case for the field study. The field study asks this
11 question. "Was the test score deceptive?" The answer is
12 yes. It had to have been scored deceptive. Because if it
13 wasn't scored deceptive we wouldn't get the confession. So
14 the problem with field studies is we are caught in the loop.
15 The things that leave the loop are the things that would
16 allow us to provide an accurate estimate of how the field
17 study works but they are not included in the field studies.
18 Instead, we get deceptive, a confession. We blindly rescore
19 the test. Since the numerical scoring is highly reliable
20 they tend to get the same scores, so it's almost always the
21 case that the original scored deceptive will be rescored
22 deceptive. That's the problem with the studies.
23 Q. Don't the studies also include people who are found
24 truthful because someone else confessed to the same crime?
25 A. Yes. That's how you get truthful polygraphs included.
72
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 So that's not included in the flow chart, but an example
2 would be this: Subject One takes the polygraph test in a
3 case that has two suspects, and that person passes. Subject
4 Two takes a polygraph test and this person fails. That
5 Subject Two now gets into this loop, and Subject One would
6 now also be selected for the same field study because
7 Subject Two's confession would confirm Subject One's
8 truthful outcome of the polygraph test.
9 But Subject Two would never have been tested in a
10 typical polygraph situation if there could only be one
11 person who committed the crime. Subject Two would typically
12 not be tested unless Subject One passed. Because if Subject
13 One -- if only one person could commit the crime and Subject
14 One failed the test, there would be no reason to test
15 Subject Two.
16 What I am trying to explain here is Subject Two is
17 being included in the study dependent on the fact that --
18 I'm sorry, Subject One -- this is confusing so I am sorry if
19 I am making it more confusing by getting the terms mixed up.
20 Subject One has passed the test. And his test gets
21 included only because Subject Two took the test, was scored
22 deceptive and confessed. So Subject One's past test is also
23 dependent on Subject Two's confession, just like Subject
24 Two's confession is dependent on it.
25 So the problem here is both truthful and the deceptive
73
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 tests that are chosen in field studies that are based on
2 confessions obtained right after the polygraph test is given
3 are almost always going to be tested as the original
4 examiner administered correctly. All the opportunities to
5 find errors are systematically deleted or hidden in the way
6 the field studies are done.
7 Q. Has anybody else done this kind of analysis, come up
8 with similar conclusions?
9 A. Yes, the National Academies of Sciences report is very
10 clear that this is the problem with confession studies, and
11 that's why they don't put any weight in them and that's why
12 they point out, like I do, the studies overestimate the
13 accuracy of polygraph. And the German analysis of the
14 control question test reaches the same conclusion for the
15 same reason.
16 Q. Have you done actual research in the area?
17 A. I have.
18 Q. Can you describe that to us?
19 A. Yes. We did a study with the Royal Canadian Mounted
20 Police when I was a professor At the University of British
21 Columbia. We reviewed all of the cases that had been
22 processed by the Vancouver polygraph unit, which serves
23 essentially all of British Columbia over a certain time
24 period.
25 This amounted to about 400 cases total. Then we
74
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 attempted to establish ground truth for these people who are
2 involved in the 400 cases, and we did something that hasn't
3 been done in other field studies. What we did was we
4 examined confessions that were obtained as part of the
5 polygraph testing process and the post-test interview, but
6 we also went into the police precincts where the actual
7 files were kept for the investigation that was being
8 examined through the use of a polygraph test to find out
9 what happened after the polygraph test was given.
10 All the other field studies involved taking a file from
11 the polygraph unit itself, as best we can tell, and
12 determining whether or not the person confessed in the
13 post-test phase of the polygraph exam. What we did was we
14 didn't stop there. We went out -- this is years later after
15 the case has essentially been retired -- and reviewed the
16 actual case files that the detectives investigating the case
17 had assembled to determine ultimately how they were resolved
18 to see if we could find evidence of ground truth.
19 An example would be two years later after this crime,
20 someone was arrested for another crime with the similar MO,
21 and that person would confess to this crime that the people
22 took the polygraph test for.
23 Well, that confession would now establish these people
24 as innocent, and that confession, made by someone who never
25 even took the polygraph test, was totally independent of the
75
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 outcome of the polygraph test. So now we have a confession
2 that serves as a criterion as independent of the outcome of
3 the test. So that would be included in the independently
4 verified cases.
5 Likewise we had instances where a woman, for instance,
6 believed that something was stolen from her house and that
7 they tested a cleaning company that cleaned the house to see
8 if they had taken the object. It turned out the woman
9 learned that the item had not been stolen so that was an
10 example of something that verified that the people who took
11 the polygraph was innocent because no crime was committed
12 and that polygraph was independent of their having taken the
13 polygraph.
14 So we used this procedure to find a criterion data
15 related to guilt or innocence that was independent of the
16 outcome of the test. That's how the study was different.
17 Q. Did you have any trouble finding enough in the
18 different categories of error?
19 A. We did. We were able to find evidence of independent
20 confirmation for truthful people, that is people who didn't
21 fail the polygraph test. But we couldn't find, except in
22 one instance, independent confirmation of a person who
23 failed the polygraph test. This reflected in part the way
24 the police used polygraph tests.
25 Typically polygraph tests are used in such a way so
76
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 that an investigation has been underway, has proceeded and
2 reached a point where it seems like it's not going to be
3 easy to resolve. At that point the examiner, the detective
4 doing the investigation, will make a request to have the
5 person take a polygraph test. If they agree, they take one.
6 The examiner is hoping the person who takes the test will
7 fail and confess and that will solve the crime. But because
8 the examiner has done a fairly thorough examination, if the
9 person takes the test and fails but doesn't confess, then
10 the examiner typically doesn't investigate the case further.
11 So when we went to the precinct files for the cases
12 where people failed polygraph tests and reviewed all of the
13 evidence added after the person took the polygraph test, we
14 found there was very little information included in the
15 files because the detective pretty much stopped
16 investigating the case once the person failed the polygraph
17 test. But for people that had truthful outcomes, the
18 detective would say, "Well, the people passed, maybe I
19 should be looking in other places for evidence." The file
20 would be left open and continue to accumulate evidence for
21 ground truth.
22 Q. So the amount of people in the test that you were able
23 to find that passed when they actually failed, how many were
24 there?
25 A. The number of people who passed when they actually
77
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 should have failed?
2 Q. Right, guilty people who got a truthful response from
3 the test.
4 A. Well, in that study we weren't able to find guilty
5 people who got a truthful response because the people that
6 -- when people fail the test, they either didn't confess or
7 they confessed -- I'm not sure I understand the question.
8 But if a person took the test and failed it, when we then --
9 and confessed, when we rescored it we basically found that
10 98 percent of the time that rescored tests matched the
11 original failed tests that elicited the confession. So this
12 is exactly what we expect to find when we have this kind of
13 ground truth contamination.
14 Again, the flow chart indicates that all of the errors
15 that we could possibly include have been eliminated from the
16 study, and in the absence of any possibility of finding
17 errors we only found one. We have 98 percent accuracy but
18 that's a totally misleading figure. In fact, the whole
19 point of doing this study was to illustrate with real data
20 this particular problem to other people doing this kind of
21 research.
22 Q. What about the innocent people? How did they do in the
23 study, the ones where you could verify ground truth?
24 A. The innocent people was different because we did get
25 independent evidence of ground truth. Their accuracy rate
78
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 for them was about 57 percent.
2 Q. Now, you said that part of the problem is the way that
3 police or investigative agencies use the test as a last
4 resort?
5 A. Right.
6 Q. Why do you think that's the way they do it?
7 A. Well, I have talked to people who run police units to
8 ask them this question, and it has to do with the fact that
9 they see polygraph testing as a limited resource. It
10 involves an expenditure of manpower in the form of the
11 polygraph operator, and they want to preserve the option to
12 use it in special circumstances and not make it routine.
13 One commander of a police unit told me that he doesn't
14 want his detectives to become lazy, for instance, and have
15 all of the detectives bring people in and have them start
16 taking polygraph tests so there's actually a procedure in
17 place that requires the detective to carry the case through
18 a certain point before they schedule a polygraph
19 examination.
20 Q. Is that a written procedure that requires sort of
21 initial work before it can go to the polygrapher?
22 A. Yes. Not only is it a written procedure that I found
23 was used in various police departments as well as by the
24 federal government, it's also a procedure that's included in
25 some of the standards that are -- that exist on how the
79
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 police are supposed to conduct polygraph tests. So, for
2 instance, the standards promulgated by -- I'm not sure what
3 it's called but the American Association of Police
4 Polygraphers, includes the standards that polygraph tests
5 are not done until a thorough examination of the case has
6 been carried out.
7 And the Department of Defense embraces that same police
8 polygraph standard in their criminal investigations.
9 Q. You said there were problems with the way the field
10 studies were conducted. What can you tell us about the
11 quality of some of the studies?
12 A. Well, the quality is all over the place, but most of
13 them are very poor quality. They are not published at all,
14 or they are published in trade journals. They don't include
15 any information about ground truth. They select cases in
16 such a way as to bias the outcome one would get in terms of
17 the way the study is done. There are all sorts of problems
18 with them.
19 Q. Does the National Academies share your concern over the
20 quality of these?
21 A. The National Academies did. In fact, they pointed out
22 that the studies that they reviewed would be unlikely to be
23 the type of research that would ever get funded by the
24 National Science Foundation or the National Institute of
25 Health because it was such poor quality, but they noted that
80
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 some of it gets published anyway, and they thought that was
2 because the topic was so important that what little anybody
3 could do that was ever any good should be out in the public
4 domain so people could see it. But they were clearly
5 frustrated by the lack of quality research in the area.
6 Q. The NIH standards that you were talking about, are they
7 the ones that you have to meet in order to get the funding?
8 A. Yes.
9 Q. This issue you talked about where the only time you get
10 confirmation of a test is when someone confesses, does that
11 have any effect on the examiner's perception of reliability
12 of the machine of their test?
13 A. It definitely does. It's extremely reenforcing to
14 examiners to learn that someone they failed confesses
15 because it helps them believe that what they are doing is
16 valid and that the test actually works. But the problem for
17 the examiners is just the same as the problem for us doing
18 the field studies. When someone passes a test and doesn't
19 confess, the when are has no way of knowing if that's an
20 error. When someone fails the test and doesn't confess, the
21 examiner has no way of knowing that's an error. So guilty
22 people who pass don't get detected and innocent people who
23 fail don't get detected but the examiner gets a great deal
24 of reinforcement from the guilty people who fail and then
25 confess. That lets the examiner know that he is doing
81
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 something right, and the examiners are trained to believe
2 that polygraph tests work, and then they have this type of
3 reinforcement that informs them of their successes but makes
4 it impossible for them to learn of their errors, and as a
5 consequence they believe strongly in the technique. They
6 don't see errors.
7 When Minnesota was considering banning polygraph one
8 examiner testified before the Minnesota legislature he had
9 given 20,000 polygraph tests and never once shown to be
10 wrong. You could say that sounds farfetched but it really
11 isn't, because in the context of how polygraph tests are
12 given, examiners learn when they are right but they don't
13 get the opportunity to learn when they are wrong, because
14 the mistakes escape detection.
15 Q. Now, this analysis that we have gone through with the
16 problems with confession, using confession cases to
17 establish field accuracy, you published on that area?
18 A. Yes.
19 Q. And that's been adopted more or less in toto by the
20 National Academies?
21 A. I don't know what you mean by adopted.
22 Q. Your analysis went through there?
23 A. They reached the same conclusion.
24 Q. They used the same fairly kind of example that we saw
25 where you follow the people in the various categories to see
82
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 where they end up in the study?
2 A. They have a similar type of example, yes.
3 Q. Now, if, in fact, there's a confession and someone goes
4 and finds physical evidence, does that give you independent
5 verification?
6 A. No, that's no different than having the confession
7 alone, except for one thing, and that is that if you have a
8 confession that then leads to physical evidence, you can be
9 certain -- at least more certain -- that the confession was
10 valid and you don't have an innocent person confessing to a
11 crime they didn't do. But the reason that having physical
12 evidence doesn't help with this problem is that you can't
13 get to the physical evidence without having the confession,
14 and you can't get the confession without the person failing
15 the polygraph test.
16 So to use the physical evidence to confirm the failed
17 polygraph test is no different than to use the confession to
18 do it. To marry the confession with the physical evidence
19 and say the standard is higher, it's marginally higher, but
20 even in the physical evidence the errors have been excluded
21 from sight.
22 Q. Let me ask you about this. These requirements that we
23 went through in Exhibit 2, are these unique to the
24 evaluation of polygraph accuracy?
25 A. No, these are the requirements that you use for
83
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 evaluating any type of test. For instance, if you had a
2 test that was designed to identify mental illness, you would
3 want to study its effectiveness by giving it to people who
4 have real mental illness, not college students who are
5 feigning mental illness, for instance. Likewise, you
6 wouldn't want the person evaluating the effectiveness of the
7 test to be the same person as the one who determines whether
8 or not the person has a mental illness, because if they are
9 doing the two tests simultaneously they are likely to lose
10 their objectivity and inflate the association between the
11 two.
12 So in the development of a new test like that, you need
13 the same type of blind evaluation we are talking about for
14 polygraph tests.
15 Q. Are there studies that verify these criteria?
16 A. No. The exception would be our study for the Royal
17 Canadian Mounted Police but only as it relates to the
18 truthful outcomes. Even there the sample is fairly small.
19 Q. That's because of the selection process?
20 A. It's because of the selection process.
21 Q. Is that unusual in these studies where you start out
22 with a large number of tests and there's very few where you
23 can establish ground truth?
24 A. It's not unusual. That's another problem with the
25 confession studies that we haven't gotten into, but it
84
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 involves the fact that people who confess to polygraph tests
2 are not necessarily the same as people who don't. So when
3 you go through this long process -- like we started out with
4 400 cases in the RCMP study. Even using the cases where
5 people confessed right after taking the polygraph test that
6 they failed, we still only ended up with about 75 individual
7 subjects out of 400 cases. So then you are left with the
8 question how would that group of 75 now generalize to the
9 other hundreds and hundreds of people that took polygraph
10 test that you have no ground truth criterion on.
11 So even working with confessions you have a problem
12 with only a select number of people are going to confess,
13 and because of the problem we talked about, the confession
14 studies are further compromised by the fact that they
15 themselves are dependent on the outcome of the polygraph
16 test.
17 Now when you go to get independent criteria from police
18 files, you are talking about a teeny fraction of the cases
19 they have to work with, so you need a lot of resources and a
20 lot of patience to go through lots and lots of material,
21 most of which is discarded in finding the tiny number of
22 cases you can actually use.
23 Q. Is it possible, given the studies that you have seen,
24 to derive an unequivocal accuracy of the polygraph tests?
25 A. No.
85
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 Q. Why not?
2 A. It's not possible because no study meets the criteria
3 you need to satisfy to have an acceptable field study and
4 the laboratory studies don't do the trick.
5 Q. Do you know what the National Academies concluded about
6 the same question?
7 A. They have the same conclusion.
8 Q. I would like to show you that and ask if you agree with
9 them in the way they phrased this. I am looking here at the
10 section down below. The part that I have highlighted, does
11 that express your opinion?
12 A. It certainly does.
13 Q. Do you know what the Germans thought about that?
14 A. They reached the same conclusion.
15 MR. COX: I'm sorry, that was Page 115 for the
16 record.
17 A. In the German study they went so far as to suggest that
18 people haven't really tried seriously to get an answer to
19 the question, given the way the studies have been conducted.
20 Q. And the conclusions that they reached and the Germans
21 reached, are these also confirmed by your individual
22 research?
23 A. Yes.
24 Q. Now, we talked a little bit about the quality of the
25 publications. Are all peer review publications equal?
86
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 A. No, they are not.
2 Q. How can we tell whether a journal is one of the top
3 echelon journals or one of the lesser journals?
4 A. Well, there are several criteria. The top journals,
5 for one thing, are difficult to publish in. They have a
6 very competitive peer review process, so typically they have
7 multiple evaluators who anonymously and independently review
8 your work, and typically most of the papers submitted to
9 such a journal are actually rejected, so rejection rates of
10 85 percent, for instance, would be common for the top
11 psychology journals.
12 Another thing that's used to consider the quality of
13 the journal is whether or not the articles that are ever
14 published in the journal are actually cited and there's an
15 organization that actually publishes the library reference
16 system called citation index, and there's a science citation
17 index and social science citation index. Any university
18 reference library has access to this. It's online.
19 What you can do is determine when articles are actually
20 published in journals whether or not anyone cites them, the
21 idea being that the best journals are going to publish
22 research that draws a great deal of attention, and that
23 other scientists will find worthwhile and important and,
24 therefore, will cite. So each journal has citation impact
25 factors that indicates how often articles that are published
87
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 in it are actually read and cited by other scientists. So
2 those two factors are two objective indices of journal
3 quality.
4 Q. The Journal of Applied Psychology, is that where the
5 study came from?
6 A. Yes.
7 Q. How does that follow in the citation frequency?
8 A. It falls within the top echelon of psychology journals
9 in the citation index.
10 Q. You also consider it one of the top journals?
11 A. Yes, it is one of the top journals for people doing
12 polygraph research. If you have a good study, that's
13 probably the journal you would aspire to get it published in
14 because it has the toughest standards.
15 Q. What about Psychophysiology?
16 A. Psychophysiology is a similar journal, high impact
17 factor and high rejection rate. But the audience for
18 Psychophysiology is a little bit different. The Journal of
19 Applied Psychology, as you might guess from its title, is a
20 journal that's a little bit broader and it goes to people
21 who are interested in applications of psychology or
22 psychophysiology. Goes to people who are more specifically
23 interested in the psychophysiological measures.
24 Q. How would Psychophysiology compare with the Journal of
25 Applied Psychology?
88
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 A. Big difference. Psychophysiology has a citation impact
2 factor more than 20 times greater than the Journal of
3 Applied Psychology.
4 Q. Cited 20 times more often?
5 A. The typical article, the way they make the calculation,
6 is cited 20 times more often.
7 Q. We talked about problems with the field studies related
8 to using confessions. Are there other issues that make
9 field studies less reliable when we are talking about
10 accuracy?
11 A. Yes.
12 Q. What are some of those issues?
13 A. One issue we have already talked about is
14 counter-measures. The field study research that's been done
15 has been based on cases that have been accumulated before
16 the information on how to use counter-measures has become
17 widely available. So if we were to use, for instance, the
18 RCMP study that I did, these cases came from the '80s, and
19 other field studies involved cases from the '80s and even
20 the '70s. And the information -- for instance, the web
21 based evidence available on how to beat polygraph tests and
22 the publications on the topic and the research that's shown
23 that it might be effective, these things have essentially
24 merged in the '90s and the last few years.
25 So it's more likely now that counter-measures could be
89
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 used to defeat a test than it would have been, say, five
2 years ago, ten years ago or fifteen years ago. So if we
3 were going to try somehow and use the existing research to
4 estimate the accuracy of the polygraph testing we would need
5 to qualify that by the fact that data were collected before
6 information on counter-measures had become more available.
7 Q. Is there also an issue concerning a test called the
8 directed lie?
9 A. Well, the directed lie test is -- to the extent it's an
10 issue, it's an issue because it's a different test. There
11 is no field study that I am aware of that has ever included
12 the directed lie test. There's one field study that was
13 done that included a directed lie question as part of a
14 test, but that's different than having a whole directed lie
15 test. So if you were going to use these field studies to
16 estimate the accuracy of anything, we could estimate the
17 accuracy of a control question test but not any other type
18 of test, not a directed lie test.
19 Q. For example, it wouldn't be appropriate to use your
20 study and try to apply that to the directed lie?
21 A. No.
22 Q. What is the effect of base rates --
23 THE COURT: Mr. Cox, let's break for lunch.
24 That's going to be a whole new area, I suspect. 1:30. We
25 are in recess.
