by Jim Wygant
The question remains, why would the C.I.A. misrepresent the results of its own polygraph tests – if, in fact, it did? Why would it claim that the examinations revealed nothing, while a review by the F.B.I. determined that deception was evident?
The C.I.A. may have regarded polygraph as a convenient scapegoat. The polygraph instrument itself is a silent witness. It can not defend itself, and any examiner’s attempts to justify his conclusions only sound self-serving and defensive. An agency may be inclined to blame polygraph, rather than its own administrative decisions. With regard to Ames, just because the C.I.A. claims his polygraph results were wrong does not mean they were, particularly when members of another agency dispute that claim.
–––
We may never learn with certainty which version of Ames’ test results is correct, the one told by the C.I.A. or the alternative claimed by some members of the F.B.I. Lacking that, we can at least speculate about the kinds of errors that might have occurred while testing Ames – either errors within the test itself, as the C.I.A. would have us believe, or in human judgment that mistakenly refused to accept accurate test results, as F.B.I. agents claimed.
I have already suggested, in discussing the Harding case, some of the factors that might contribute to an error in a polygraph examination. We need to expand upon those a bit when talking about the tests administered to the employees of secret government agencies.
First, polygraph testing has never been 100 per cent accurate. Although research indicates accuracy of about 95 per cent with certain procedures, such a high degree of correct results only means that it is likely, but not certain, that results of any given test are accurate. In other words, mistakes will occur – regardless of how competent and experienced the examiner is, regardless of the issue, regardless of how appropriate the questions and the test procedures were.
Second, not all polygraph test procedures are the same. The kind of test that would have been administered to Tonya Harding to resolve a specific issue is not the same test administered by the C.I.A. on a periodic basis to discover spies. The spy test typically takes a shotgun approach, covering a broad range of general issues. The accuracy rate previously mentioned really only applies to the other kind of test, the one that addresses a single, specific issue. It is that kind of test that has itself been tested. Mock crimes with volunteers can approximate the specific issue test and reveal how accurate it is. It is nearly impossible to create a laboratory paradigm for the non-specific screening test that government agencies must rely upon.
We must also consider the probability that moles within secret agencies have been trained by their handlers in ways to beat polygraph tests. It is possible to deliberately beat a polygraph test, though never easy and usually obvious. To succeed, an examinee would need at minimum two advantages: prior knowledge of the test questions and an opportunity to practice with a real polygraph instrument. Controlling the physiological functions monitored in a polygraph examination is extremely difficult, even with practice. The average person is without the resources to satisfy those requirements, but a spy agency for a foreign government might have an easier time. Allowing for the possibility of that kind of training, a lying double agent would still find it difficult to escape detection, of both his lies and his attempts to subvert the test.
The final consideration, though not the least, is the examiner himself. He confronts his own natural bias when he tests a man who has a 20-year career behind him. The examiner knows that a negative finding can destroy that career. Most of us would want to be certain before we took such an irrevocable step. Regrettably, polygraph charts are not always clear enough to permit a high degree of certainty. Of any 100 C.I.A. employees an examiner tested, almost all of them would pass the test. Of those few who did not pass, most would make minor admissions and be retested. In that environment, the examiner may unconsciously begin to direct his efforts to getting everyone to pass the test, through their own admissions. When a major spy comes along, the examiner may be unprepared for something so removed from his usual experiences.
Even if polygraph failed to detect one major spy, should spy agencies cease using it? Those agencies would tell you that they use polygraph both as an investigative aid and as a deterrent. With each test administered, a potential spy may have been dissuaded from that course by the threat of discovery.
The Ames affair was settled quickly when he pleaded guilty to espionage and tax evasion and was sentenced to life imprisonment. Rosario pleaded guilty to lesser charges and received a much lighter sentence, presumably in consideration of Ames’ agreement to tell the CIA everything he had done and all that he could remember about his Russian handlers.
–––
I have heard of examiners who claim they have never made a mistake. The more cautious only assert that they have “no known errors.” The truly honest examiner who has administered at least a few hundred tests will say that he or she has probably made at least one error.
Like doctors, most polygraph examiners are reluctant to even talk about errors. However, a doctor who makes a mistake is apt to be confronted with its effects, evident in the condition of the patient, which provides speedy verification that something went wrong. Polygraph examiners do not often obtain verification of mistakes. If the error consists of reporting someone as truthful when he was really lying, the only verification would be a subsequent confession by the same person and enough additional evidence to corroborate his admission. That rarely happens. If it is the other kind of mistake, calling someone a liar when he is actually telling the truth, verification of the error would consist of someone else confessing with enough corroboration to make it credible. That, too, rarely happens.