90
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 (Note: Court stood in recess at 12:00 to 1:30).
2 MR. COX: Your Honor, for planing purposes, if it
3 turns out we are close to 5:00 and have to go over, is that
4 a possibility?
5 (Note: A discussion was held off the record).
6 Q. When we left, I just asked the question about base
7 rate. What is base rate?
8 A. Base rate refers to the proportion of people in a
9 population. You have a certain characteristic. So, for
10 instance, it could be the base rate of guilty that we would
11 be concerned about, and it might vary. If we took a random
12 sample, not many would be guilty of an undetected criminal
13 offense versus people being investigated by police. They
14 have a higher base rate of guilt, it would be a higher
15 percentage versus people in prison. Presumably the base
16 rate of guilt rate in prison is high, hopefully approaching
17 100 percent, for instance.
18 Q. Is that related in any way in terms of accuracy or the
19 kind of confidence we put in the outcome of the polygraph
20 test?
21 A. It is, because tests work best in terms of classifying
22 people when the base rate of the characteristic you are
23 interested in, in this case guilt, is 50 percent. So the
24 closer the base rate gets to zero, the more errors you are
25 going to make where innocent people are diagnosed as
91
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 deceptive, and the closer the base rate gets to 100 percent,
2 the more errors you will have where guilty people are
3 diagnosed as truthful.
4 Q. Now, I would like you to take a look at State's Exhibit
5 4. It's called The Problem With Base Rates. Is this
6 another one you created for us?
7 A. Yes.
8 Q. Could you go through this and use the pointer if you
9 need to explain, in kind of a practical scenario, what you
10 mean by base rate affecting the confidence in the outcome?
11 A. Okay. Well, this is just a hypothetical example. It
12 requires, to understand it, at least three assumptions
13 listed up here. So the first is that 90 percent of those
14 that have hired an attorney and come before the Court are,
15 in fact, guilty. So these would be people that come before
16 the Court, say, with a past polygraph test.
17 The second is that the polygraph test is 90 percent
18 accurate. And the third one is that we have 100 people who
19 come before the Court having taken a polygraph. And the
20 flow chart just shows how the Court would see the outcome of
21 tests given this way in terms of the percentage of errors
22 that would be associated with truthful or past polygraph
23 tests that would come before it.
24 So if you start -- see, where is this thing? Start
25 here. Just hypothetically to make the example easy to
92
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 follow, we have 100 people. Because the base rate is 90
2 percent, that means 10 of them will be innocent and 90 of
3 them will be guilty. Because the polygraph test is 90
4 percent accurate, for the ten innocent people, one would be
5 diagnosed deceptive and nine would be found truthful.
6 Again, the polygraph test is 90 percent accurate so among
7 the guilty, nine of them would pass and 81 of them would be
8 found deceptive.
9 Now, in terms of cases that come before the Court, the
10 81 deceptive cases will never come before the Court, because
11 no one would want to introduce as evidence on their behalf
12 that they failed. The one innocent person that failed
13 wouldn't come before the Court either. So the only ones
14 that the Court would see under the circumstances are the
15 nine truthful who are actually from the innocent group, so
16 these are nine truthful innocent people, and the Court would
17 see the nine truthful guilty people.
18 So 18 truthful people come before the Court, but 50
19 percent of the guilty polygraphs are from guilty people. So
20 even though the polygraph test is 90 percent accurate,
21 because the base rate is so high, getting close to 100
22 percent, there's going to be a disproportionate number of
23 guilty people who pass. Therefore, when the truthful tests
24 come before the Court, even though it's 90 percent accurate,
25 50 percent of the results will be in error.
93
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 Q. I would like to show you another chart which wasn't
2 prepared by you. Are you familiar with Dr. Raskin, The
3 Polygraph in 1986 Scientific Profession and Legal Issues
4 Surrounding Application and Acceptance of Polygraph
5 Evidence?
6 A. Yes.
7 Q. Utah Law Review article. I would like to show you a
8 chart out of that book. Does that look familiar?
9 A. Yes.
10 Q. I believe for the record that's from Page 56. Just to
11 illustrate how this change affects things, can you tell us
12 what we are looking at here? Let's start on the left side.
13 This column rate, the base rate of guilt, how do you read
14 the chart? Can you give us an idea of how it works?
15 A. Actually, I haven't seen the chart, so let me look at
16 it for a second. But the base rate of guilt that would
17 correspond to what I am talking about is right here at 90
18 percent. And then looks to me like there's two sets of
19 statistics here. One is estimating the accuracy from
20 laboratory studies where at this point in 1986 apparently
21 Dr. Raskin was making the argument that 97 percent of the
22 people in laboratory studies who were guilty were accurately
23 detected and 92 percent of the innocent were accurately
24 detected.
25 Over here we have a different estimate that comes from
94
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 the Office of Technology Assessment report, which was -- I
2 think it came out in 1983, and here the assumption would be
3 that 9 percent of the people would be guilty, and 80 percent
4 would be innocent.
5 And then under these circumstances, if the test is this
6 accurate and you give it to people, then this is the actual
7 rates which you would accurately classify people apparently
8 as guilty or innocent.
9 I am actually not sure -- you know, I am -- not having
10 looked at this and not looking at the caption that explains
11 how the table works, I'm not sure exactly that I could
12 confidently say what the columns of figures are here.
13 Q. Well, in general, let's look at it this way. As the
14 base rate of guilt falls, does that affect the rate of
15 deception or confidence in the deceptive outcome? Without
16 even thinking about his chart, let's talk about it that way.
17 In other words, at 90 percent rate of guilt in the base
18 rate, you indicated that it was about a 50/50 proposition?
19 A. The problem is I'm not sure what the numbers mean,
20 because it's sort of in the opposite direction. When the
21 base rate of guilt is high, most of the errors are going to
22 be with guilty people. So I'm not sure if deceptive means
23 guilty or deceptive means the outcome of the polygraph test
24 given to some number of people. I'm not quite sure what it
25 means.
95
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 Q. We will ask Dr. Raskin about that. Now, in general
2 though, is it necessary to have a base rate, at least to
3 know a base rate if you are going to express confidence in
4 the outcome of the test result?
5 A. Yes, it is. Because it's not just the accuracy of the
6 test, the known error that's important. It's the known
7 error as it pertains to the likelihood of you actually
8 finding people who are guilty given the rate at which they
9 occur in the population of the study. So, for instance, in
10 the National Academies of Sciences report, one of the things
11 that they are worried about is even if the test was highly
12 accurate, if you give it in the employment screening
13 connection where hardly anyone is guilty, most of the errors
14 will be innocent people who failed.
15 In fact, because there are so few guilty people
16 presumably that have national security clearances, you might
17 detect virtually nothing but innocent people who failed.
18 But in a context where people have been accused of a crime,
19 there's evidence against them, there's an indictment, they
20 have hired a defense attorney, one would expect the rate of
21 guilt to be high. The exact number, I don't know what it
22 would be, but it's certainly higher than it would be in a
23 random group of people selected off the street, higher than
24 it would be for people just being investigated by the
25 police, lower than people already adjudicated as guilty and
96
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 in prison, but if you know that number and it's high and
2 it's something like 90 percent, and the polygraph test was
3 90 percent, then the flow chart you presented would indicate
4 how the past tests that come before the Court, what
5 proportion come for innocent and what proportion come for
6 guilty. That number would be the errors. If you change
7 either of the two numbers, either the accuracy of the test
8 or the base rate of guilt, the numbers would vary slightly.
9 But the test would work the way you think it was if the
10 base rate was 50 percent. In other words, then it would be
11 a balance between errors between guilty and innocent people,
12 and the test would classify people in a way you would feel
13 comfortable with. The bottom line is you shouldn't use a
14 test when the base rate deviates markedly from 50 percent
15 because you will be generating mostly errors and not knowing
16 it, even if the test is really accurate.
17 Q. In your experience does the base rate explain why?
18 A. No.
19 Q. Is it known, is it accepted --
20 A. It's ignored.
21 THE COURT: I want to make sure I am clear on
22 that. If the base rate is 50 percent, with your chart in
23 State's Exhibit 4, the base rate is 50 percent meaning 50
24 percent of those folks who are accused and hired an attorney
25 are in effect really guilty?
97
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 THE WITNESS: Uh-huh.
2 THE COURT: And if we assume a 90 percent accuracy
3 rate for our purposes, then, therefore, 90 percent of the
4 folks who PASS the polygraph test saying, "I successfully
5 passed" would, in fact, be truthful? Assuming that 90
6 percent is an accurate figure.
7 THE WITNESS: One way you could look at it is if
8 you had fifty people who came before the Court who had
9 passed and the test was 90 percent accurate, five of them
10 would be guilty and 45 of them would be innocent. Okay? So
11 in other words, that's exactly what you would expect because
12 the test is 90 percent accurate, that the 90 percent figure
13 is now reflected in the people who actually show up in
14 court. Then as the figure approaches -- I'm sorry, that's
15 50 percent base rate. As the base rate rises, more and more
16 guilty people who pass would come forward.
17 THE COURT: Of the things I was able to follow in
18 the articles, base rate was one that I was not sure I was
19 following so that helps a lot. I appreciate it.
20 Q. (By Mr. Cox) Now, among other issues that you
21 mentioned that must be taken into account considering
22 accuracy is something called the friendly polygrapher. What
23 is that?
24 A. This refers to the fact again that the research that's
25 been done on polygraph test in the sense of field studies
98
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 has been done with adversarial tests. These are tests
2 administered by police officers, and the investigation of a
3 crime.
4 In those circumstances, when a person takes a test,
5 they are likely to have fear of the consequences of being
6 found guilty because the person, if they fail, the results
7 will be known to the police, known to the prosecution.
8 THE COURT: Could I interrupt? I'm sorry. I have
9 another base rate question. We talked about a base rate of
10 90 percent, assuming 90 percent guilty. We talked about an
11 assumption that the polygraph test is 90 percent accurate.
12 My understanding, from reading your articles and from
13 reading the materials in the National Academies of Sciences,
14 their publication, is a lot of folks don't necessarily go
15 along with the figure, but in the National Academies of
16 Sciences, as I recall, they agree -- if you look at the best
17 case that chances are a little better than chance that it's
18 going to be accurate --
19 THE WITNESS: Uh-huh.
20 THE COURT: Let's say it's a 55 percent accuracy,
21 just for argument sake. That's the working assumption. You
22 have a base rate of 90 percent. What does that do to the
23 liability of the polygraph examination that the factfinder
24 would be likely to see?
25 THE WITNESS: You would see nothing but guilty
99
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 people who passed polygraph tests. It would be so skewed in
2 that direction that you would see hardly any people from
3 truly innocent people that passed because so many of the
4 people they even bother to test at all are guilty. You
5 would see almost nothing but guilty people who passed.
6 THE COURT: I mean, in your example of the 90
7 percent base, rate 90 percent accuracy, when you ran through
8 the hypothetical with 100 test-takers, it worked out to
9 about 50 percent of the ones you would see were basically
10 lying and not caught. They were guilty and not being found
11 deceptive.
12 THE WITNESS: Right.
13 THE COURT: What would that -- how would that -- I
14 haven't had a chance to work out the number or do a chart,
15 but how would that number be affected with the 90 percent
16 and say 55 percent accuracy or 60 percent accuracy?
17 THE WITNESS: I made a few calculations, but
18 suppose the 90 percent accuracy dropped to 80 percent. It
19 would turn out that 67 percent of the cases that came before
20 the Court would be guilty people who passed. If the 97
21 percent dropped to 70 percent, then it would be 75 percent
22 of the cases that came before the Court would be in error.
23 So as you keep dropping down, you would have a larger and
24 larger fraction of cases coming before the Court that would
25 be errors.
100
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 THE COURT: Thank you.
2 MR. COX: Does that make things clearer (handing
3 witness document)?
4 THE WITNESS: That's actually about something else.
5 Q. (By Mr. Cox) This is a different change?
6 A. That's a different change but it would produce the same
7 result, in terms of the percentages.
8 Q. You started talking about the friendly polygrapher and
9 how that might affect accuracy?
10 A. Right. So as I mentioned, the field studies are based
11 on adversarial tests but when tests are admitted by defense
12 attorneys they are not adversarial tests, they are arranged
13 through attorney/client privilege. In these situations the
14 psychological factors that are likely to be at work when an
15 adversarial test is administered are going to be different.
16 Because when a defendant fails the polygraph, friendly
17 polygraph test, no one is going to know. It's protected by
18 attorney/client privilege. So the only test that would come
19 forward would be the ones passed in the first place.
20 When a police officer gives a test, he is going after a
21 person and maybe not sure how it's going to come out, but
22 certainly if it comes out that the person has failed, he is
23 trying to solve a crime and that's an expectation he might
24 have, that the person might be guilty and I will get him
25 with the test.
101
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 Whereas when a friendly polygraph test is conducted,
2 the issue is not to capture someone or take a suspect and
3 turn them into someone who has confessed or something like
4 that. The issue is to determine whether or not they can
5 pass the test that the person taking the test has hired them
6 to give or who has been hired on behalf of the defense
7 attorney, at least, where the hope is that the person would
8 pass.
9 So this sort of heated adversarial atmosphere that
10 exists in the field studies where the police give tests
11 doesn't exist with the friendly polygraph test.
12 Particularly the fear of the consequences of being found
13 guilty isn't going to be the same under the circumstances.
14 Another difference relates to the Rosenthal effect we
15 mentioned earlier, because the expectation in the friendly
16 polygraph test could lead the examiner who is giving the
17 test to, again not deliberately, but to unintentionally bias
18 the outcome in such a way that the person would actually
19 pass the test.
20 Q. Is there any way to know what effect this might have on
21 the accuracy of that?
22 A. Well, the presumption would be that it would make it
23 more likely for people to pass, but there's no literature,
24 no studies that have examined the friendly polygrapher per
25 se, so we don't know for sure what the effect would be.
102
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 Q. Are there any field studies which sought to compare
2 ones from the police department with the friendly
3 polygrapher series?
4 A. No, that's the problem with, I think, accepting
5 friendly polygraph tests is that the research literature is
6 on adversarial tests so we have some idea even with the
7 flaws the studies have how they work. With friendly tests
8 we have no idea how they work and we can question on
9 theoretical grounds that they would work the same way.
10 Q. Can't we just use laboratory studies to determine that?
11 A. You can't use the existing laboratory studies to
12 determine that, because the existing laboratory studies are
13 different in other ways from both adversarial to friendly
14 tests.
15 For instance, in a laboratory study, the people who are
16 innocent are basically playing a game. They didn't take the
17 $20 that was the alleged crime, and was is likely to be most
18 disturbing to them is the control questions, so there's no
19 real reason for them to be concerned about the relevant
20 questions. The people who are guilty are not really
21 deceptive in the way they are in real life because they are
22 told to act the behavior. They know that the people running
23 the study know whether they did it or not. One of the
24 factors that triggers autonomic responses is an unordinary
25 response. So in a mock study where you have stolen $20 and
103
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 you take a laboratory test, you are responding to the
2 relevant questions because you know those are important.
3 There's an expectation built in, you have the $20 in your
4 pocket, these are the questions important to me, so you
5 respond to those questions possibly in the same way you show
6 a recognition on a guilty knowledge test instead of because
7 you are lying when you answer the question.
8 The problem is that because we don't have a clear
9 theory of what it is that is essential to the operation of a
10 polygraph test, when you go across the different contexts of
11 adversarial test to friendly test to laboratory procedure,
12 we have no way of knowing to what extent the psychological
13 processes that a person goes through were the same or
14 different. So it's very difficult to go from one to the
15 other and make generalizations.
16 Q. You also mentioned that the competence of the
17 polygrapher should be taken into account when looking at
18 accuracy?
19 A. Yes, the competence of the polygrapher involves the
20 fact that polygraph examiners don't actually receive a stamp
21 of approval from a testing organization, for instance, who
22 says this polygraph operator is in general 87 percent
23 accurate. What we can say about polygraph operators is that
24 they practice their profession ethically, do continuing
25 education, have had training at such and such a school and
104
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 what have you, but there's likely to be quite a bit of
2 variability in examiners at how good they are in carrying
3 out the procedures.
4 In any given case, we have no idea in that particular
5 case what accuracy to assign to the particular examiner.
6 That's independent of the problems related to
7 standardization and objectivity and what have you. It's
8 just examiner's skills will vary quite a bit from one person
9 to the next. Because there's no way of knowing who the
10 really great examiners are versus the ones who aren't, you
11 don't have a way of knowing in a particular case whether
12 this particular test administered by this particular
13 examiner is one that you should put a certain amount of
14 weight in.
15 Q. In general, given the complex nature of the testing
16 that you described and the psychological manipulation that's
17 required for success, do you think eight weeks of school for
18 somebody with, say, a high school education, would be
19 sufficient to understand and apply the various concepts that
20 we have talked about?
21 A. No. In particular, I don't believe they have any
22 opportunity to really understand the psychological basis of
23 what they do.
24 Q. Let's say that they spent 44 hours during that eight
25 weeks studying psychology, psychophysiology. Is that enough
105
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 to educate them so that they can make the kind of choices
2 that we've been talking about that have to be made through
3 the process?
4 A. No.
5 Q. Now, in some of the standards and the rules that we
6 have seen, the examiners also are required to determine
7 suitability of examinees in terms of psychological, medical
8 drug issues, etc. etc. Is that kind of training going to
9 give them the basis they need to make decisions like that?
10 A. No.
11 Q. Why is that?
12 A. For one thing, we don't know for sure how to use the
13 information they even collect. Then they are not skilled or
14 trained with enough background to know what the information
15 they collect means. In other words, if people say they have
16 certain types of mental disorders, they have certain
17 neurological problems where they take a medication, it's not
18 necessarily something the polygraph examiner would
19 understand or know anything about.
20 Q. How would you contrast that kind of training, when you
21 describe the eight weeks, with the kind of training that,
22 for example, is necessary for someone to give an MMPI and
23 interpret the Minnesota Multi-Phasic Personality Inventory
24 test?
25 A. To give the MMPI in Minnesota and interpret it you have
106
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 to be a licensed psychologist, which would require a Ph.D.
2 Degree with at least two years of full-time clinical
3 supervised work and, you know, specific training in test
4 construction and administration and interpretation of the
5 particular test.
6 Q. Have there been major advances in the polygraph
7 recording technology that you are aware of?
8 A. There have been changes in the equipment that's used,
9 so that, for instance, they are more likely now to
10 computerize and digitalize the data. Software is used to
11 score it and some of the recording apparatus is different in
12 the sense that now it attempts to use more sophisticated
13 electronics than it used to, say 20 years ago.
14 Q. Any difference in the quality of the information
15 captured by the cuff or the band around the chest and that
16 sort of thing?
17 A. No, there isn't. We actually did a study of this
18 sometime ago where we used -- we attached people to very
19 expensive state-of-the-art psychophysiology laboratory
20 polygraph that uses the best instrumentation available. We
21 attached sensors to the person using that polygraph and also
22 the sensors from a conventional field polygraph at the same
23 time, and that particular field polygraph, for instance, two
24 of the channels that we recorded which were the respiratory
25 channel and the cardiovascular channel on the equipment
107
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 didn't even require an electronic sensor. In other words,
2 if you unplug the equipment, you could still measure
3 respiratory activity and blood pressure because those older
4 polygraphs used mechanical means to record those signals.
5 Yet when we examine how those signals from the more
6 primitive instrument compared to the state-of-the-art
7 laboratory machine, the correspondence was remarkably good.
8 So even though the electronics of the polygraph equipment
9 have improved, the fact is the old equipment wasn't bad at
10 doing what it was supposed to do.
11 Just like the computer programs are good, the computer
12 programs still don't outperform numerical scoring by highly
13 trained examiners. So although the advances have improved
14 the equipment and the standardization of the scoring, they
15 really haven't changed anything otherwise over the last
16 couple decades because these were not where the problems
17 were. Twenty years ago the equipment still provided a good
18 recording of the cardiovascular activity, the respiratory
19 and the palmar sweating. The equipment was good in the
20 sense that examiners well-trained provided consistent
21 results. So those things have been improved but they were
22 good to begin with.
23 Q. Does the computer scoring make it more accurate?
24 A. I don't believe so. It standardizes it. It might
25 remove a little bit of inconsistency but the problem with
108
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 polygraph tests goes into the fundamental three assumptions
2 that we talked about earlier about how the questions are
3 formulated and what factors lead a person to respond more to
4 one question than the other. Computer scoring and better
5 instrumentation don't really have any effect on the
6 fundamental fact that the psychological foundation of the
7 procedure is weak.
8 Q. I want to ask you about standards in the polygraph
9 field. I would like to show you State's Exhibit 11. Did
10 you have a chance to review that?
11 A. Yes.
12 MR. COX: I have given a copy of State's Exhibit
13 11 to the defense, Your Honor. This is the same standards
14 that are scattered throughout the materials provided to the
15 Court. I put them in one volume.
16 THE COURT: Okay.
17 MR. COX: I have a copy for the Court. I'm not
18 sure we will go into that detail. The only addition, Your
19 Honor, is I have added Rule 11-707 of the New Mexico
20 statutes and the New Mexico rules on polygraph.
21 Q. Have you reviewed these in preparation for today?
22 A. Yes.
23 Q. I'm not going to go through them item by item but we
24 have, I believe, at least ten different kinds of standards
25 here?
109
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 A. Yes.
2 Q. Did you look at all of them?
3 A. Yes.
4 Q. Is there anything in these that sets a standard for
5 even the technique that must be used in terms of whether you
6 use relevant, irrelevant, guilty knowledge or comparison
7 question?
8 A. I couldn't find it. It's not that specific in any of
9 the standards that I read. The closest you could come to
10 something like that would be statements to the effect that
11 they use procedures that are commonly accepted in the field,
12 something like that. But the specific procedures -- for
13 instance, how a test should be constructed, what elements it
14 should have, how many questions it should have, whether you
15 should even use control questions, whether the control
16 question test should be used, it's not that specific. None
17 of the standards that I read through pointed out details at
18 that level.
19 Q. So you could use, for example, the guilty knowledge
20 test as opposed to the comparison question test under some
21 of the standards?
22 A. As near as I can tell, you could. I don't see any
23 differentiation in them. I think some of the standards
24 mention, in particular, that you have to have relevant
25 questions. So the concept of the relevant question being
110
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 the important element of the test is brought up, but I
2 couldn't find anything in them about the need to have other
3 types of questions or how many you would need to have.
4 Q. Are there any standards about control questions, when
5 they are used or how they are formulated?
6 A. I couldn't find them, other than, you know, the general
7 comment that you should use practices that are embraced by
8 the field, but nothing specific about how to use them, when
9 to use them.
10 Q. What about the issue of when somebody is suitable and
11 when somebody is not suitable for a test? Any standards you
12 could find that would tell the polygrapher based on what you
13 heard, this person should not be tested?
14 A. No, the way the standards are tested they have
15 statements like this: "Don't test unsuitable people" as
16 opposed to saying "someone would be unsuitable for testing
17 based on the following."
18 Q. Is that true for medical issues, psychological issues,
19 drug issues?
20 A. As near as I can tell reading through this.
21 Q. Did you find any standards specific to what scoring
22 system should be done or how it should be done?
23 A. No. Actually, I think there might have been one
24 standard, I'm not sure where it was, that said numerical
25 scoring should be used but didn't specify the type.
111
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 Q. These, provided under the heading of standards, for
2 example, description No. 10 by Mr. Bell, the Utah numerical
3 scoring system?
4 A. What's the question?
5 Q. No. 10 on your list, is that a standard that's adopted
6 by the industry or is this a standard that's used within
7 this system?
8 A. The way I look at it, it's a standard in the sense that
9 it explains how the Utah scoring system works.
10 Q. Are there other systems?
11 A. There are other systems.
12 Q. In your opinion, are any of these true standards in the
13 sense that they control the operation of the polygraph?
14 A. Well, the Bell standards that are part of No. 10 would
15 at least explain how to score a polygraph test, but they
16 don't explain how to have standards for what goes into the
17 test up front.
18 Q. I want to ask you about some surveys that you have done
19 to try to determine what the community thinks of the
20 polygraph?
21 A. Okay.
22 Q. Have there been studies or how many have been done, I
23 guess is the question.
24 A. To my knowledge, there have been four surveys done so
25 far, four surveys. Two of them include, I guess, two
112
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 surveys, so you could say there's been six total, but four
2 attempts to survey and two of the attempts involved two
3 different surveys each.
4 Q. Who did the studies that we are talking about?
5 A. The Gallup polling organization did the first study,
6 and then a student named Susan Amato did the second as her
7 master's thesis at the University of North Dakota. I did a
8 study with David Lykken and Honts and I'm not sure if there
9 were collaborators or any done more recently.
10 Q. Any of them published in the peer review literature?
11 A. One of them, and that's ours.
12 Q. Where was that published?
13 A. Journal of Applied Psychology.
14 Q. What were you trying to survey?
15 A. We surveyed members of the Society for
16 Psychophysiological Research and Fellows in Division One of
17 the American Psychological Association regarding their
18 opinions about the science of polygraph testing, about the
19 accuracy claims of proponents of polygraph tests, about
20 whether tests are standardized and objective, whether or not
21 it made sense to have them admitted as evidence in court.
22 Q. Now, you said you surveyed people in the Society for
23 Psychophysiological Research?
24 A. Yes.
25 Q. What kind of people are in that group?
113
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 A. These are people who are trained in psychophysiology.
2 We only surveyed American members of that society who had
3 full status, so these would be people who would understand
4 the basic science on which polygraph testing was based.
5 Q. You mentioned Fellows in the American Psychological
6 Association?
7 A. Yes. These were Fellows from a division of the
8 American Psychological Association that's referred to as the
9 General Psychology Group, so they are broadly versed in
10 principles of psychology. Because they have been elected to
11 the status of Fellow, they are distinguished in their field,
12 and they would typically be, for instance, the type of
13 people who would write textbooks on psychology and
14 introductory courses or be identified as psychologists with
15 general knowledge to serve on panels like the NAS panel and
16 that sort of thing.
17 Q. Why did you pick the two groups?
18 A. We thought it would be good to pick two that were
19 different but nonetheless had expertise that could bear on
20 this issue. So the psycho physiologists in particular would
21 be familiar with the physiological aspect of polygraph
22 testing and the members of the Division of General
23 Psychology who are Fellows we could expect to be
24 sophisticated in issues relating to testing and test
25 construction, standardization, objectivity of the test, that
114
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 sort of thing.
2 Q. How many people did you manage to survey in each of the
3 groups?
4 A. In each one we surveyed a little over 200 people.
5 Q. How did you do that?
6 A. Through the mail.
7 Q. Did you send them surveys?
8 A. We sent them surveys and followed up the surveys with
9 prompts and additional surveys if people failed to respond.
10 Q. What kind of a response did you get?
11 A. For the Society for Psychophysiological Research we got
12 a little over 90 percent, and for the Fellows in the
13 American Psychological Association we got about 74 percent.
14 Q. Is there any evidence, aside from your personal
15 opinion, that this was a valuable or well constructed study?
16 A. Well, it was published in the top peer review journal.
17 Both the surveys got a high response rate, which I think
18 shows that these people that we tried to survey thought the
19 survey was valuable and took it seriously. The president of
20 the American Psychological Association division that we
21 surveyed sent a letter to us indicating that he endorsed our
22 survey and thought it was very important and that he wanted
23 to encourage people completing the survey to actually do it.
24 So he lent his name and enthusiasm to the survey because he
25 thought the topic was important and the survey was
115
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 important.
2 Also the article that we published and the results of
3 the survey were selected by the publisher of a book on
4 research methodology to be used in the book as an exemplar
5 of how to do this kind of research.
6 Q. What kinds of polygraph tests were you surveying about?
7 A. Most of the survey questions were about the control
8 question test but also questions about the directed lie test
9 and the guilty knowledge test.
10 Q. How did you describe the test? What words did you use?
11 A. We thought it was very important to make it clear to
12 the people that we were surveying what we were asking them
13 about, so we wanted them to understand exactly what a
14 control question test was and how that was the same or
15 different from a directed lie and guilty knowledge test. So
16 we provided text from publications by proponents of these
17 procedures using their own words that explained what the
18 tests were and how they were run and how they were based.
19 Q. So, for example, the description of the directed lie
20 would be by Dr. Raskin?
21 A. Yes, and the control question test description was by
22 Dr. Raskin.
23 Q. Did you ask questions that appeared on both surveys?
24 A. Yes. There were some questions that overlapped on the
25 two surveys.
116
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 Q. I would like to take a look at State's Exhibit 6 and
2 see if these are some of those. Now, are these questions
3 that both groups looked at?
4 A. Yes.
5 Q. What's the significance of the fact that both groups
6 looked at them?
7 A. I think the significance comes in the response rates,
8 which you can see are highly similar, so even though these
9 are different groups of people with sort of different
10 perspectives on the topic, they produce almost virtually
11 identical endorsements of these different statements.
12 Q. Let me ask you what we are looking at here. Does this
13 -- this question then would be specific to the comparison
14 question test?
15 A. Right. The question would be asked essentially three
16 times, so the first time it would be "would you say that the
17 comparison question test or the control question test how we
18 described it is based on scientifically sound psychological
19 principles and theory?" 36 percent agreed and 36 percent of
20 the Psychological Association members agree. We ask the
21 same question about the guilty knowledge. You can see 77
22 percent and 72 percent saw that as based on scientifically
23 sound psychological principles, and then only members of APA
24 Fellows were asked about the directed lie test, and only 22
25 percent of them saw it the procedure as sound.
117
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 Q. Why don't we have a figure here?
2 A. Because the SPR members weren't asked about the
3 directed lie test.
4 Q. Why is that?
5 A. Mainly because we only had so much space in the survey
6 that we wanted to constrain to four pages in length and we
7 couldn't fit it on there.
8 Q. So does this indicate then that for the comparison
9 question test only about a third or -- less than a third had
10 a sound psychological principle?
11 A. Yes.
12 Q. And down here we have another question asked of both
13 about admitting this into Court?
14 A. Yes.
15 Q. Was there a difference here in terms of the two groups?
16 A. Very little difference. Neither group was enthusiastic
17 about admitting either results that showed that a person was
18 deceptive or results that showed that a person was truthful
19 before a Court.
20 Q. Now, I want to ask you about the Society for
21 Psychophysiological Research in particular, the results you
22 obtained there. Now, these are accuracy questions?
23 A. The first question asked people if they thought that
24 the control question test was at least 85 percent accurate,
25 and then we asked them if they thought it was at least 85
118
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 percent accurate in tests of guilty suspects or in tests of
2 innocent suspects, and you can see about a quarter of them
3 thought it might be that accurate.
4 Q. Why did you pick 85 percent?
5 A. We picked 85 percent because it was still a high number
6 but not so high as to seem outrageously high or ludicrous,
7 so we could have put something like 95 percent in there, but
8 then we thought people would see that as pushing an idea
9 that really was untenable. In any event, most of the people
10 who are proponents of the controlled question test would
11 argue that it's at least 85 percent accurate or higher. So
12 we thought if we started there we would see how well
13 scientists agreed with that estimate.
14 Q. The same issue on the friendly comparison question
15 test?
16 A. Friendly CQTs, the scientists thought that the friendly
17 test was more likely to be passed than adversarial at a rate
18 of three to one.
19 Q. How many of them thought that counter-measures might
20 work?
21 A. 99 percent thought counter-measures might work. We
22 also included a question that asked people specifically if
23 they were familiar with the research on counter-measures and
24 we broke the question down again, just looking at those who
25 specifically had read the research on counter-measures, and
119
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 I think it was then 98 percent agreed that counter-measures
2 could beat the test.
3 Q. And a number of them thought the Court should give
4 substantial weight to the results of laboratory studies?
5 A. Only 17 percent thought the results of laboratory
6 studies should be given substantial weight.
7 Q. Now let's take a look at the American Psychological
8 Association Fellows. Let's begin with -- is this their
9 version of the accuracy question?
10 A. Yeah. Here they were asked just fill in a blank and
11 give a numerical estimate of the accuracy of the control
12 question test, so they came up with numbers close to 60
13 percent.
14 Q. They found a difference -- your results gave a
15 difference between the innocent and guilty subjects?
16 A. Barely any difference, but they thought it might be a
17 little more accurate here for innocent suspects.
18 Q. For all intents and purposes they are about the same?
19 A. They are about the same.
20 Q. The next series of questions. What did you ask them in
21 No. 2, for example, if guilty would take a friendly CQT.
22 What was the question actually?
23 A. Well, the question was put to them for them to consider
24 that personally if they were guilty of a crime would they be
25 willing to take a friendly control question test, and the
120
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 idea would be well, you know you are guilty, so would you
2 take a test under these circumstances. Essentially
3 three-quarters are saying that they would.
4 Q. Then the third question, what was that asked?
5 A. That was the same thing switched. Assuming you were
6 innocent, how would you feel about taking a test
7 administered by the police. And I think those two questions
8 together express an overall low level of confidence in the
9 polygraph tests. A low level of confidence that friendly
10 tests would actually capture a guilty person and a worry
11 that an adversarial test will make an innocent person fail.
12 Q. What were you trying to get out with the question about
13 whether criminals and spies could beat a CQT?
14 A. This was a different way of dealing with the
15 counter-measures, so we asked them if they thought it was
16 likely that criminals and spies would be able to get the
17 kind of information they need to beat a control question
18 test, which they thought it was.
19 Q. And 75 percent --
20 A. Thought they could do it themselves.
21 Q. And when you asked them about the standardized aspect
22 of CQT?
23 A. Only 20 percent thought it could be characterized as a
24 standardized procedure.
25 Q. Are psychologists in general familiar with what a
121
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 standardized procedure in testing would be?
2 A. Yes, because that's a big part of psychology.
3 Q. And the question about CQT being relatively
4 independent, what were you trying to capture there?
5 A. That has to do with whether or not polygraph tests are
6 objective and this is really the definition of what
7 subjectivity is or a lack of objectivity, and you can see
8 that these respondents felt that this is a subjective
9 procedure.
10 Q. And the very last question, what were you asking there?
11 A. Well, this involves the idea, I think, that's also a
12 concern that's expressed by the National Academies Panel
13 Report, and that's that the proponents of polygraphy are
14 making claims that it's extremely accurate, and there's no
15 sound theory that will support the claim that the reports
16 are accurate and seems implausible that they would be that
17 accurate. So we thought we should ask people -- of course,
18 we didn't state that but we thought what they thought about,
19 what kind of empirical evidence would be required. "Given
20 that we don't have a sound psychological theory, that the
21 claims that are made are very high, shouldn't we have strong
22 empirical evidence before we believe it," and they agreed.
23 Q. Now, these are the results of your study?
24 A. Yes.
25 Q. Have you had indications from other source that other
122
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 members of the scientific community agree with this in terms
2 of accuracy?
3 A. Yes. In the first place there's the whole NAS Panel
4 Report, which I think almost all of the conclusions reached
5 by the National Academies Panel express sentiments similar
6 to the judgments that our survey respondents produced.
7 Q. Let me show you one from Page 107 of the report. This
8 is one -- I better read this for the record, I guess. "What
9 is remarkable given the large body of relevant research is
10 that claims about the accuracy of the polygraph today
11 parallel those made throughout the history of polygraph.
12 Practitioners have always claimed extremely high levels of
13 accuracy and these claims have rarely been reflected in
14 empirical research." Is this the kind of quotes you are
15 talking about?
16 A. Yes.
17 Q. Are there similar quotes?
18 A. There are similar quotes in the concluding chapter.
19 They get into the fact that you can't use the existing
20 theory and data to come up with the precise error rate of
21 the polygraph test.
22 Q. I believe looking at the studies they came up with 81
23 to 91 percent? Does that sound about right?
24 A. I'm not sure where you got that.
25 Q. Let's see. This is Page 129. Does that look familiar?
123
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 A. What they are saying there is that those studies that
2 exist produce estimates in the literature of 81 to 91. They
3 are saying those are overestimates.
4 Q. They overstate the accuracy?
5 A. Right.
6 Q. Would that also be in line with the research, the
7 surveys that you have done?
8 A. Yes.
9 Q. Now, is there also an article that surveyed
10 psychological textbooks?
11 A. Yes. There was a study done by Devitts and others that
12 examined how the polygraph testing and the theories of lie
13 detection are treated in introductory psychology texts and I
14 believe they examined over 35 introductory psychology texts
15 which could be seen as broadly representative of the views
16 of psychologists who write the texts and are trying to
17 capture what psychology is all about and present it to the
18 students who would first be exposed to it. And they
19 analyzed the nature of the material that appears in these
20 texts and found that these textbook writers overwhelmingly
21 present polygraph information to indicate that it's flawed
22 and overestimated.
23 There are different ways that they got this, but I
24 think they pointed out that negative characterizations of
25 polygraph testing exceeded positive characterizations at a
124
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 rate of 15 to 1, and hardly any of the texts cited research
2 on polygraph that indicated that the tests were accurate.
3 Q. You said there were other surveys?
4 A. There were other surveys that were done, yes. The
5 Gallup poll, for instance, and Amato thesis done recently.
6 Q. What can you tell us about whether those are helpful?
7 A. Well, each of them has different problems. I think the
8 Gallup and the Amato surveys don't really ask questions that
9 get at the heart of the issue related to people's views on
10 the theory of controlled question test or how accurate they
11 find it or whether they think it should be presented in
12 court. So they don't really give us a sense of what the
13 view of the scientific community is. I think even with
14 that, based on the results that they reported, there's not
15 evidence of a great deal of enthusiasm among scientists for
16 the polygraph given the results of their survey as they
17 present them.
18 I think there's also issues related to sampling. In
19 our survey we got a very high return rate and we
20 systematically surveyed members of the society. They either
21 got a low return rate or didn't explain how they sampled
22 members of the society.
23 Q. We saw a big difference between acceptance of the
24 guilty knowledge test and the probable lie or comparison
25 test?
125
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 A. Yes.
2 Q. Let's talk first of all about the first two surveys,
3 the Gallup and the Amato. Do you know whether that was
4 specific or not?
5 A. In those surveys there were various questions, but the
6 key question referred to all the time doesn't talk about the
7 type of polygraph test.
8 Q. Is there any way to know whether those people are
9 responding to the guilty knowledge or the control question
10 test?
11 A. No.
12 Q. Why does a response rate matter?
13 A. Well, if you have a high response rate, you can
14 reasonably generalize the data from those who did respond to
15 those who didn't, and then presume that the results you have
16 are reflective of the opinion of the entire membership of
17 that body.
18 If you have a really low response rate, then you start
19 to worry about who is it that's responding and the potential
20 for bias that might exist as a consequence. So, for
21 instance, busy people might not respond. Only people who
22 are interested in the topic might respond. It's hard to say
23 who is responding and who isn't, but it's hard if you don't
24 know. Because the response rate is really low you can't
25 figure that out, and then you can't extrapolate from that to
126
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 the views of the entire group.
2 Q. Now, did you include in your survey questions that had
3 been asked by other surveys?