Some polygraph examiners would like to promote the notion that they are infallible. They fear that any admission of error will destroy their credibility. Many years ago I knew a police examiner who had perfected his “delivery” of results. District Attorneys and police investigators loved him. He was always confident and absolute in saying that someone lied or told the truth. None of the law enforcement agencies that used him could remember ever getting an inconclusive result from him. I reviewed one of his tests early in my career and realized immediately, even as a novice examiner, that he did not know what he was doing. The questions were poorly worded. The charts reflected poor control of his instrument. And the result could only be reported as inconclusive, not the definite deception that he had claimed.
I cautioned a deputy district attorney that I knew well. His office relied heavily upon that examiner. The prosecutor listened but remained unconcerned. The examiner was delivering the results that the prosecutor’s office wanted to hear.
A few years later that examiner retired. He began doing tests for defense attorneys. He produced a series of improbable results, all claiming that the persons he tested were telling the truth. His former law enforcement colleagues began to insist that his tests be reviewed. When that happened the tests were regularly rejected. Other examiners could not find evidence in the charts for the results that the original examiner claimed. His reputation plummeted. Finally, his own former law enforcement agency decided to review a number of his old police cases, many of them murders, in which he had reported definite results. The re-evaluation produced startling results. Most of the tests were so poorly done that only inconclusive results were possible. One of the offices most affected by that discovery was that of the deputy district attorney I had warned. That office subsequently adopted a policy of refusing to consider polygraph results under any circumstance from any examiner. Another District Attorney’s office followed the more reasonable course of refusing to accept any results from that particular examiner.
What everyone concluded was that the examiner had been telling people what he perceived they wanted to hear. Often that meant that his decision was correct, but most of the time there was no basis in his work for any decision at all. Fortunately there was little fear that anyone had been wrongly convicted. The criminal tria
l process assures that the prosecutor must submit adequate evidence to obtain a conviction, regardless of polygraph results. What remained was concern that some avenues of investigation and prosecution may have been prematurely abandoned because of questionable test results.
The agency where that examiner had been formerly employed quickly instituted a program of quality control to regularly review the work of its other examiners.
This tale represents an extreme example of the infallibility syndrome. It was so important to the examiner that he appear infallible, he simply pretended to have obtained results from his tests. It was a remarkable failure within my profession, unlike anything else I have ever encountered. It happened more than thirty years ago and I can not imagine anything like that occurring today, when training is better and the widespread practice of chart review assures that incompetence becomes apparent quickly. There are, of course, doctors and lawyers who represent within their own professions that same level of incompetence, or worse. Such discoveries do not lead us to conclude that we should abandon all use of medicine or law.
I believe that the polygraph examiner serves himself, his profession, and his clients to the best extent possible if he never even implies that he is right-on certain from test results that someone is either lying or telling the truth. A little skepticism is prudent.
When I have analyzed polygraph charts and reached a conclusion, either indicating lying or truthfulness, I can make only one claim for my result. The person I tested responded in a manner characteristic of most people who are lying, or telling the truth. It is likely, based upon validity indicated by research, that when someone responds a certain way in a polygraph test he is lying, or – given different indications – telling the truth. It is likely. It is not certain. In fact, that is what my reports say, that “responses correspond to usual indication” of lying or truthfulness. I never say that someone lied or told the truth. How would I know, without some incredible psychic ability? I can offer my opinion, based upon my experience and the procedures I use, but I cannot offer irrefutable fact.
Judges fear that polygraph testimony will not be presented as opinion, but rather as fact. That concern probably represents the biggest impediment to admission of polygraph testimony, as documented repeatedly in court opinions. The courts do not want any witness or machine rendering absolute, unqualified verdicts on the truthfulness of someone standing trial, or someone who is a witness for or against that person. Those decisions about veracity are supposed to be left to the judge or jury. If judges recognized that any polygraph examiner is ultimately only expressing his or her opinion, the same as any other forensic expert, the courts might change their attitude toward admissibility.
Ironically, while the courts do not want irrefutable conclusions from polygraph examiners, law enforcement does. Police, prosecutors, and defense attorneys do not want to hear qualified opinions. They want hard, uncompromised facts in order to resolve their own difficult decisions. However, when they convince polygraph examiners to provide that kind of result, they feed a myth and contribute to one of polygraph’s most self-destructive aspects.
Those examiners who are most subject to stress are those who are themselves convinced that they must always avoid inconclusive results and must achieve a level of personal certainty that may not always be permitted by the circumstances of their work.