4 A. We did.
5 Q. I would like you to take a look at State's Exhibit 9.
6 Does that reflect what you were talking about?
7 A. Yes. This is the one question that has been pointed
8 out the most as relevant to this issue of general
9 acceptance. So the stem of the question says, "Which one of
10 these four statements best describes your own opinion of
11 polygraph tests interpretation by those who have received
12 systematic training of the technique when they are called
13 upon to interpret whether a subject is or is not telling the
14 truth?" It says "sufficiently reliable method to be the
15 sole determinant" and hardly anyone picks that. "Useful
16 diagnostic tool when considered with other available
17 information." That's the popular response. And "of
18 questionable usefulness with other available information."
19 That's the other popular response.
20 It's ambiguous what the responses mean in that the
21 polygraph is a diagnostic tool, for instance, used as an
22 investigative tool from the police used to get a confession.
23 Many members of the Society for Psychophysiological Research
24 are familiar with the guilty knowledge test. So when they
25 are thinking about the procedure, they could be thinking
127
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 about the guilty knowledge test. The problem is the
2 question isn't very specific.
3 Also you could say if 60 percent see it as a useful
4 diagnostic tool, that's more than half. But it's not a
5 ringing endorsement. In our survey, it's a little bit
6 different, because we ask this question and said, "When you
7 are answering this question, we want you to answer it as it
8 specifically applies to the control question test," so we
9 made it clear that we were actually referring specifically
10 to the control question test, and because we did that, the
11 percentage of people choosing Option B dropped a little bit
12 and the percentage picking C indicating it was of
13 questionable usefulness increased.
14 Q. Now, you said your study -- was it done before or after
15 the National Academies report came out?
16 A. It was done before, about five years before.
17 Q. Was there a study after yours? When did yours come
18 out?
19 A. Ours was published in '97.
20 Q. Was there a study after that?
21 A. Well, the most recent one was the one that I mentioned
22 done by Honts.
23 Q. Do you know when that was done?
24 A. I think it was done last year, maybe the year before.
25 Q. Do you know if it was before or after this polygraph
128
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 report came out from the National Academies?
2 A. Before.
3 Q. Was that published?
4 A. It's not published as far as I know.
5 Q. Anywhere?
6 A. Not that I am aware of.
7 Q. Peer reviewed or not?
8 A. No.
9 Q. I would like to ask you about that a little bit. This
10 was a telephone survey by Professor Honts?
11 A. Yes.
12 Q. Look at State's Exhibit 10. Why is this of interest to
13 us?
14 A. You might shrink it down a little bit because some
15 lines are missing.
16 Q. How about that?
17 A. I guess it was done in 2002 since it says that there.
18 MR. DANIELS: Excuse me, is this supposedly
19 something Dr. Honts prepared?
20 MR. COX: This is a summary.
21 MR. DANIELS: All right.
22 Q. This is your analysis?
23 A. Yes.
24 Q. What are we looking at here? First the response rate?
25 A. Well, in the first place, two groups were surveyed,
129
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 members of the American Psychology and Law, and members for
2 the Society for Psychophysiological Research, so this is the
3 fourth time this group is surveyed. This is -- the two rows
4 indicate that the number of people who they tried to get
5 survey data from. For example, from the American Psychology
6 Law Society they contacted -- tried to contact 205 people.
7 They reached on the telephone 76. 21 refused, 55 were
8 surveyed. So the overall response rate in terms of the
9 number of people who were eligible was 27 percent.
10 With SPR, there were 366 people they tried to reach.
11 They surveyed 38, which is 16 percent of the total. So
12 these are very low numbers and would make it difficult to
13 generalize from these responses to the group as a whole.
14 Q. Why did you label this one a key question?
15 A. It was one of the questions that they used that I think
16 gets at an important issue, which is -- I think this is
17 something that you asked me about earlier in terms of does
18 the polygraph add value to the legal proceeding or does it
19 add validity to the whole procedure? So this question was
20 "would the accuracy of judicial verdicts be increased or
21 decreased if experts could produce polygraph test results in
22 courts of law?" This is the percentage of people responding
23 that presenting polygraph tests in court would either have
24 no effect on the procedure or actually decrease the
25 accuracy. And it was 45 percent in the American Psychology
130
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 Law Society and 56 percent in the Society for
2 Psychophysiological Research. So approximately half of the
3 people in these organizations felt that adding polygraph to
4 the judicial process would not improve it.
5 Q. Can you tell us, based on your experience, is polygraph
6 used very much outside of the United States?
7 A. It's not. It's primarily used -- the two countries
8 that really embrace the polygraph, at least in terms of the
9 control question test, are Canada and Israel. One question
10 uses the guilty knowledge test extensively and that's Japan.
11 Q. When you get out of those countries is it widely used?
12 A. As best I can tell it's not widely used. There may be
13 places in different countries where it's been tried. In
14 western Europe and the western industrialized countries
15 outside of the United States and Canada it doesn't seem to
16 be used.
17 Q. How about Australia?
18 A. Not used in Australia.
19 Q. And you indicated you brought us a report from Germany?
20 A. Right.
21 Q. Which indicates Germany is not using it either?
22 A. Correct.
23 Q. That's all the questions I have.
24 THE COURT: Let's take a short recess, about 15
25 minutes.
131
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 (Note: Court stood in recess at 2:30 to 2:45).
2 CROSS-EXAMINATION BY MR. DANIELS
3 Q. Good afternoon, Dr. Iacono.
4 A. good afternoon.
5 Q. Your primary specialty is in schizophrenia, isn't it?
6 A. I'm not sure I would say that now. I'm doing more
7 studies of adolescents and twin mental health than I have
8 schizophrenia.
9 Q. Last time we met your primary specialty was
10 schizophrenia and it's now shifted?
11 A. Shifted a bit.
12 Q. With adolescence and mental health?
13 A. Right.
14 Q. But you have done some work on polygraph study,
15 correct?
16 A. Yes.
17 Q. Just to set the stage a little and let the Court know
18 the nature of the battleground here, it's fair to say that
19 you and your mentor and friend and colleague, Dr. Lykken,
20 have been identified with the anti-polygraph camp if we are
21 talking about the control question polygraph on one end of
22 the debate. Is that fair to say?
23 A. I think so.
24 Q. Then there are others that are Ph.D.
25 psychophysiologists who are in the camp that believes there
132
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 is validity in the comparison question polygraph exam; is
2 that fair to say?
3 A. Yes.
4 Q. And, in fact, there's been the clash of the Titans
5 before this Court -- this case in this Court where you
6 testified in one side and various others have testified on
7 the other side before, correct?
8 A. Yes.
9 Q. You and Dr. Lykken actually favor a form of polygraph
10 scoring that you referred to as the guilty knowledge test?
11 A. That's a type of test, not a type of scoring.
12 Q. All right. We will get to the details of it in a
13 moment, but it uses the same instrument, doesn't it?
14 A. Yes, it could.
15 Q. Measures the same kinds of tracings on the graphs or
16 the computer printout?
17 A. Yes.
18 Q. And you and Dr. Lykken have occasionally competed for
19 grant monies with your studies of the GKT, guilty knowledge
20 test, and other people applying for grant money for the CQT
21 or comparison question test; is that fair to say?
22 A. I'm not sure it is. We have applied for grant money,
23 but I'm not sure that you would look at it as sort of a
24 contest between getting money for two different techniques.
25 Q. Do you remember agreeing to that characterization
133
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 several years ago in testimony?
2 A. I don't.
3 Q. All right. But it's fair to say that you have applied
4 for grants that promote your use of the polygraph and
5 Dr. Lykken's use of the polygraph, the GKT, correct?
6 A. Well, I wouldn't say to promote it but to study it. To
7 understand how it works.
8 Q. Because it's your position that that test, the method
9 of administration, method of scoring, has scientific
10 validity, correct?
11 A. Yes.
12 Q. You have suggested to the federal government and others
13 that they use the GKT instead of the CQT in their efforts of
14 detection of deception and finding spies and determining who
15 is a criminal and who is not; is that fair to say?
16 A. I don't think it would be fair to say instead of. I
17 suggested they try to use it more, yes.
18 Q. And those efforts have been largely unsuccessful,
19 haven't they?
20 A. Yes.
21 Q. Generally, the comparison or control question polygraph
22 test is the numerically scored test of choice in the
23 countries that use polygraph evidence as a diagnostic tool;
24 isn't that fair to say?
25 A. Except Japan.
134
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 Q. Except Japan?
2 A. Right.
3 Q. The United States?
4 A. Canada, Israel.
5 Q. And the United States, the biggest user of CQT
6 polygraph testing is the United States government; is that
7 fair to say?
8 A. I believe it is.
9 Q. Every single federal law enforcement agency has a
10 full-time paid staff of polygraphers administering that
11 particular kind of examination?
12 A. I can't say for sure every single, but there's maybe 25
13 of them that do it, yeah.
14 Q. Well, the Secret Service?
15 A. Secret Service does.
16 Q. FBI?
17 A. They do.
18 Q. Do you know how many polygraphers they have?
19 A. I don't.
20 Q. You don't know whether it's in the neighborhood of 100?
21 A. I'm sorry, I don't know.
22 Q. You were on the Advisory Committee for the Department
23 of Defense Polygraph Institute for a number of years?
24 A. Yes.
25 Q. Are you still?
135
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 A. No.
2 Q. When did that end?
3 A. I'm not quite sure. I think it's the early 90s.
4 Q. Could you tell the Court what the Department of Defense
5 Polygraph Institute is, please.
6 A. It's a facility that trains polygraph examiners. It
7 provides their education and their training in polygraph
8 technique. It's also a government institution that conducts
9 its own research program on the polygraph. They also have a
10 program where they fund other people to do research.
11 Q. Have you ever done research for the institute?
12 A. Yes.
13 Q. And to your knowledge have other experts, such as
14 Dr. Raskin, Dr. Honts?
15 A. I'm pretty sure Dr. Honts has. I'm not sure about
16 Dr. Raskin.
17 Q. And Dr. Gordon Barland was, for the years you were on
18 the Advisory Committee, the Director of Research at that
19 institute?
20 A. Yes.
21 Q. And that institute teaches the validity and use of the
22 comparison question test, does it not?
23 A. It does.
24 Q. And that institute not only does research but it trains
25 most of the federal polygraph examiners in various agencies?
136
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 A. Yes.
2 Q. Not just the armed forces?
3 A. That's correct.
4 Q. Do you know how many exams the Department of Defense
5 does in a given year?
6 A. I don't, but it's on the order of thousands.
7 Q. Around 20,000, you testified before?
8 A. I think it might be something like that.
9 Q. 20,000 a year? And those are comparison question
10 tests, correct?
11 A. No, those would be counter-espionage and employee
12 screening tests.
13 Q. You understand this hearing is not related to screening
14 kinds of tests. We are talking about specific incident
15 comparison question tests in the New Mexico rule here,
16 right?
17 A. Correct.
18 Q. But you know that the Department of Defense Polygraph
19 Institute administers comparison question polygraph tests,
20 correct?
21 A. Yes.
22 Q. And --
23 THE COURT: I'm sorry to interrupt, but is there a
24 difference between controlled question and comparison
25 question tests?
137
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 MR. DANIELS: I can tell you or explain it through
2 the witness. Maybe the proper procedure is to have a
3 dialogue with him.
4 THE COURT: Doctor, there a difference?
5 THE WITNESS: I think there is a difference, in the
6 sense that you can get confused. Some people use the term
7 "comparative question test" to refer to what we have been
8 talking about as the control question test and the directed
9 lie test because both procedures involve comparison
10 questions. But originally the inventor of what's called the
11 controlled question test, when he originally presented it he
12 called it something more like the comparison question test,
13 so if we are not careful, we can all get confused.
14 But today when we have been talking about the control
15 question test, we have been talking about a test that has a
16 probable lie control and a directed lie is different where
17 the person is actually told to lie. The comparison question
18 test could be the two of those together. That's how I would
19 characterize it.
20 THE COURT: Thank you.
21 Q. (By Mr. Daniels) We have a different understanding of
22 it then, Your Honor. Let me show you Petitioner's Exhibit
23 A, one of the documents submitted in the terms. This may
24 help a little. Do you recognize what that is?
25 A. It's a polygraph chart.
138
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 Q. Let's talk about what's represented in Exhibit A. Is
2 this a fairly typical set of readouts in a polygraph chart?
3 A. I don't know what you mean exactly by typical, but you
4 can see two cardio channels, two respiratory channels and
5 the GSR channel and those are the components you would
6 typically see.
7 Q. This is what you would see in the normal control
8 question test?
9 A. The patterns are a little unusual but the signals
10 recorded there are typical.
11 Q. This is also the kind of tracing that would be done in
12 the GKT, the guilty knowledge test, correct?
13 A. It could be, yes.
14 Q. Because both of them measure the thoracic respiration,
15 that's the top one here around the chest?
16 A. Well, in fact, very few GKT studies measure
17 respiration. Most of the GK studies measure GSR and nothing
18 else, and the more recent ones measure brain potential so
19 they don't measure any of these signals, but there are very
20 few -- GKT that measure the cardio and the respiratory
21 channel.
22 Q. Let's go through what's on here and narrow it down to
23 your use of the GKT. The top one you recognize as thoracic
24 respiration. That's breathing measured of the chest,
25 correct?
139
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 A. Yes.
2 Q. The second is abdominal respiration, correct?
3 A. Yes.
4 Q. Breathing measured around the abdomen?
5 A. Yes.
6 Q. The third one, which you referred to as GSR, this says
7 SC, is a measurement of the electric impulse as transmitted
8 through moisture on the palmar glands, correct?
9 A. It's actually the resistance of the sweat gland or the
10 conductance. The SC stands for skin conductance.
11 Q. How well or poorly an electrical signal will be
12 transmitted on the palmar surface, correct?
13 A. That's right.
14 Q. The theory here being that the concerned or agitated or
15 fearful or whatever response it is, there will be more
16 perspiration activity, correct?
17 A. Yes.
18 Q. The fourth line here is the blood pressure, which is
19 measured with a traditional blood pressure cuff of the kind
20 we use in the doctor's office, correct?
21 A. Yes.
22 Q. And the final line here, could you tell us what that
23 is?
24 A. I will guess from the abbreviation that it's the photo
25 plethysmograph, a measure of the finger pulse volume from a
140
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 sensor attached on the fingertip.
2 Q. Blood volume of the fingertip?
3 A. Yes.
4 Q. Now, you mentioned that the GKT is often now just based
5 on the skin conductance?
6 A. Yes.
7 Q. What's sometimes called the galvanic skin response?
8 A. Yes.
9 Q. That's why you refer to it as GSR?
10 A. Yes.
11 Q. Because actually that, of all of these, is often the
12 clearest indicator of all of the various sensor devices;
13 isn't that correct?
14 A. That's what the research has shown.
15 Q. And so you may get initial information from the others,
16 but if you were forced to rely on one, the GSR probably is
17 the best one to do, correct?
18 A. Yes.
19 Q. Now, will you agree that the polygraph is a sensitive
20 instrument?
21 A. Yes.
22 Q. It's a scientific instrument?
23 A. Yes.
24 Q. And it's reasonable to expect that device to register a
25 response to the type of emotional turmoil associated with
141
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 lying?
2 A. Yes.
3 Q. In fact, you have written that before, haven't you?
4 A. Sounds like it.
5 Q. I'm sorry?
6 A. Sounds familiar.
7 Q. Do you remember writing "Scientific Status of
8 Polygraph"?
9 A. Yes.
10 Q. In any event, that statement is true?
11 A. Yes.
12 Q. When you say there's emotional turmoil associated with
13 lying, could you tell the Court what you mean by that?
14 A. Well, when you lie you respond with physiological
15 arousal.
16 Q. Why?
17 A. Why?
18 Q. Yes.
19 A. There could be multiple reasons, but these would
20 involve the types of emotions which you experience when you
21 lie which could include, for instance, shame or guilt or
22 simply conflict. For instance, maybe you are lying to
23 somebody who you feel especially bad about lying to and you
24 feel disappointment in the fact that you would lie to such a
25 person. So there would be a complex set of emotions in any
142
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 given instance that might vary from person-to-person that
2 would characterize the emotional turmoil.
3 Q. As a psychophysiologist you would agree there is
4 emotional turmoil, whatever label you attach to the turmoil?
5 A. Yes.
6 Q. And that is associated with lying?
7 A. Yes.
8 Q. And you mention a number of possible options as to
9 whether it's guilt or shame or other components of the
10 emotional turmoil? Correct?
11 A. Yes.
12 Q. But the bottom line is there is emotional turmoil?
13 A. Yes.
14 Q. And that is picked up by this sensitive scientific
15 instrument, the polygraph?
16 A. Yes.
17 Q. In fact, don't we use our own common experience in
18 daily life in sort of picking up those kinds of clues from
19 other people and sensing them in ourselves?
20 A. I think that's fair.
21 Q. We notice a difference in the feeling in our face, for
22 example?
23 A. Yes.
24 Q. We notice sweaty palms if we are lying to somebody, or
25 if we see somebody lying to us we may see somebody exhibit
143
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 that kind of behavior?
2 A. Yes.
3 Q. Have you testified in court enough to know about the
4 kinds of things jurors are told to take into account in
5 evaluating whether a witness is telling the truth?
6 A. I actually have never -- well, I seldom ever testify at
7 a point where there's a jury so I am not familiar with that.
8 Q. You are not familiar with demeanor evidence that a jury
9 is told to take into account?
10 A. No.
11 Q. Now, isn't it a fact that the polygraph can detect
12 what's going on in the body causing those reactions a lot
13 more accurately than someone can do just looking at someone
14 and trying to see how they are reacting to the emotional
15 turmoil associated with lying?
16 A. Yes.
17 Q. Now, let's talk about the difference between the guilty
18 knowledge test that you and Dr. Lykken favor, and the
19 comparison question test. Do you prefer I use the word
20 "control question test"?
21 A. I do.
22 Q. Let's use this chart, Exhibit A, to talk first about
23 the guilty knowledge test, all right? Do you recognize this
24 as a numbers chart or not?
25 A. I see numbers across the bottom so that makes sense.
144
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 Q. Right. And numbers chart is the kind of thing that is
2 the same as what the guilty knowledge test is based on,
3 isn't it?
4 A. It's similar.
5 Q. All right. Basically this is the kind of thing you
6 were talking about some polygraphers using to test a
7 subject's reactivity before the main test, and also to show
8 the subject that the test will work on them, correct?
9 A. Yes.
10 Q. Someone is typically told, "Pick a number between 2 and
11 10 and don't tell me what it is," correct?
12 A. They are told not to tell you or sometimes they are
13 told to write it down on a piece of paper. Dr. Raskin and
14 Honts say, "Write it down on a piece of paper and show me
15 what it is."
16 Q. They are told, "I'm going to ask you, one by one, is it
17 number 2, is it number 3, is it number 4, and I want you to
18 say no each time," correct?
19 A. Yes.
20 Q. Now, and there are no questions in the numbers test
21 that relate to the specific incident that's being
22 investigated, correct?
23 A. Right.
24 Q. And there are no control questions that have to be
25 devised for this test?
145
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 A. Right.
2 Q. It just basically compares your reactions to the target
3 number and your reaction to, say, the other number, correct?
4 A. Right.
5 Q. Now, the first question in any polygraph exam is never
6 user-scored, correct?
7 A. Right.
8 Q. Tell us why that is.
9 A. Because usually the first stimulus you are exposed to
10 produces a large response just because it's the first in a
11 sequence.
12 Q. It's a habituating response, getting used to the whole
13 process?
14 A. The first response isn't habituating but it's a
15 novelty, a I'm-not-used-to-the-process.
16 Q. Nonhabituating response, is that fair to say?
17 A. Right.
18 Q. So no responsible practitioner of polygraphy would
19 score the first one?
20 A. Right.
21 Q. It's a throw-away. Now, by looking at this, can you
22 tell us which number the test subject would have picked?
23 A. Looks like 5.
24 Q. No. 5, because you can see clearly the sweat in the
25 palms being registered by the peak there, correct?
146
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 A. Right.
2 Q. And you also see a decline in the breathing right after
3 that peak, correct?
4 A. Yes, but what's especially distinctive is the increase
5 in the breathing.
6 Q. All right. But then there's a stairstep decrease after
7 that, correct?
8 A. Right.
9 Q. And that's consistent with the deer caught in the
10 headlights phenomenon that we see in the polygraph exams
11 that breathing becomes a little bit repressed, right?
12 A. Right.
13 Q. You see the blood pressure going up until the subject
14 gets past No. 5 and then gradually come back down, right?
15 A. Right.
16 Q. And I never have been good at reading the
17 plethysmograph. What is it that you see in this tracing on
18 this test?
19 A. You don't see much, but what you would expect to see is
20 a constriction in the amplitude of the oscillating line.
21 Q. That's why it's -- the amplitude is not as great here?
22 A. Yeah, so it gets reduced a little bit.
23 Q. That's also consistent with being No. 5, right?
24 A. Right.
25 Q. So those various measures detect the emotional turmoil
147
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 that the person has when we get to his saying, "No, my
2 number is not 5," correct?
3 A. Yes. Excuse me. I should qualify that because here
4 there are other things going on besides emotional turmoil in
5 the sense that he is told to answer the question no. But
6 there's also kind of a logical buildup of arousal to this
7 point, which I think is why you see the blood pressure
8 increase. So there's probably other things going on here,
9 too, like a recognition of the fact that the number 5 is
10 coming up, but I think those things are working together.
11 Q. They work together also in the comparison question
12 test, don't they?
13 A. Yes.
14 Q. Because if this were a comparison question test, we
15 would have the relevant specific question being asked and
16 measured, but we would also have next to it on one side or
17 the other the control question, correct?
18 A. Yes.
19 Q. And the control question is designed to elicit a
20 response also, correct?
21 A. Yes.
22 Q. The theory being that if you are lying about the
23 specific incident that's being investigated, the crime you
24 are charged with or whatever, and you are guilty of that,
25 you will have a greater reaction on that than you will to
148
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 the control question to which it is compared, correct?
2 A. That's the idea.
3 Q. And the theory also, the scientific theory that these
4 various lab studies and field studies have been directed at,
5 is that if you were innocent of this, then your reaction to
6 a control or comparison question designed to elicit a
7 response to some other kind of lie in the nonrelevant
8 situation will be greater, correct?
9 A. Yes.
10 Q. Now, the way you use this underlying science in the
11 guilty knowledge test that you and Dr. Lykken prefer is that
12 if there are distinctive things that only the guilty person
13 would know, for example items at a crime scene, you ask a
14 series of questions about -- for example, if the murder
15 weapon has never been publicized, "Was it an ice pick, was
16 it a kitchen knife, was it a hammer," those kinds of
17 questions designed to see whether the true murder weapon
18 will show up like this No. 5 on the numbers test, correct?
19 A. Yes.
20 Q. And people react in a way that can be measured on the
21 polygraph, correct?
22 A. Yes.
23 Q. Now, is there anything in this tracing, if you were
24 using this for your guilty knowledge test, which would
25 indicate just by looking at the tracing whether it was fear,
149
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 shame, arousal, knowledge or any particular thing that
2 caused that peak right there?
3 A. No.
4 Q. So the guilty knowledge test, like the comparison
5 question test, causes reactions, although you cannot look at
6 the polygraph and say, "I can tell you what specific form of
7 emotional turmoil this person was having," it can just tell
8 you they were having a reaction, correct?
9 A. That's correct.
10 Q. You talked about several different kinds of scoring of
11 polygraph tests that are done, and I wanted to eliminate the
12 ones that aren't related here. You talked about the
13 relevant/irrelevant test. That's not used by any
14 responsible polygrapher in the United States today, is it?
15 A. I hope not.
16 Q. Okay. It's long been left behind, correct?
17 A. Yes.
18 Q. Because the relevant/irrelevant test is just a whole
19 lot like this numbers test. The only thing you are looking
20 at is whether you react to the question about the crime. So
21 the innocent question is going to have a reaction, and you
22 have nothing else to compare it to, right?
23 A. That's correct.
24 Q. So that really has a high rate of innocent people being
25 found deceptive, correct?
150
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 A. Yes.
2 Q. The control question test, and maybe I ought to clarify
3 why I am referring to it occasionally as comparison
4 question, there's been some criticism of using the term
5 "control" in this test because I believe you have written
6 before that the comparison question is not a true control,
7 in your view, correct?
8 A. That's correct.
9 Q. Well, it does provide a basis for comparison, doesn't
10 it?
11 A. It does.
12 Q. If you were asked, "Did you steal the $20," in the
13 matter being specifically tested on, and you were asked a
14 question to compare that response to, "At any time before
15 last year have you ever taken anything that didn't belong to
16 you," that would be the comparison question, correct?
17 A. Yes.
18 Q. It does give us an indication of how a person responds
19 to a lie, doesn't it? The comparison question?
20 A. Well, that's an assumption.
21 Q. Right. And in order to test that scientific assumption
22 or that hypothesis you would conduct studies, field studies
23 or laboratory studies to see if it works that way in real
24 life, correct? Isn't that the scientific method?
25 A. Yeah, but what your question is, as I understood it, is
151
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 you want to find out if people are lying to the control
2 question, so that would be a different thing to study than
3 whether the control study works the way people say it does.
4 Q. The studies have shown that most people will show a
5 reaction to those kinds of general questions of wrongdoing,
6 such as, "Have you ever hurt someone who trusted you," or,
7 "Have you ever taken something that didn't belong to you,"
8 or, "Have you ever lied to someone in authority," isn't that
9 fair to say?
10 A. I'm not sure. People respond to control questions. Is
11 that what you are asking me?
12 Q. Control questions of that general nature about prior
13 conduct, that's very general, "Have you ever basically done
14 a dishonest thing, have you ever done a hurtful thing, have
15 you ever lied to someone in authority," people do respond to
16 those questions, don't they?
17 A. Yes.
18 Q. And to test whether the hypothesis is correct that the
19 arousal created by the lie to the subject under inquiry is
20 going to be greater for the guilty and the converse for the
21 truth-tellers, one would want to test it in either a
22 laboratory or a field setting or preferably both, correct?
23 A. Yes.
24 Q. Because that's the ultimate proof of whether a
25 scientific hypothesis is correct, isn't that fair to say?
152
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 A. Yes.
2 Q. Now, you are familiar with the New Mexico rule that's
3 been in effect here a couple of decades, correct?
4 A. Vaguely familiar with it.
5 Q. It requires numerical scoring to be used, correct?
6 A. I believe that's the case.
7 Q. I have it here if you --
8 A. No, I think I remember reading it last night.
9 Q. And it also requires that the examination be
10 quantitatively scored in a manner that is generally accepted
11 as reliable by polygraph experts, right?
12 A. If that's what it says.
13 Q. And the method generally accepted as reliable by
14 polygraph experts is the control question, numerical scoring
15 system; isn't that correct?
16 A. I would say so.
17 Q. All right. And you have never heard of anyone coming
18 in trying to introduce a relevant/irrelevant test in New
19 Mexico?
20 A. No.
21 Q. Or global scoring, for that matter, as opposed to
22 numerical scoring?
23 A. I am not aware of such a case.
24 Q. You had mentioned global scoring in your direct
25 examination, but that is not what is meant by numerical
153
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 scoring, correct?
2 A. That's correct.
3 Q. The global scoring is where an examiner will look not
4 only at the numbers that you can look at and measure with a
5 ruler or by reference to the dots on the chart, but takes
6 into account the examiner's impressions and the examiner's
7 impression about outside information and so on in deciding
8 whether to pass the polygraph, correct?
9 A. Yes.
10 Q. But the New Mexico rule in requiring numerical scoring
11 does not require that kind of subjective decision-making?
12 A. It's designed to prevent it, yes.
13 Q. From your studies, you know that there is a high
14 correlation between examiners in numerically scoring charts?
15 A. Yes.
16 Q. One examiner will agree substantially with another
17 examiner's scoring of those charts, correct?
18 A. Right.
19 Q. And you are aware that in New Mexico the rule is that
20 all the charts have to be reserved and produced to the other
21 side for their review and their expert analysis to do an
22 independent blind scoring?
23 A. Right.
24 Q. And that's a good way of avoiding the problem of an
25 incompetent or crooked examiner, if you want to hypothesize
154
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 that, coming up with the wrong scoring of the chart,
2 correct?
3 A. Yes.
4 Q. And you are familiar with the New Mexico rule that
5 requires that all polygraph examinations be either
6 audiotaped or videotaped, right?
7 A. Yes.
8 Q. That's unusual in the United States and in Canada and
9 in Israel, isn't it?
10 A. Yes.
11 Q. Because most standards that exist don't go as far as
12 New Mexico standards go?
13 A. True.
14 Q. And that also is a good standard to use to avoid any
15 kind of misconduct or incompetent administration of the
16 exam, correct?
17 A. Yes.
18 Q. So you can have somebody else look at it or listen to
19 it and see if the questions were formulated in a responsible
20 way, if the pre-test interview was done correctly, and if
21 the questions were asked in the right sequence and in the
22 right manner without undue emphasis or distortion?
23 A. Well, I guess I would say another person could look at
24 it and see if it's done according to the way people are
25 taught to do it properly.
155
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 Q. Right. We will get to that next, but first I want to
2 talk about your comments before about the competency of the
3 examiners and the variations in which the exam can be
4 administered. This recording requirement, the requirement
5 that the recordings be turned over to the other side along
6 with all of the charts and all of the score sheets, is a
7 safeguard against those kinds of problems, isn't that true?
8 A. It would be.
9 Q. And have you inquired into how that's worked over the
10 years in the State of New Mexico?
11 A. No. At least I don't know if I have inquired. No one
12 has told me.
13 Q. Now, let's look at the specific objections you have
14 made to the concept of the controlled question test. Are
15 you aware that a number of people in the field are calling
16 them comparison questions in the same way that many refer to
17 them as control questions?
18 A. That started in the last year or two.
19 Q. All right. I will use the term "control question"
20 here, but there may be other witnesses who refer to the
21 comparison question, but we are talking about the same
22 thing.
23 A. Okay.
24 Q. The use of the question about the specific incident in
25 comparison to the other lie, the irrelevant lie that may be
156
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 on the exam, whether you call it the directed lie or the
2 probable lie, correct?
3 A. Now you are using the word "control question" to refer
4 to the directed lie, too?
5 Q. Isn't the directed lie a form of control question?
6 A. Well, the answer I gave earlier, for the sake of
7 clarity I was trying to use the control question test for
8 the probable lie and the directed lie is the directed lie
9 test.
10 Q. Okay. There are people who view the control question
11 -- I'm sorry, the probable lie and the directed lie as just
12 two variations on the control question, aren't there?
13 A. See, that's where I thought the word "comparison
14 question" was designed to embrace the two variations but
15 "control question" was designed to talk about the probable
16 lie.
17 Q. Whatever you call the irrelevant lie, some examiners
18 have used the directed lie, which is to say, "I want you to
19 lie to me about this subject. I'm going to ask you if you
20 ever told a lie, and I want you to deny it, and I know you
21 are lying to that." That's the directed lie, correct?
22 A. Correct.
23 Q. The other form is, "I want to find out how honest you
24 are. Have you ever told a lie before the incident in
25 question," and that is called the probable lie, because
157
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 everyone is going to react to that, correct?
2 A. That's the assumption.
3 Q. Whichever way you ask that question about the
4 irrelevant incident, it is used as the basis of comparison
5 against the reaction to the specific incident that's being
6 tested on?
7 A. Right.
8 Q. Okay. So when you talk about control question, you
9 mean only with the comparison question -- I'm sorry, the
10 probable lie?
11 A. The probable lie, right.
12 Q. I will try to remember to use it in that context. Do
13 you know of anyone who is actually using the directed lie in
14 New Mexico?
15 A. You mean a polygraph operator?
16 Q. Yes.
17 A. No.
18 Q. Do you know if Dr. Honts or Dr. Raskin uses that
19 anymore?
20 A. I believe they do.
21 Q. You are not aware of any current polygraphers in New
22 Mexico using the directed lie form of the irrelevant
23 question? The irrelevant lie question, I should say?
24 A. No. Okay.
25 Q. Now, I assume your testimony here is not making any
158
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 kind of distinction between the directed lie and the
2 probable lie, is it?
3 A. Well, it is, in the sense that my testimony was about
4 the probable lie test, unless I was asked specifically about
5 the directed lie test.
6 Q. I will leave the directed lie aside, because I don't
7 think anybody is using it. If you had a criticism about it,
8 it can certainly be addressed by fine-tuning the rule and
9 avoiding that, although there is data to support it, but
10 your criticisms are directed to the probable lie and in that
11 regard you are not making any separate criticisms of the
12 directed lie. Is that fair to say?
13 A. Well, unless I was asked about it, but now I am not
14 sure -- almost everything I discussed in my testimony was
15 about the probable lie test, but I think there were one or
16 two questions in there specifically about the directed lie
17 test.
18 Q. Do you have any opinion about the directed lie that is
19 not the same as your opinions about the probable lie so we
20 can go ahead and get that resolved?
21 A. I do.
22 Q. All right.
23 A. I think they are largely overlapping, but the directed
24 lie is different in the sense that one of the advantages of
25 the probable lie test is that if the person is naive and
159
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 truly believes that all the questions are important, then it
2 would be more difficult to use counter-measures in that test
3 than in the directed lie test where the question that you
4 are using as the comparison is specifically identified as
5 such, and you are basically told that the point of asking
6 the question is to see what your response looks like in a
7 lie. It's much easier when I think the question is
8 presented that way to understand the significance of why you
9 want a strong response to that particular question than in a
10 probable lie test.
11 Q. One of the things I would like to do as we go along is
12 to distinguish between those things which are your opinion
13 in the absence of data and your opinion based on specific
14 test data. Is it fair to say that impression you have is
15 one in the absence of data?
16 A. Yes.
17 Q. Because there have been no data that show the directed
18 lie is any less effective than the probable lie, correct?
19 A. You mean for counter-measures?
20 Q. For counter-measures or any other test purpose.
21 A. Well, there isn't a lot of data on the effectiveness of
22 the directed lie that's in the literature.
23 Q. So your opinion is not based on data, it's based on
24 basically a hypothesis, correct?
25 A. Yes.
160
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 Q. As we go through the various areas, I will try to make
2 sure we make those distinctions so we can talk about the
3 things you are talking about on the basis of data and others
4 on what you think the data would show if there were tests on
5 it.
6 A. Okay.
7 Q. Now, you had some criticisms of the probable lie test,
8 and I think we addressed some of these. You said the
9 formulation of probable lies is not standardized. How would
10 you standardize the formulation of the probable lies?
11 A. I'm not sure. I think that's a fundamental problem
12 with the test. You could imagine having standards that for
13 certain types of people with certain backgrounds who commit
14 certain types of crimes there's a list of ten questions that
15 you choose from or things like that, but I think it would be
16 very difficult.
17 Q. Now, you have done a few studies, both field and
18 laboratory studies, of the control question polygraph,
19 haven't you?
20 A. Yes.
21 Q. Two or three?
22 A. I believe three.
23 Q. Three?
24 A. Yes.
25 Q. The study of the psychopaths you did, that was a
161
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 laboratory study?
2 A. Yes.
3 Q. That showed that psychopaths can't beat the test,
4 correct?
5 A. That would be the conclusion from that study.
6 Q. And there was about an 89 percent hit rate for the
7 subject in the test of yours?
8 A. Sounds a little high. Maybe it was 87. I'm not sure.
9 Q. You think it was 87? All right. I think you are
10 right. 87. I was thinking of a different one. 87 percent
11 hit rate on psychopaths, correct?
12 A. I think that was the overall hit rate for guilty, so
13 I'm not sure how it broke down but it didn't differ.
14 Q. How were the probable lies conducted in that study that
15 you conducted?
16 A. In a that study they were standardized because everyone
17 committed the same crime, so they were -- the probable lie
18 questions themselves were the same, but if people made
19 admissions during the pre-test interview that would be the
20 way they would have been changed.
21 Q. If they made admission in the pre-test interview you
22 would change the wording of the probable lie question?
23 A. To, "Other than what you told me about blah blah blah."
24 Q. Do you remember what the probable lie questions were in
25 the study you did?
162
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 A. No, they are in a table in the study but I don't
2 remember what they were.
3 Q. Did you find that the formulation of the probable lie
4 question with which you compared the specific incident made
5 any difference in the hit rate?
6 A. We didn't specifically do a study of different types of
7 probable lies.
8 Q. You also studied a number of polygraph exams conducted
9 by the Royal Canadian Mounted Police, correct?
10 A. Yes.
11 Q. Did you look at the comparison questions used in that
12 study?
13 A. Not in particular. I mean, they were used, but we
14 didn't make an inventory of them and examine them.
15 Q. Did you determine that some forms of probable lie
16 questions had a different accuracy rate than others in that
17 study?
18 A. No.
19 Q. And there was a high rate of detecting the guilty in
20 that study, too, wasn't there?
21 A. I think there was a high rate of detecting the guilty
22 in the sense that I discussed earlier, which had to do with
23 the fact that for guilty people who confessed after they
24 took a polygraph test, the confession then matched the
25 failed test, but that's not the same as detecting guilty
163
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 people.
2 Q. Well, we will get to that analysis shortly. You have
3 also tested the effect of drugs on people's responses, have
4 you not, to these polygraph responses?
5 A. Yes.
6 Q. And you found that the use of drugs will not defeat the
7 test?
8 A. Yes.
9 Q. That's consistent with what's been found by other
10 people, right?
11 A. No. In fact, that's the reason we did those studies is
12 other people found they did find drug effects so we wanted
13 to follow them up and try some different drugs and see if we
14 could improve on the procedure, but our studies, which were
15 based on the guilty knowledge test, actually did find an
16 effect.
17 Q. And basically there's no way to administer a drug
18 that's going to be selective in allowing you to respond to
19 one question and not another, correct?
20 A. That's why we did the study, because the first study
21 published on the topic actually showed the selective effect
22 that you are saying you can't show. What they argued is
23 because the drug affected anxiety, it affected anxiety
24 specifically to the relevant item on the test, so there was
25 a selective effect.
164
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 Q. Are you saying there was a study that showed you could
2 reverse the results of the test?
3 A. I am saying there was a study that showed that using
4 the guilty knowledge test, if you take an anti-anxiety drug
5 you can selectively reduce the response to the key item.
6 Q. Do you think that was a valid study?
7 A. What I can say about it is we tried to replicate it and
8 did not.
9 Q. Who was it who came up with that?
10 A. It was -- I believe it was Wade and Oran, maybe just
11 Wade. I can't remember who all of the authors were, but the
12 first author was Wade. This would have been 20, 25 years
13 ago. It's an older study.
14 Q. Has there ever been any study that's ever determined
15 that drug use can defeat a comparison question test and
16 cause different results?
17 A. Well, again, yes. There is one. There's one that
18 suggested that alcohol could have such an effect, and the
19 other study that we did that included the control question
20 test to examine alcohol effect and we couldn't replicate
21 that result either.
22 Q. You found that wasn't valid?
23 A. We couldn't replicate the result.
24 Q. The alcohol use did not reverse the effects of the
25 test?
165
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 A. Right.
2 Q. Who was it that did the study that showed that alcohol
3 reversed the results of the test?
4 A. Bradley.
5 Q. Bradley?
6 A. Yes.
7 Q. Where was that published?
8 A. I think it might have been published in
9 Psychophysiology.
10 Q. Is that a study that's relied on by anybody in the
11 field?
12 A. Not that I am aware of.
13 Q. I'm sorry?
14 A. Not that I am aware of.
15 Q. So you would agree that the use of drugs, alcohol, or
16 the existence of a condition of being a psychopath or
17 sociopath will not cause the opposite result on polygraph
18 tests?