In my career I have made no secret of the fact that I have made mistakes. In nearly 20 years of work those errors I know about or even suspect are few, but they have always demanded concern and careful review. I have wanted to know how and why I failed, so I could attempt to avoid the same problems in the future. I can say that my review of those cases never revealed any evident mistake in either my test questions or my chart analysis. And when I have asked other examiners to review those same tests, my colleagues too were unable to identify any source of trouble. Still, I recognized that something had gone wrong, even though standard procedures were followed. The dilemma is not relieved by the knowledge that most other persons accused of similar issues and asked similar questions produced accurate results.
Ultimately I am forced to acknowledge that guilt, and fear, and deceit, and threat of discovery are things we do not understand well. Research tells us only that accuracy of 95 per cent can be achieved in polygraph testing. It does not tell us why the other 5 per cent were wrong.
16. Betrayers Of Trust
When I finish a polygraph test and have to tell someone that the results indicate lying, the usual response is resignation. No argument, no denials, no anger towards me. The attitude is that of someone who has been defeated and is only worrying about what lies ahead. I recall a notable exception to that, a time when the client’s reaction to her test results surprised both me and her employer. The case was a theft of money at a car dealership.
The general manager told me on the phone that over $1,000 in cash was missing from the safe. This was a classic case for polygraph, a single theft in a substantial amount, a very black and white issue. The suspects were those who had access to the safe. They included the general manager himself, the sales manager for new cars and his counterpart in used cars, the parts manager, and two women who worked in the vicinity of the safe and who each functioned as a combination of receptionist, secretary, and bookkeeper.
The loss had been discovered by one of the women, Lori Benedict, when she opened the safe in the morning to begin the day’s activities. She immediately notified the general manager. An entire bank bag and its contents were gone. Left behind in the safe were paperwork from car sales and a bag containing a much smaller amount of money.
I took my polygraph to the dealership and set up in an office where I would not be disturbed. The general manager insisted that he be tested first. We did that, and his test produced clear indications of truthfulness. Each test took about an hour and a half, so I spent the entire day working through part of the list of suspects. I returned the following day to finish up. When I was done, no one had produced definite results indicating lying. Only two tests had been problematic, producing results that did not match normal patterns.
One of those was the test of the parts manager. He was unusually nervous and uncommunicative during the pretest interview. His voice was so quiet that I had to strain to hear what he said, and he spoke only in response to my direct questions. He volunteered nothing. I suspected before the test that he might be the thief. However, his test results indicated truthfulness about the safe.
There was one test question that asked if he was afraid I would inquire about anything else. It was a question intended to discover something that was distracting the client from the issue. If the client answered “yes” I would reassure him that there would be no surprise questions. When he was confident enough to answer “no” we could begin the test.
I no longer use that question because it often seemed to invite concerns that did not previously exist, but in this one test it produced the greatest reaction that I have ever seen with that kind of question. Clearly the parts manager was concerned that I might ask about something else, something other than the theft of money from the safe. What I saw on those charts correlated well with the employee’s guarded attitude.
I asked him about that question afterwards and he became even more wary. I never found out exactly what caused that reaction, but knowing a little about the common abuses in the auto parts business I cautioned the general manager that he ought to check his parts inventory. The general manager was not surprised. He said he already had concerns in that regard.
The only other test of interest among all of the suspects was the one instance of an inconclusive result. An inconclusive result is really no result at all. It means that the examiner cannot reach a determination that there are patterns on the charts corresponding either to lying or truthfulness. In a typical test the same questions are read to the client three or four times. In order to reach any determination, the examiner must find that the person’s body responded in a similar manner ea
ch time the questions were asked, even though the sequence was changed. In other words, there must be a consistent pattern. No one-time response to any particular question will by itself determine test results. When results are inconclusive, a retest is usually scheduled if the client agrees and the charts do not suggest that the test was compromised by deliberate countermeasures. The inconclusive result at the car dealership was from the test of Lori Benedict, the young woman who had reported the loss. She consented to my request that she come to my office for a retest.
When I first met Lori Benedict I saw her as her employer did, a polite, attractive woman in her mid-twenties who had worked at the dealership for about three years, practically a career in the automobile trade where high turnover is typical. I was willing to attribute the inconclusive result to her likely concern about being the person who had discovered the loss, and whatever suspicion she might think had accrued to her because of that. Those circumstances did not actually cause me to be suspicious of her, though I knew that many commercial thefts are in fact committed by the very person who reports them. In this case, so many people knew the combination to the safe and it was unattended so frequently during the hours the business was open that it meant little to me that Ms. Benedict had been the one who reported the loss.
When she arrived at my office I spent some time chatting with her before we began her retest. I did not expect to learn anything new. I was only hoping to provide an opportunity for her to become accustomed to my office and to burn off some of her expected nervous anticipation. What she spontaneously told me was surprising. She began complaining about her employer.