19 A. Well, that's a bit too all-encompassing. Certainly the
20 research that's been done so far would suggest that the
21 specific drugs don't have the effect, and a psychopath in
22 the context of the laboratory studies has no advantage.
23 It's not to say that it would work the same way for a
24 psychopath in a field operation where they might be able to
25 use their pychopathic talents, if you will, to greater
166
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 advantage, and there's literally hundreds if not thousands
2 of drugs that could affect the autonomic nervous system and
3 the responses we considered, so it's hard to say based on
4 three or four of them that we can discount the effect from
5 any of them.
6 Q. You are not here to suggest to the Court that drugs or
7 alcohol can reverse the effects of the test, are you?
8 A. No.
9 Q. Incidentally, you mentioned the Rosenthal effect.
10 That, I guess, is something that you study in your field.
11 It's new to me. Does that apply to all of us human beings?
12 A. Yes.
13 Q. Including anti-polygraph witnesses as well as pro
14 polygraph witnesses?
15 A. Sure.
16 Q. Basically you might tend to confirm your own theories
17 and design tests in a way that support them?
18 A. Yes.
19 Q. Subconsciously, correct?
20 A. Yes.
21 Q. Now, you have specialized in schizophrenia good part of
22 your career, haven't you?
23 A. Yes.
24 Q. Can you tell us what the theoretical basis is of
25 schizophrenia? Can you give us a theoretical understanding
167
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 of what it is and why it works the way it does?
2 A. The major theory at present is that there are a
3 combination of genes that make people vulnerable for the
4 development of schizophrenia which when they then encounter
5 certain environmental triggers will ultimately lead them to
6 develop this disorder. Currently the prevailing theory is
7 that schizophrenia is a neurodevelopmental genetic disorder,
8 the idea being that perhaps the environmental events that
9 are responsible for the ultimate unfolding of the disease
10 are present early in life. They might reflect intrauterine
11 phenomenon or possibly effects related to birth trauma at
12 the time of birth. For instance, obstetric complications,
13 these sorts of things.
14 So people with the pregenetic disposition who
15 experience the environmental stressors in developing a
16 disease that typically emerges in adolescence in part
17 because the frontal lobes of their brain which show the most
18 pronounced development in adolescence takes place during
19 that period. It's this part of the brain that's believed to
20 be genetically vulnerable and damaged by the environmental
21 stressors and that accounts for most of the symptoms that
22 you see in schizophrenia.
23 Q. Have you been able to conduct laboratory tests to see
24 if this theory is 100 percent accurate?
25 A. Laboratory studies have been studied with schizophrenia
168
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 patients, yes.
2 Q. I understand that, but my question is to determine
3 whether this theory of what causes schizophrenia, how it
4 works in the human brain, is absolutely accurate? Or is it
5 --
6 A. I have to answer no.
7 Q. Is it still -- I'm sorry?
8 A. I have to answer no if you put it that way.
9 Q. Yet, people in the field do make diagnoses of
10 schizophrenia, don't they?
11 A. Yes.
12 Q. You have done so?
13 A. Yes.
14 Q. And people in the field testify about it in court?
15 A. Yes.
16 Q. Based a lot on observation and empirical evidence,
17 correct?
18 A. Yes.
19 Q. Because there is no way you can take an x-ray of a
20 brain and say, "This person has schizophrenia?"
21 A. That's right.
22 Q. -- in any kind of foolproof way. That's true with a
23 lot of is scientific knowledge, isn't it, Dr. Iacono?
24 A. What's true?
25 Q. What's true is that sometimes we can't take a
169
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 photograph and prove with 100 percent accuracy what is the
2 cause of the condition or to diagnose with 100 percent
3 accuracy a condition?
4 A. That's right.
5 Q. But we make a lot of scientific decisions based on
6 studying empirical evidence and conducting tests as best we
7 can?
8 A. Right.
9 Q. Do you know what the hit rate on diagnosis of
10 schizophrenia is?
11 A. Can't be determined.
12 Q. No way to validate that?
13 A. No.
14 Q. Do you know if anyone would say it's even as much as 80
15 percent?
16 A. They can't say that.
17 Q. Now, you talked a little bit about a chart of what you
18 called assumptions underlying the control question test.
19 And let's talk briefly about these assumptions. If I
20 understood your testimony correctly, you have agreed that
21 the third assumption is correct and you don't quarrel with
22 that?
23 A. No, I don't agree with that.
24 Q. You don't agree with that? Have you changed your
25 position over the years on that?
170
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 A. Not that I am aware of.
2 Q. Do you recall authoring "Assessing Deception Polygraph
3 Techniques" in 1988 with Christopher Patrick in Clinical
4 Assessment Of Malingering And Deception?
5 A. Yes.
6 Q. Incidentally, a number of your articles weren't
7 included in the materials. Did you bring those with you,
8 the prior articles you have written?
9 A. I did not.
10 Q. Do you recall writing in that article about the
11 assumptions underlying the control question test? Does that
12 look familiar? Would you like to look at the printed copy?
13 A. No, I see it. It's familiar.
14 Q. You recognize this, don't you?
15 A. Yes.
16 Q. At that time you said there were four assumptions that
17 were basic to the control question test?
18 A. Well, the first assumption on the exhibit has been
19 split into two there.
20 Q. It's been shortened to -- I'm sorry, the first one --
21 A. Nos. 1 and 2.
22 Q. These two have been reversed, correct?
23 THE COURT: They are combined.
24 A. One and 2 are combined.
25 Q. All right. The first one in your 1988 article deals
171
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 with relevant questions, correct? That guilty people will
2 answer deceptively?
3 A. Yes.
4 Q. And the second one is constructing control questions
5 rather than relevant questions, correct?
6 A. Yes.
7 Q. The third is the assumption that an innocent person
8 will be more disturbed by the control question than the
9 relevant, and that's the one you keep in both charts,
10 correct?
11 A. Right.
12 Q. And the fourth is the guilty person will be more
13 disturbed by the relevant questions.
14 A. Right.
15 Q. And you go on and say, "The first and fourth of these
16 assumptions which determine primarily how well the test will
17 work with the guilty, are not unreasonable, although there
18 may be circumstances where they do not hold," correct?
19 A. Yes.
20 Q. Do you still agree with that?
21 A. I think that's consistent with what I said today.
22 Q. I want to clarify that. Basically, the control
23 question test structure works well with the guilty. Your
24 main concern is how it works on the innocent; isn't that
25 correct?
172
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 A. No, I don't think so. I have a concern about how it
2 works on the innocent, but I also have a concern about how
3 it works on the guilty, especially because the guilty can
4 use counter-measures to augment the response of the control
5 question. That's a major problem for the third assumption.
6 Q. Is it fair to say at least this. Your biggest problem
7 is polygraph control question test is you believe it's
8 significantly biased against the innocent?
9 A. I think that's a fair statement of my belief prior to
10 all the information becoming available about
11 counter-measures, so I still think that's a big problem, but
12 I guess what I would say now is in addition, I am much more
13 concerned about how straightforward it is to use
14 counter-measures so that guilty people have the potential to
15 beat the test.
16 Q. So the bases of your objections have grown over the
17 years since you have testified that your main concern is
18 with the innocent?
19 A. Yes.
20 Q. Let's talk about what your current belief is as
21 reflected in this chart. These -- without belaboring the
22 point -- are basically assumptions a your part. I mean,
23 these -- your criticism of these theories underlying the
24 test are based not on data but on your belief that these are
25 incorrect assumptions, correct?
173
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 A. I'm not saying they are incorrect assumptions, I am
2 saying that for the test to work properly, they have to be
3 true, and I believe that it's not likely that these
4 assumptions are going to hold in so many of the cases that
5 you can get accuracy rates on the order of 90 to 95 percent,
6 which are the rates, for instance, that we are asking in
7 Honts report. Because they seem psychologically implausible
8 that they would hold that fast and steady so you could
9 guarantee such a high accuracy rate.
10 Q. So it's your belief that these are assumptions that
11 won't bear out in actual practice?
12 A. No, that's not fair either. It's that they won't bear
13 out all the time. What we could quibble about is what all
14 the time means, but I would say not enough of the time to
15 produce the accuracy rate on the order of 90 to 95 percent.
16 Q. Have you ever testified as a psychological or medical
17 expert in court?
18 A. No.
19 Q. Are you familiar with the term "reasonable scientific
20 probability"?
21 A. Vaguely.
22 Q. And "reasonable medical certainty"?
23 A. I have heard them used.
24 Q. And you know that those terms mean before you can give
25 an opinion as to a medical or scientific matter, you have to
174
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 be -- you have to believe it is more likely than not true,
2 51 percent true?
3 A. Uh-huh.
4 Q. Are you establishing a different standard here for
5 polygraphs than that more likely than not standard?
6 A. I believe I am.
7 Q. Do you believe, whether or not based on data -- I will
8 get to that in a moment -- that these assumptions will not
9 bear out in actual practice at least 50 percent of the time?
10 A. I think the assumptions will be accurate more than 50
11 percent of the time.
12 Q. All right. So as a matter of reasonable scientific
13 certainty, by the 51 percent standard, these will bear out?
14 A. Yes.
15 Q. All right. Now, isn't the way to resolve the
16 disagreement between you and those folks who agree with you
17 and the other polygraphy researchers who agree with the
18 assumptions to test it and see how it works in actual
19 practice?
20 A. Yes.
21 Q. That is the scientific method to conduct both lab
22 studies and field studies, correct?
23 A. Yes.
24 Q. Now, we're going to get a counter-measure separately,
25 but is that your only reservation about the third
175
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 assumption, that the guilty person will respond more
2 strongly to the relevant than the control questions?
3 A. No.
4 Q. What is the other objection you have?
5 A. I had a couple of other objections. One involved the
6 phenomenon of habituation, which involves the fact that a
7 guilty person who has been living with the accusation that
8 he committed a crime over an extended period of time is
9 likely to find the accusation, when it's made, less arousing
10 than it was initially, which would reduce the response to
11 the relevant question.
12 Another reservation I have is that there is no effort
13 made to determine, in fact, what the control question
14 covers, and it's possible, at least for some people, that
15 the control question would cover something that is of
16 greater concern to the person than the information that the
17 relevant question deals with.
18 Q. Now, those are your concerns. Those are not objections
19 based on data that you can cite; is that correct?
20 A. That's correct.
21 Q. There are no test data supporting the habituation
22 theory that it will change the way the guilty person
23 responds to the relevant as compared to the control
24 question, correct?
25 A. No one has specifically done a study like that.
176
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 Q. Now, in the laboratory studies that have been
2 conducted, there's a pretty high hit rate, even though there
3 is not a whole lot at stake in those tests; isn't that
4 correct?
5 A. The laboratory studies show quite a range of accuracy
6 figures. Some of them get higher rates than others.
7 Q. But most of them in the 80 to 90 percent range, don't
8 they?
9 A. A lot of them are in that range.
10 Q. All of them were then 51 percent?
11 A. Yes.
12 Q. And that's where something like $10 or $20 might be at
13 stake, correct?
14 A. Right.
15 Q. In general, you have taken the position that there is
16 no laboratory test that you will accept to confirm these
17 assumptions; is that correct?
18 A. Again, the issue isn't confirming the assumptions. The
19 issue is how far can you go with them. So my objection to
20 the laboratory studies is that you cannot use them to
21 estimate accuracy in the field. It's not that you can't use
22 laboratory studies to answer interesting experimental
23 questions that you can formulate hypotheses about, but if
24 your hypothesis is the polygraph will detect criminals at
25 this rate, and then you go do a laboratory study, you cannot
177
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 answer that hypothesis and not test it with a laboratory
2 study.
3 Q. Even if you use criminals in the laboratory study?
4 A. Even if you use criminals in the laboratory study.
5 Q. So basically you are saying there is no test that any
6 scientist in the world could devise in a laboratory setting
7 that will satisfy your misgivings about the assumptions?
8 A. I don't want to say that, but I will say there's no
9 test that's been done so far that would be suitable for
10 evaluating the accuracy of the polygraph in the field.
11 National Academies' conclusion is exactly the same.
12 Q. How would you formulate one? How would you suggest to
13 any researcher who is working, formulate a laboratory study
14 in a different way than they have done it that would have
15 you come in and say, "Now I agree with you?"
16 A. There is an example of such a study that was done years
17 ago, and you could imagine extending this kind of research.
18 I don't remember all of the details of it, but the gist of
19 it was they allowed people to take examinations and they had
20 rigged the way the responses to the examinations were
21 reported. So that if someone who gave an answer tried to
22 change it, that could be detected and it would be known that
23 they were cheating. These were very important examinations
24 that the people had to take. And so then later they could
25 then evaluate the people with polygraph tests to see whether
178
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 people cheated or not and they actually had ground proof
2 evidence as to whether or not they were cheating.
3 The problem with that study was, though, when they
4 confronted people about their cheating they all admitted it,
5 so they basically didn't have anybody to go give polygraph
6 tests to. But that would be an example of the kind of thing
7 that could be done that would involve a laboratory exercise.
8 It would be difficult to do for a number of reasons,
9 including the ethical reasons that one has to worry about
10 when you essentially entrap people that way, but that was an
11 example of a study done in Israel.
12 Q. That was one that didn't work, right?
13 A. Didn't work because people confessed.
14 Q. Not because there was anything proven wrong about the
15 assumptions?
16 A. That's correct.
17 Q. So other than that one kind of test that you say won't
18 work, is there anything you could suggest to the polygraph
19 researchers that they could do that would satisfy your
20 objection to the assumptions in the form of a lab study?
21 A. I didn't say that procedure won't work. I just said
22 that one study didn't work. You asked me could I think of a
23 type of study that would deal with this concern, and that
24 would be an example. So that particular study, which
25 involved not a large group of people, didn't work out the
179
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 way they had hoped.
2 Q. Have you ever conducted a study that would satisfy your
3 standards for a lab test?
4 A. No. I'm sorry, a lab test equal --
5 Q. Of the tests that we are talking about here, the New
6 Mexico control question test.
7 A. I am saying -- I have done laboratory research and
8 other people have too, and I support the conduct of
9 laboratory research and its publication, but not to estimate
10 field validity.
11 Q. And that gets back to the question that I was asking.
12 There's basically no laboratory study that you will accept
13 as a confirmation of the theoretical basis for the control
14 question test; is that correct?
15 A. Well, no, I thought we just went through that. I just
16 gave an example of a study that could have provided evidence
17 that would at least give you an idea of how the thing would
18 work in the field.
19 Q. And you would accept the results of the study if
20 anybody could figure out a way to make it work?
21 A. Well, in theory -- every study has problems, but that
22 type of study, if it showed the polygraph worked well, and
23 it was done well, then I would have to rethink -- I would
24 accept it, yes.
25 Q. You would. All right. Are you familiar with how
180
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 laboratory studies in psychology generally translate into
2 the field?
3 A. Generally they translate fairly well, but, you know, I
4 would say these laboratory studies translate fairly well.
5 There's just a difference between a study providing you with
6 information that's useful to the conduct of a field
7 polygraph test versus a study that says in the field, a
8 polygraph test has 93 percent accuracy or 94 percent
9 accuracy or whatever.
10 Q. All right. So you would agree the lab studies show
11 that this theory works, it's just a question of what
12 accuracy figure?
13 A. Yes.
14 Q. All right. And you agree that these laboratory studies
15 show it's more likely than not that the result will be the
16 correct one. That is, if someone is called innocent or
17 deceptive, and not in that inconclusive range, it's more
18 likely than not that's a good call?
19 A. At least slightly better than chance.
20 Q. At least better than chance?
21 A. Yes.
22 Q. You are just saying that you are not comfortable with
23 someone assigning an 80 or 90 percent figure for the
24 likelihood?
25 A. That's right.
181
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 Q. Are you comfortable with assigning a percentage figure
2 to the diagnosis of schizophrenia?
3 A. That just can't be done. It's not a matter of being
4 comfortable with it. It makes no sense.
5 Q. Are you familiar with other kinds of diagnoses that
6 psychologists and psychiatrists make regularly in courts of
7 law?
8 A. I am familiar with the fact that they make diagnoses.
9 Q. And testify about them in courts?
10 A. Yes.
11 Q. Based on such things as Rorschach test?
12 A. They probably do.
13 Q. And do you know what the accuracy rate of the Rorschach
14 test is?
15 A. It's -- you know, the studies on the validity of the
16 Rorschach indicate it has serious problems.
17 Q. More serious than your criticisms of the control
18 question test would reflect, right?
19 A. Yes.
20 Q. How about the thematic apperception test?
21 A. Same thing.
22 Q. Do you know what the accuracy rate of that test is?
23 A. You can't even estimate really an accuracy rate,
24 because there's no criterion. That's part of the problem
25 with these examples that you give. With a lie detector,
182
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 somebody is guilty or not. In theory it's possible to know
2 if somebody is guilty or not. With schizophrenia or giving
3 a Rorschach ink blot test, there is no equivalent to guilt.
4 There's nothing we can point to that serves as a criterion
5 that allows us to know if we are right or wrong. It's like
6 you said earlier, you can't take the picture of the brain of
7 the person with schizophrenia and say, "There is the
8 schizophrenia demon that makes the person that way." As long
9 as that's missing you can never say a diagnosis of
10 schizophrenia is 73 percent accurate or that the thematic
11 apperception test is 27 percent accurate because you have
12 nothing to compare it to coming up with the accuracy.
13 Q. You can say that with a number of psychological tests?
14 A. True.
15 Q. The MMPI?
16 A. Yes.
17 Q. The draw-a-person test?
18 A. Yes.
19 Q. Prediction of dangerousness?
20 A. That, though, it's at least in theory I possible to do
21 studies about. But I don't claim to know -- you could
22 release people that you made predictions about and see what
23 they do and ten years later see if you were right or wrong
24 but I don't know if they do those studies.
25 Q. Do you know what the accuracy rate of predictions of
183
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 dangerousness is?
2 A. I don't.
3 Q. Are you familiar with the U.S. Supreme Court case a few
4 years ago that said 30 percent accuracy of prediction of
5 dangerousness is enough to allow a psychologist to testify
6 in support of executing the defendant?
7 A. I am not.
8 Q. And do you know what the accuracy rate of stranger
9 eyewitness identification test is?
10 A. No.
11 Q. How about informant testimony?
12 A. No.
13 Q. You are aware there are a lot of things in the human
14 life and the courtroom that we base judgment on that we
15 can't know to a specific percentage degree or we can't
16 guarantee as 100 percent positive with no possibility of
17 error, right?
18 A. Yes.
19 Q. And what we do is take everything that may give us more
20 information than we had before and look at the whole,
21 correct?
22 A. Yes.
23 Q. And that's a reasonable way to make human judgments?
24 A. It's a reasonable way to make human judgments, but
25 these are not all the same thing in the sense that if you
184
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 give a test that's not -- has unknown validity and used by
2 someone along with other instruments to formulate their
3 opinion, the issue is that person's opinion. When you take
4 a polygraph test, the test itself gives you a verdict, and
5 it's not assisting someone's opinion. It's like a
6 laboratory test, like a DNA match or fingerprint match. So
7 it's a little bit different from the ink blot test which may
8 not have much validity but is used by someone that says,
9 "When I gave this person the ink blot the responses were
10 bizarre and it helps me decide that the person has
11 schizophrenia."
12 Q. Are you familiar with the accuracy rate of handwriting
13 comparisons?
14 A. No.
15 Q. How about ballistics identification?
16 A. I am not familiar with it. I assume it's high.
17 Q. Let me show you a document that's in Volume 2 of the
18 Honts affidavit references.
19 MR. DANIELS: Your Honor, I have marked those, to
20 help at least me remember if not everyone else, H-1, H-2 and
21 H-3, the H signifying Honts, and the 1, 2 and 3 signifying
22 the volume of the affidavit it references. These were on
23 the data CD and printed out and produced in the three
24 volumes. This is the second No. 2.
25 THE COURT: Let me inquire, Mr. Daniels, there's
185
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 no wrong answer, I am trying to get a feel for timing. If
2 we are going to go as far as 5:30 I was going to take
3 another short recess just to let Jan stretch her fingers.
4 MR. DANIELS: I will be through in another hour.
5 THE COURT: Why don't we take ten minutes.
6 (Note: Court stood in recess at 4:10 to 4:20).
7 Q. (By Mr. Daniels) Dr. Iacono, I think we were talking
8 about the applicability of laboratory studies to real life
9 in the field of psychology. Are you familiar with the
10 research that's been done on the correlation between
11 laboratory studies in psychology and their application to
12 the real world?
13 A. It's not a literature I keep up with.
14 Q. I'm sorry?
15 A. It's not a literature I keep up with.
16 Q. It's not a literature you keep up with?
17 A. I have seen this paper that you have here.
18 Q. Are you familiar with Current Directions in
19 Psychological Science?
20 A. Yes.
21 Q. That a peer-reviewed publication in your profession?
22 A. It's not exactly a peer-reviewed publication, because
23 all of the articles are invited, but it's a little bit
24 different than a publication where you take a paper that you
25 have written and you submit it and it's reviewed. This
186
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 journal, the people who make contributions to it are invited
2 to submit articles.
3 Q. Doesn't it require that the editor send out the article
4 for review by two or three independent specialists in the
5 field before publishing it?
6 A. They do, but it's a friendly review.
7 Q. A friendly review?
8 A. Yeah.
9 Q. Instead of a rubber stamp?
10 A. It's not a rubber stamp. It's a constructive effort to
11 try to improve the paper, but I think if you wrote a
12 completely unacceptable paper it would be rejected.
13 Q. The peer-reviewers comment on the quality of the
14 research and the statistical analysis and the reasonableness
15 of the conclusions drawn and the appropriateness of the
16 article, correct?
17 A. That's what peer review it.
18 Q. But is this a publication, one that you read at all?
19 A. Yes.
20 Q. Are you familiar with the study by Craig Anderson,
21 James Lindsey and Brad Bushman from the Department of
22 Psychology of University of Missouri and Iowa State
23 University in 1999?
24 A. I remember seeing it when it came out, but I haven't
25 gone over it recently so I wouldn't want to say I am
187
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 familiar with it in the sense that I could tell you now the
2 details.
3 Q. Would you agree with their conclusion that the
4 psychological laboratory generally produced psychological
5 truths rather than trivialities?
6 A. I agree with that conclusion, and I would agree it
7 applies to laboratory studies on polygraph testing. When
8 you consider why the laboratory studies were done and what
9 kinds of questions they try to answer. An example, a study
10 by Horowitz, Honts and Raskin published recently was a
11 laboratory study. It's one of the ones they include in
12 their tables that show that laboratory studies have high
13 accuracy, but the study was done to show how different types
14 of comparison questions work. It, in particular, showed on
15 the relevant/irrelevant test the irrelevant comparisons
16 worked very poorly.
17 So the way I interpret what these people have in this
18 paper is that if you take that Horowitz paper and you say
19 "how does this generalize to real life," you can generalize
20 to real life that the different types of comparison
21 questions that Horowitz studied are going to work in real
22 life in a similar fashion, but that's different than saying
23 the Horowitz study showing that the polygraph is highly
24 accurate in real life. I think we are confusing two things
25 here.
188
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 Q. I'm not confusing it. Aren't there a number of studies
2 that have been -- laboratory studies that have been
3 published in peer-reviewed and other professional literature
4 relating to the validity of the assumptions that you
5 criticized here in theory?
6 A. I wouldn't dispute that. I don't think that the point
7 of the studies is to show how the polygraph works in the
8 field. I think they ask other questions.
9 Q. Let's take it one step at a time. Do you agree that
10 there has been a large number of publications of studies of
11 laboratory tests on the polygraph? The control question
12 polygraph?
13 A. Yes.
14 Q. You do?
15 A. Yes.
16 Q. And --
17 THE COURT: You may or may not have intended this,
18 but the previous question that the doctor disputed had to do
19 with peer review publications and the last one you left the
20 term "peer review" out. So you were asking if there's a
21 large number of publications.
22 MR. DANIELS: I will get to both of those.
23 Q. There are basically two ways for peers to review the
24 work of others in scientific fields, isn't there, Doctor?
25 A. Probably dozens of ways.
189
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 Q. Probably lots of ways. A central feature, and the one
2 talked about in the Supreme Court in Daubert, is you ought
3 to be suspect of a theory someone came up with in a dark
4 basement and not to let the world look at and review. Would
5 you agree with that as a scientist?
6 A. I didn't interpret what they were saying to mean that.
7 I interpreted it to say if you had a novel procedure and
8 someone is arguing it works, you can have more confidence
9 that scientists accept it if the study is published in a
10 peer-reviewed journal that has undergone scientific scrutiny
11 in the process.
12 Q. There are two ways for scientists to examine each
13 other's work. One is in the strictest kinds of peer review
14 journals, correct?
15 A. Yes.
16 Q. And those are -- would include -- let me give you the
17 names of some and see if you agree with me on that. The
18 Journal of Applied Psychology is a peer-reviewed journal,
19 correct?
20 A. Yes.
21 Q. And there have been studies reporting positive results
22 for control question tests and laboratory studies in that
23 publication; isn't that correct?
24 A. Yes.
25 Q. The Journal of General Psychology; is that a
190
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 peer-reviewed journal?
2 A. As far as I know.
3 Q. You don't know about it?
4 A. I don't really know exactly how the journal works, but
5 I assume it is.
6 Q. Psychophysiology?
7 A. That's a peer-reviewed journal.
8 Q. There have been articles about laboratory studies
9 showing positive results for the validity of the controlled
10 question polygraph in Psychophysiology; isn't that correct?
11 A. Well, again, I guess I dispute your interpretation of
12 why the study was published. You are saying the study was
13 published because it deals with the validity of the
14 polygraph. I believe the studies are published because they
15 deal with issues that are related to how the polygraph
16 works. I mean, we could go through each study and talk
17 about what it says the hypotheses were that were tested and
18 the conclusions, but again, I don't think the studies were
19 done because the hypothesis was by doing the laboratory
20 study we can estimate how it works in the field.
21 Q. The Journal of Police Science Administration. Is that
22 a peer-reviewed publication?
23 A. As far as I know, no. The peer reviewers are certainly
24 not the scientists one would think of as competent to review
25 in this field.
191
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 Q. Current Directions in Psychological Science?
2 A. That's the journal we are talking about with the
3 projection.
4 Q. Psychological Bulletin?
5 A. Peer-reviewed.
6 Q. The Journal of Research and Personality?
7 A. Yes.
8 Q. Law and Human Behavior?
9 A. Yes.
10 Q. Now, as I understood your direct examination, you
11 didn't think there was a consensus of the scientific
12 community on the polygraph but you did agree there has been
13 publication in peer-reviewed journals, correct?
14 A. There have been publications of peer-reviewed journals
15 and there's no consensus, right.
16 Q. Now, let's separate the lab studies and the field
17 studies here. By the way, were you -- you go to the
18 meetings of the Journal of the Society for
19 Psychophysiological Research?
20 A. Yes.
21 Q. The society is probably the premier organization for
22 experts in the field of psychophysiology; isn't that
23 correct?
24 A. Yes.
25 Q. Psychophysiology is the parent science of the polygraph
192
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 techniques that we are talking about here?
2 A. Yes.
3 Q. The experts in the field in the relevant scientific
4 field would be psychophysiologists, correct?
5 A. It would be one group, yes.
6 Q. Is there another group?
7 A. Well, again, I think, for instance, the people in the
8 National Academies Panel are experts in the ability to
9 review the methods that a scientist uses to produce
10 psychological tests or tests that are purported to do the
11 sorts of things that a polygraph test does, so I think there
12 are different relevant scientific communities. But the
13 Society for Psychophysiological Research is certainly an
14 obvious one because the technique is a psychophysiological
15 technique.
16 Q. Do you recall the presentations that were made at the
17 1994 annual meeting of the society?
18 A. No.
19 Q. Do you recall the presentation by Dr. John Kircher,
20 Dr. Steven Horowitz, Dr. Raskin and Dr. Honts on the
21 validation of laboratory studies for field application?
22 A. No, but those are just abstracts. Those don't undergo
23 peer review.
24 Q. I am not asking if this is peer-reviewed. I am asking
25 about the usefulness of lab studies for application to the
193
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 field as a matter of what you rely on in your profession?
2 A. I'm sorry, I don't know what the question is exactly.
3 Q. Well, is it accepted in the field of psychology that
4 laboratory studies have a great deal of effectiveness in
5 determining how real people function psychologically and
6 psychophysiologically in this case?
7 A. Yes, I think that's true.
8 Q. And the Horowitz, Kircher and others presentation
9 specifically applied that to laboratory studies of
10 applicability in their field, correct?
11 A. Again, I'm not quite sure I understand the question. I
12 just don't want to be seeming to say that I think that this
13 abstract shows that polygraph testing is accurate in the
14 field. I don't even think it addresses that issue.
15 Q. Believe me, I don't have any belief that I am going to
16 change your mind here.
17 A. I am saying I don't think the study, as I understand
18 it, says what you think it says, so I don't want to say
19 something that I am not comfortable with.
20 Q. Let's look at some of the lab studies that have been
21 conducted. Did you review the Honts affidavit in this case?
22 A. I never saw a Honts affidavit.
23 Q. All right. Have you seen the brief that was filed by
24 the Committee of Concerned Social Scientists in the Supreme
25 Court in Scheffer?
194
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 A. I have seen it.
2 Q. And have you seen -- that was submitted by a number of
3 researchers in the field, correct? Do you recall going
4 through the --
5 A. I thought it was written by Dr. Honts.
6 Q. You see the list of names of people that signed on to
7 that?
8 A. I saw that they signed on to it, but I assumed it was
9 his work.
10 Q. Did you see the names of the scientists who signed on
11 to this?
12 A. Yes.
13 Q. And those are researchers in this field of
14 psychophysiology and its application to control question
15 polygraph, correct?
16 A. I don't think that's true.
17 Q. And do you recall the summary of some of the laboratory
18 studies of the comparison question test that have been done?
19 A. Well, I see it right there. You presented it for me.
20 I don't have to recall it.
21 Q. I wondered if you have seen it before.A. Yeah.
22 Q. Now, no matter what the lab studies show, though, your
23 theory is that they are of no value in predicting what
24 happens when you polygraph a person in a criminal case,
25 correct?
195
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 A. No, again, that's not what I want to say. What I want
2 to say is they are of no value in arriving at a precise
3 quantitative estimate of how accurate the polygraph test is
4 in the field. I don't believe you can take what's in the
5 bottom row of the table, which might say for guilty people
6 the polygraph is 80 percent accurate in the laboratory
7 studies, and now say in the field the polygraph is 80
8 percent accurate. I don't believe you can say that. I
9 believe that the studies provide an overestimate of the
10 accuracy of polygraph testing in the field. I don't even
11 believe the laboratory studies deal with the same
12 psychological processes going on in the field.
13 Q. Well, we are talking about how someone would ever
14 respond to your theory that the underlying assumptions don't
15 bear out, not specific percentages.
16 A. No, that wasn't my testimony. My testimony was that
17 the underlying assumptions don't bear out all the time.
18 Q. I agree with that.
19 A. What we quibble about is what all the time means. What
20 the exact number is. That's what this is about. The people
21 in the polygraph profession say this happens 90 percent of
22 the time. I don't believe it happens 90 percent of the
23 time, and I don't believe the evidence that exists supports
24 that conclusion.
25 Q. All right. So your objection is not to the
196
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 propositions listed on this chart of assumptions but to any
2 one that's trying to give a specific percentage figure of
3 what is called a hit rate. Now, you understand what hit
4 rate is because you used that, correct?
5 A. Yes, but to be clear, what I am saying --
6 Q. The hit rate is correctly calling someone truthful or
7 deceptive, right?
8 A. Right.
9 Q. All right. Go ahead.
10 A. So for the test to work -- if these assumptions were
11 true for every single human being, the accuracy of the
12 polygraph test would be 100 percent.
13 Q. It would be the most amazing evidence we have presented
14 in court then.
15 A. But if those were true, you would never make an error
16 on the polygraph test, so they are not always true, because
17 the assumptions can't always hold. I find the assumptions
18 implausible that they would hold a significance enough so
19 that these accuracy rates that people refer to as 90
20 percent, for instance, could possibly be true. I don't
21 believe human nature works that way, and I don't believe the
22 evidence supports that conclusion.
23 So that's my point, is that you can't -- for a
24 procedure that's based on these kinds of assumptions, for me
25 it's absurd to think they would produce a test with 90
197
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 percent accuracy, and I don't believe the evidence supports
2 that conclusion.
3 Q. All right. I may understand this better than before.
4 So to read this the way you are reading it, we should put in
5 here "always formulate questions" or that people will
6 "always react", or that will respond with "100 percent
7 accuracy", those kinds of things. That's what your
8 criticism is, a belief that the assumption is that this
9 works in every case, correct?
10 A. Yes.
11 Q. Oh. All right. Do you know anybody researching in the
12 field who claims that this is an error-free procedure and
13 that people will, in every instance, 100 percent of the
14 time, react in those ways?
15 A. Do I know people in the field who say that? Yes, I am
16 sorry to say I do.
17 Q. The psychophysiologists?
18 A. No, not psychophysiologists. Practitioners say that.
19 Q. That would be just wrong, wouldn't it?
20 A. It would.
21 Q. And none of the published studies by the
22 psychophysiologists have claimed they a 100 percent accuracy
23 rate without a chance for error?
24 A. On the overhead you just put up from the Committee of
25 Concerned Scientists, whatever they are called, I thought
198
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 there was 100 percent figure in some of those studies.
2 That's why I don't think you can use the studies that way.
3 I agree with what you said in the question, but there are
4 studies that people float these numbers out there that say
5 things like 100 percent.
6 Q. And argue from that that because there's 100 percent in
7 this particular study it's always 100 percent?
8 A. I am saying that's not the case. You said is there
9 anything out there that says it's 100 percent and I am
10 pointing out on your overhead it said it was 100 percent
11 accurate so that's one example of 100 percent accuracy.
12 Q. The study didn't say that polygraph is 100 percent
13 accurate. That was the percentage in that particular study,
14 correct? Correct decisions? In fact, you had 100 percent
15 accuracy in one of your studies, didn't you?
16 A. For original examiners.
17 Q. And you wouldn't hypothesize from that that in every
18 instance the hit rate is going to be 100 percent?
19 A. I would not, you are right.
20 Q. And did you read anything in here that said because one
21 study up here happened to have 100 percent as compared to
22 the others, that we can assume that those basic assumptions
23 will always hold true 100 percent of the time?
24 A. I'm sorry, these are laboratory studies.
25 Q. These are laboratory studies. We will go to field
199
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 studies next.
2 A. I am just saying that the laboratory studies don't
3 provide an adequate opportunity to answer this question.
4 Q. I understand that that's your theory. I'm going into
5 what has been done both in laboratory and field studies.
6 Let's talk about the field studies. You have said that the
7 laboratory studies simply can't be relied on for field
8 studies of people accused of crimes, correct?
9 A. Again, just to be clear, to come up with a precise
10 quantitative estimate, the laboratory studies are useful for
11 answering questions about things related to how polygraph
12 tests are conducted. But early laboratory studies that were
13 done in Dr. Raskin's lab, for instance, showed that the test
14 would even differentiate guilty from innocent people at all.
15 In the early papers that was the point of doing the research
16 because no one did that before. That's why it was
17 published, it showed you could differentiate guilty from
18 innocent, not because it showed it had such and such
19 accuracy.
20 Q. Because it showed you differentiated.
21 A. Yes. From that I would agree that in real life you can
22 make these kinds of differentiations. That's how I see you
23 use a laboratory study that you can generalize the results
24 to real life. I don't think you can use them to come up
25 with a precise quantitative estimate.
200
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 Q. I understand that's the point you are making. You
2 can't just place a number on it.
3 A. Right.
4 Q. Now, in designing a field study, would you want to use
5 actual criminal suspects?
6 A. Yes.
7 Q. In fact, some of those laboratory studies used criminal
8 suspects in case they somehow functioned differently from
9 the other people in the laboratory studies; isn't that
10 correct?
11 A. I'm not sure. They definitely didn't use people who
12 were being tested because they were suspected of being
13 criminals in a crime that I am aware of. I mean, they
14 weren't being tested on an issue related to their own
15 criminal activity.
16 Q. All right. Let's take those one at a time. First,
17 Dr. Honts and Dr. Raskin and others have gone out to prison
18 populations and conducted laboratory studies with those
19 people, correct? Just as you did on one occasion, correct?
20 A. Yes.
21 Q. But you criticize those studies because they weren't
22 being tested with respect to the crime they were charged
23 with. It was a hypothetical situation, correct?
24 A. I'm not criticizing for that reason. I am pointing it
25 out.
201
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 Q. You find fault with them, saying that we can't rely on
2 the studies?
3 A. No, I'm sorry. I thought what you were saying was that
4 in these studies, laboratory studies test criminal suspects
5 in a way that implies they are being tested on the crime
6 they are suspected of. I want to be clear that's not what
7 they did.
8 Q. I am taking it one at a time. First we take criminal
9 suspects and design lab studies to try to determine if they
10 react somehow differently?
11 A. No, they are prisoners. To make it simple we say they
12 take people convicted of crime and put in prison and ask
13 them to participate in a polygraph study.
14 Q. Just as you did?
15 A. Just as I did.
16 Q. That's one you find fault with, correct?
17 A. What one?
18 Q. Using real in people in prison as the subjects for the
19 laboratory study?
20 A. No, I don't --
21 THE COURT: Time out. Nothing personal. The only
22 reason I am interrupting at all is I am concerned about a
23 time crunch. I think you all are sort of at cross-purposes.
24 This is what I heard. You asked the doctor if you are going
25 to design a field study you would want to involve folks that
202
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 are charged with crimes or criminals. Then somehow it got
2 to a question of well, people who use criminals in lab
3 studies, and his only point, I think, was well, not in the
4 sense of investigating the individual crime. That's the
5 only point.
6 MR. DANIELS: Maybe I didn't make it clear, Your
7 Honor. What I wanted to get to was those lab studies, which
8 had used real prison populations, and bring out the fact
9 that that doesn't satisfy his objections, and then get into
10 where you study actual cases.
11 THE COURT: The only thing I heard in terms of
12 objections or satisfying an objection is I didn't hear the
13 doctor, correct me if I am wrong, say he was critical of
14 them in the sense that -- well, this is what I am hearing
15 from the doctor. You do the lab study, you get interesting
16 information that doesn't necessarily mean that the accuracy
17 rate in the lab study will accurately translate to real
18 life.
19 MR. DANIELS: The exact percentage. That's right.
20 I think that's what he said.
21 THE WITNESS: That's right. That would apply if
22 the lab study is done on prisoners.
23 MR. DANIELS: That's what I wanted to establish.
24 Those are also -- now I will get into the real life case
25 studies and go with what's wrong with those.
203
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 Q. All right. So the people who have been researching in
2 the field have gone out to study real life cases in the
3 field, correct?
4 A. Yes.
5 Q. And, in fact, you participated in one of those?
6 A. Yes.
7 Q. What was the -- that was the one with the Royal
8 Canadian Mounted Police?
9 A. Yes.
10 Q. With the high detection rate of guilty?
11 A. Again, it didn't show a high detection rate of guilty.
12 It showed that only the guilty -- that the ground truth
13 criterion that we were looking for, which was a confession,
14 was strongly linked to guilty people who failed the
15 polygraph test and confessed. That's not the same thing as
16 saying that the test was showing a high accuracy rate.
17 Q. There have been several ways to conduct field studies,
18 the validity of the procedure that was worked out in the
19 laboratory; isn't that fair to say? The control question
20 test procedure?
21 A. Well, I don't -- there have been field studies with
22 control question tests but not of the control question test
23 that was worked out in a laboratory study.
24 Q. The same theory that we have been talking about all
25 along here? Hasn't that been tested in the field?
204
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 A. I'm not sure what you mean. But the control question
2 test has been tested in the field.
3 Q. That's what I am talking about. The one whose
4 theoretical underpinnings you are objecting to.
5 A. Correct.
6 Q. And that's what I am talking about throughout the day
7 here, all right? One way that has been done, has been to
8 set up panels of experts to review all of the case materials
9 and come up with a judgment as to whether the person was
10 guilty or innocent, deceptive or truthful, if you will, and
11 then compare those judgments to the results of polygraph
12 exams in the case, correct?
13 A. Yes.
14 Q. And those had a general correlation to the results of
15 the laboratory studies, didn't they? The kinds of ranges of
16 numbers in the laboratory studies?
17 A. They report -- they tend to report high accuracy rates.
18 One of them rated a particularly low accuracy rate for
19 innocent people, but those studies, I don't think anybody
20 accepts panel studies as valid anymore.
21 Q. Because you think that human beings make mistakes in
22 making judgments about things like guilt or innocence?
23 A. Well, there's actually a couple of issues. One is that
24 in the panel studies there's been two panel studies that
25 have been done that deal with this issue. One of which has
205
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 been published and the other wasn't. In both of them the
2 people in the panel were given the confessions that were
3 made if the person failed the polygraph and confessed. So
4 as soon as you do that, you turn the panel study into a
5 confession study. It really isn't any different.
6 Another factor that one of the grants that I got from
7 the Department of Defense Polygraph Institute was actually a
8 grant to determine how accurate panels are. So what we did
9 is we took polygraph case files where there was a confession
10 and we removed all of the evidence of the confession from
11 the case file and gave all of the information to the panel.
12 We had three types of panels. Defense attorneys,
13 prosecuting attorneys and laypeople. We asked them to
14 review the evidence now that was in these files to determine
15 ground truth based on a panel consensus, and they were
16 accurate less than 25 percent of the time.
17 What was unique about our study was that we actually
18 did a study where you could verify exactly how good panels
19 work and it showed they don't work, so I don't think panel
20 studies add anything to clarity in terms of the accuracy of
21 lab.
22 Q. There were panel studies where judges participated,
23 weren't there?
24 A. I think so.
25 Q. Basically you have taken the position that the panels,
206
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 which sort of replicate what happens in the courtroom, have
2 such low accuracy rate that they are not useful to compare
3 the polygraph results?
4 A. That's right, because they are even less accurate than
5 the polygraph test unless they get confessions or something.
6 The point is as a criterion they are highly flawed, so you
7 can't accurately estimate how well the polygraph works if
8 you are comparing it to something that itself doesn't work
9 well.
10 Q. That's right. The panel judgments of all of the
11 remaining evidence in the case --
12 A. Right.
13 Q. -- come out to be less reliable than the polygraph?
14 A. Well, I think so, yes.
15 Q. So then there have been other studies then that you use
16 the confession comparison that you had addressed before, and
17 you found that to be lacking.
18 A. Yes.
19 Q. Now, there were a number of assumptions you made, and I
20 don't know if we have time to get to all of these.
21 THE COURT: Let me inquire timing-wise. Your
22 flight leaves?
23 THE WITNESS: 8:00 tomorrow morning.
24 THE COURT: Current flight.
25 MR. COX: We sent somebody out to make a later
207
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 flight.
2 THE COURT: Listen, I have talked to the staff.
3 We are prepared to go past 5:00. There's not a problem.
4 But the thing is I don't want to be cutting Mr. Daniels off.
5 This is crucial stuff. Mr. Rein hasn't had a chance to ask
6 questions, and you get to redirect.
7 MR. DANIELS: How much redirect do you want?
8 MR. COX: Until I have heard the cross I can't
9 say.
10 MR. DANIELS: I will try to wrap it up in 15
11 minutes. I don't suspect I will have a Perry Mason moment.
12 THE COURT: I'm not sure I have seen one, but in
13 any event, if you want to take another hour, Mr. Daniels, I
14 don't want to cut you off, is the thing, and I don't know
15 what Mr. Rein has. As I said, we can go for a while more
16 tonight but --
17 MR. DANIELS: Either way. I can wrap this up in
18 15 minutes because we have other witnesses to address the
19 same data anyway.
20 MR. COX: You mentioned Mr. Rein. I was under the
21 impression this was going to be the cross-examination.
22 THE COURT: Well, as I understood the Supreme
23 Court order, it said one person for the State. It did not
24 so limit the petitioners.
25 MR. DANIELS: We will do one person per witness,
208
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 Your Honor.
2 MR. COX: That's all I need to know.
3 MR. DANIELS: Do you want me to do a 15-minute
4 wrap-up, I will.
5 THE COURT: If you want to, that's fine. If the
6 doctor will be here tomorrow morning anyway, I'm not sure
7 that the timing of tomorrow -- why don't you go for another
8 15 minutes. I don't mind staying late but I don't want to
9 plan on going until 7:00 or 8:00.
10 Q. (By Mr. Daniels) Just a few questions on your
11 confession analysis. You and other researchers in the field
12 have debated the underlying assumptions in this confession
13 analysis before, haven't you? This is not a new theory that
14 you are suggesting?
15 A. No, but I don't know that there's much of a debate. I
16 don't know of any debate.
17 Q. There are a number of assumptions that you make about
18 what the polygraphers did in those particular cases in the
19 course of examining suspects and deciding not to investigate
20 further or not attempting to get a confession or whatever,
21 correct?
22 A. No, they are not assumptions, they are facts. That's
23 what we learned in the study with the RCMP. The whole study
24 is meant to describe the problem.
25 Q. Whatever it is that you found out there would be in the
209
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 article you published about it?
2 A. Yes.
3 Q. Now, one of the things you mentioned earlier is that
4 you theorized that where there was physical evidence that
5 that was of no value, because you assumed any physical
6 evidence was obtained from the confessions; is that correct?
7 A. Right.
8 Q. And are you telling us the article reflects that in
9 every case where the police have physical evidence that the
10 source of it was the confessions rather than an independent
11 investigation?
12 A. As I understand, there's only one study that did this,
13 and as I understand the way the study was done, they used
14 the physical evidence to corroborate the confession. There
15 were no instances where there was physical evidence in the
16 absence of a confession. I assume that means the confession
17 led them to the physical evidence because the person not
18 only confessed but confessed the particulars in a way that
19 provided for the recovery of the physical evidence.
20 Q. Let's take a look at the base rate argument, some of
21 the assumptions underlying that. I believe that was Exhibit
22 4; is that correct?
23 THE COURT: That's what my note showed.
24 Q. Now, that also has a number of assumptions that aren't
25 established by empirical data; is that correct?
210
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 A. There are two assumptions, that's it. The base rate is
2 90 percent and the polygraph test is 90 percent accurate.
3 Q. All right. As to the first one, where do you get that
4 base rate?
5 A. I made it up.
6 Q. And that's basically presuming guilt with an arbitrary
7 percentage, correct?
8 A. Yeah.
9 Q. Of people who were accused in criminal cases?
10 A. It's an arbitrary percentage, yes.
11 Q. And it's a presumption that if you are charged, there's
12 a 90 percent likelihood you are guilty?
13 A. That's what that presumes.
14 Q. And then also implicitly assumes that all of the people
15 who choose to take a polygraph under the New Mexico
16 procedure, that all of the people charged take a polygraph
17 rather than there being some subset of the accused people
18 who make a decision that they want to take a polygraph,
19 correct?
20 A. I guess you could say that. I mean, it's not that
21 complicated. It's -- the point of the example is to make an
22 illustration, but it's true, people that are asked to take a
23 polygraph test could be ones who are more likely to be
24 innocent than people who are not asked to take a polygraph
25 test. We have no way of knowing. That's another assumption
211
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 that makes this more complicated.
2 Q. You don't have any data on what percentage of accused
3 people in criminal cases in New Mexico make a decision to
4 submit themselves to a polygraph examination?
5 A. No.
6 Q. And you don't know how many of those are, in fact,
7 guilty or innocent, correct?
8 A. No.
9 Q. And that presumption is -- or assumption -- is pretty
10 central to this whole analysis down here, isn't it?
11 A. Not really. It's -- the real reason I picked 90
12 percent as the base rate is because I did, I guess, your
13 side the favor of picking the polygraph as 90 percent
14 accurate. We could pick a lower accuracy figure, say 70
15 percent, base rate of 70 percent and it works the same.
16 Half the people who passed the test are guilty. You can
17 adjust the numbers any way you want to get anything you
18 want. This is a convenient way to illustrate the principle,
19 not the fact of how this works out.
20 Q. The bottom line conclusion will flow from the kind of
21 figures you plug in the assumptions up above, isn't that
22 fair to say?
23 A. That's right. If you lower the 90 percent accuracy for
24 the polygraph it works out worse in the sense that more
25 mistakes are brought before the Court.
212
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 Q. Is that sort of also related to the friendly
2 polygrapher hypothesis, that if a defendant and his attorney
3 arrange to hire a polygraph examiner, that the defendant is
4 not going to feel as much of a threat as those people did
5 with the $10 or $20 at stake in the laboratory experiments,
6 correct?
7 A. No, I am saying they will not experience as much of a
8 threat as they would if the police officer was doing the
9 test.
10 Q. Now, that has no published data to support the theory;
11 isn't that correct? That's a theory unsupported by any
12 study?
13 A. I think the accurate way to put it is that all of the
14 studies that have been done, have been done with police
15 officer polygraph examiners so we draw the line there in
16 terms of knowing how the test works in any other
17 circumstances. So if you extend it then to the friendly
18 polygraphs, you have no evidence to make the extension. You
19 have to make the presumption it's just the same.
20 Q. But you know of no study, and you certainly conducted
21 none, to support the friendly polygrapher theory, it's just
22 a theory?
23 A. It's just a theory.
24 Q. And have you seen any publications by people who have
25 actual experience in testing criminal defendants to
213
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 determine whether that assumption bears out?
2 A. I am not sure what you mean. People like Raskin and
3 Honts have written about their opinion about the friendly
4 polygrapher.
5 Q. Now, with regard to your base rate argument, doesn't
6 that apply to all diagnostic tests with any error rate that
7 would be conducted on a criminal defendant?
8 A. It would apply to any test.
9 Q. Let's say handwriting comparisons are 75 percent
10 accurate when you have a base rate of 50 percent. If you
11 assume that 90 percent of the criminal defendants are
12 guilty, then your same analysis would apply whether or not
13 the defendant should be able to use handwriting comparisons,
14 correct?
15 A. That's right.
16 Q. Or blood tests of various kinds. They have a
17 percentage of error, correct?
18 A. That's exactly true.
19 Q. Any medical or scientific diagnostic test with any
20 error rate would apply that way where you would multiply the
21 90 percent presumption of guilt against it and determine
22 whether you end up with a figure below 50 percent, correct?
23 A. I'm not sure what you mean by a figure below 50
24 percent. The point is the base rate argument would apply
25 for any test. There's nothing peculiar about polygraph
214
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 tests that makes it unique.
2 Q. Right. And isn't that something that a finder of fact
3 ought to take into account in evaluating any kind of
4 diagnostic test, instead of having the diagnostic test kept
5 from him?
6 A. That, I'm not sure about. I guess the way I would
7 think about it is that if we knew what the base rate was,
8 and we could tell that it was going to bias the results of
9 the way the evidence is submitted in court, and it seems to
10 me it would make more sense as a point of law for the Court
11 to make a decision about the appropriateness of it than to
12 have a jury try to figure out whether or not all the
13 statistical stuff makes sense, but I guess that would be my
14 opinion.
15 Q. And won't the base rate vary according to the other
16 evidence in the case?
17 A. The base rate is just a fact. I mean, in theory we
18 could determine, let's say if you believe in God, God knows
19 what the base rate is. It doesn't vary with anything. We
20 could go through how many criminal prosecutions there have
21 been in New Mexico in the last ten years and God knows how
22 many of the people were guilty and how many were innocent.
23 If God would tell us that number and we could plug that into
24 it example we could have it, but we don't.
25 Q. Without the assistance of the Almighty we have to deal
215
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 with what humans can do, right?
2 A. Right.
3 Q. And make human judgments?
4 A. Right.
5 Q. And you don't understand polygraph to be used in New
6 Mexico to be the sole determinant of guilt or innocence, do
7 you?
8 A. No.
9 Q. Just one piece of evidence to be considered along with
10 all of the rest of it?
11 A. Right.
12 Q. The same with all of the other imperfect evidence,
13 correct?
14 A. Yes.
15 Q. And would you agree that the likelihood of guilt in a
16 case with good DNA evidence is stronger than the likelihood
17 of guilt in a case with stranger eyewitness identification,
18 for example?
19 A. Yes.
20 Q. So if you were trying to figure percentages of guilt,
21 the DNA cases would have a higher percentage of guilt in the
22 abstract, correct?
23 A. I don't know what you mean by a percentage of guilt.
24 Q. Percentage of likelihood before you take the polygraph
25 that the person is guilty. In other words, are there six
216
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 out of ten in DNA cases that are guilty or nine out of ten
2 or 99 out of 100 where you have good DNA evidence linking
3 the defendant.
4 A. I am kind of lost here. I'm not quite sure I follow
5 how DNA evidence seems to be linked to the base rate. I am
6 not getting that connection.
7 Q. What you were assuming was 90 percent of criminal
8 defendants are guilty?
9 A. Right.
10 Q. Wouldn't that vary depending on what else is in the
11 case?
12 A. No, it doesn't. That's what I was saying about God
13 knows what the number is. Doesn't vary on anything.
14 Q. You wouldn't say there's a higher percentage of
15 defendants in DNA identification cases that are guilty as
16 opposed to defendants in eyewitness identification cases?
17 THE COURT: Just to help, another area where I'm
18 not sure there's a connection, what I am hearing him say is
19 the base rate is the percentage of folks who are guilty.
20 You are talking about folks who are likely to be convicted.
21 MR. DANIELS: He says 90 percent are guilty.
22 THE COURT: Correct me if I am wrong, but what I
23 am hearing is the base rate is an absolute. Whether or not
24 they are convicted is another issue.
25 THE WITNESS: Whether or not they have DNA evidence
217
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 is another issue. Whether or not there is an eyewitness is
2 another issue.
3 MR. DANIELS: I am getting at his picking of a
4 figure of how many defendants come in before the Court are
5 guilty and using that to multiply.
6 THE COURT: All I heard him say is it's an
7 absolutely arbitrary number. He picked it out of thin air
8 to illustrate the point that it is not -- correct me if I am
9 wrong -- but he is not signing off on saying his prediction
10 is as a voice for the Almighty that that's what the number
11 is. Just that if it's this and that, this is how the
12 concept works. I saw it as explaining the concept.
13 MR. DANIELS: My point is this: It's a matter for
14 the jury or a judge to decide in light of all of the other
15 cases.
16 THE COURT: I understand that's the point. Like I
17 say, when you talk about the strength of the case, that just
18 goes to whether or not the person is likely to be convicted
19 as opposed to affecting the base rate, whatever that might
20 be. I am not hearing an assertion from the witness what the
21 base rate, in fact, is. So all I am saying is I think you
22 are talking about two different things at this point.
23 MR. DANIELS: I believe what he is saying is that
24 you have to discount the accuracy rates of a polygraph by
25 the percentage of likelihood -- the number of people out of
218
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 100 who are actually guilty. That affects the base rate
2 because you have to multiply that. I think you get a
3 different multiplier depending on the kind of a case. It's
4 the kind of human judgment we make intuitively based on the
5 strength of the evidence.
6 THE WITNESS: I'm not saying what you said.
7 THE COURT: I'm not hearing him say that either.
8 THE WITNESS: The base rate has no impact on the
9 accuracy. The accuracy stays at 90 percent all the way
10 through. It's 90 percent accurate. That's the way it is.
11 It's that even if it's 90 percent accurate when the base
12 rate is that high, the cases that come through the Court
13 would be wrong 50 percent of the time on the hypothetical
14 example. That's what the example shows. If you decrease
15 the base rate the test produces fewer false negative tests.
16 Q. There's no way you are suggesting a jury could know
17 this hypothetical base rate?
18 A. No. I think that's why I said earlier I think it's the
19 kind of thing that the legal system or the Court should
20 decide whether this is important or not, and not something
21 that a jury should decide. So I think judges listening to
22 this argument should decide is this something they would
23 worry about or not. As a psychologist who gives tests, this
24 is something that we should worry about when the federal
25 government gives tests to employees, this is something they
219
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 should worry about. You can't be -- you can't ethically
2 practice clinical psychology, for instance, and give tests
3 under circumstances where the base rate deviates
4 substantially from 50 percent and use that test to make
5 important decisions about people, because more often than
6 not you will be wrong than rate if the base rate is 90
7 percent or 95 percent. You have an ethical obligation as a
8 psychologist to understand that and not use the test in
9 those circumstances.
10 Q. To make sure we are understanding each other, you apply
11 that same position to other kinds of diagnostic tests,
12 whether it's schizophrenia or the other things I suggested
13 here?
14 A. Right.
15 Q. Basing it on the base rate of the number of guilty
16 people who come before the Court, correct?
17 A. Yes.
18 MR. DANIELS: Now, I think my 15 minutes is up,
19 Your Honor. If we are going to finish today, I will skip
20 the surveys and bring those out through examination of our
21 witnesses.
22 THE COURT: Up to you, Mr. Daniels. If you have
23 more you want to do, my guess is we can figure out a way to
24 do it. If you think it will come through another witness
25 anyway, it's your call.
220
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 MR. DANIELS: Co-counsel suggests I not quit if I
2 truly have an option here.
3 THE COURT: I am giving you an option, I think.
4 What did we do in terms of plane reservations?
5 MR. COX: We sent somebody out to make a later
6 reservation. It's been done.
7 THE COURT: So it's changed to what?
8 MR. COX: 2:30, I believe. We have the morning.
9 MR. DANIELS: We should break then.
10 THE COURT: We could. Are you okay with breaking
11 now?
12 MR. COX: Yes.
13 THE COURT: Give you a chance to sleep. See you
14 tomorrow morning
15 (Note: Court stood in recess at 5:15)
16
17
18
19
20
21
22
23
24
25
221
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 STATE OF NEW MEXICO )
) ss
2 COUNTY OF BERNALILLO )
3 I, JAN BASSAREAR, official Court Reporter for the
4 Second Judicial District of the State of New Mexico, hereby
5 certify that I reported the foregoing proceedings to the
6 best of my ability; that the pages numbered 2 through 221,
7 inclusive, are a true and correct transcript of my
8 stenographic notes, and were reduced to typewritten
9 transcript through Computer-Aided Transcription; and that on
10 the date I reported these proceedings, I was a New Mexico
11 Certified Court Reporter.
12 Dated at Albuquerque, New Mexico, this 25th day of
13 June, 2003.
14
15
16
17
JAN BASSAREAR
18 New Mexico CCR No. 194
Expires: December 31, 2003
19
20
21 The total cost of this
22 Transcript is $___________
23
24
25
222
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
1 INDEX TO TRANSCRIPT OF PROCEEDINGS
2 PAGE
3 THE WITNESS: WILLIAM IACONA
4 Direct Examination by Mr. Cox 8
Cross-Examination by Mr. Daniels 132
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
223
JAN BASSAREAR, CRR-CCR-RPR
OFFICIAL COURT REPORTER
