cedure which would produce the answers which I think come from any reasonably probing examination. I don't think the questions I have asked are unreasonable or unfair.

I believe that if we are to use these devices and operators rely upon them to interpret human responses, then we should have the scientific notebooks full of the data that will fully support their reliability.

Mr. INBAU. I suggest to you that they are full to the extent that they can be filled for this reason.

Mr. Moss. No; to the extent that you have done it, not to the extent that they can be. You could follow each and every case. You could follow each and every case from the day of examination to the day of ultimate death or resignation. That could be done. If it were to be scientifically undertaken, you would follow it perhaps more extensively than you have. You follow it now within the framework of the economics of the activities in which you are engaged. But it is being treated as a science. I have serious doubt that it is a science.

Mr. INBAU. The limitation that I was referring to is this. Here is some money stolen. There are six people that have likely had access to it. You test them all and report that they are all innocent. Now, nothing ever comes to light disclosing who it was that stole that. There is ability in practical limitation here. You cannot take the conscience of these people apart on an operating table and find out that one of the six you said were innocent actually was guilty. It seems to me you just have a limitation.

Yet, I don't know that you can overcome it in any other way. You can't subject this technique to laboratory experiments. You can't set up hypothetical crime situations. That is too far removed from actual criminal cases.

Mr. Moss. You are not only using this in criminal cases, you remember.

Mr. INBAU. Yes; personnel.

Mr. Moss. A great number of instances of use in Government relates to personnel screening. So, we do conduct extensive experiments in human behavior and the human psyche, don't we?

Mr. INBAU. Yes.

Mr. Moss. We spend millions of dollars every year attempting to learn why people behave as they do. In this field of personnel screening, I think it is extremely important, if we are going to employ a technique that is very reliable. We are supposed to protect everyone to the best of our ability, and a system that might catch 94-percent accuracy, we will say, but condemn the other 6 percent to some disadvantage because of its failures is not at all in accord with our principles, is it?

Mr. INBAU. I don't know that I followed your statement carefully enough to give it the answer. But again I say, and this is about the only answer I know we can give, that here we have a certain type of test procedure, a certain type of testing, that can only be checked upon to the extent that there are disclosures which ultimately develop that either prove or disprove the accuracy of the diagnosis that has been made. It seems to me that if you have 30,000 cases and you have these developments establishing nothing more than 1 percent of error in that diagnosis, that you have a foundation there upon which to build the conclusion that it is a reliable aid.

Mr. Moss. I thought Mr. Reuss developed the fact yesterday that we didn't have the 99-percent accuracy. Actually, your cards could be more complete. We have discussed two things here today. I discussed the matter of cigarette smoking. Your cards don't reflect this. I don't know if you have assurances from men well versed in the field of human physiology that there is no important reaction here. I don't know whether you have gone to the trouble of determining from the best psychiatric sources that perhaps there might not be some possibility of an impact of an emotional nature of this if a person was subjected to a lengthy test and deprived of his cane. But your records don't reflect whether he is a heavy user or not. We talked about the immediate environment, the hot room with high humidity and the cool room with relatively low humidity. This is not reflected in your cards. These are factors that are unevaluated.

Mr. INBAU. I think there is one factor that is being overlooked here, that you are recording not only involuntary physiological reactions but very voluntary ones, such as in respiration. If you give a man a test, you have the control questions-let us say this happens-you ask him the control question and you get this suppression in respiration characteristic of a lie reaction.

Mr. Moss. Is it characteristic of any other reaction?

Mr. INBAU. The suppression of respiration?

Mr. Moss. Yes.

Mr. INBAU. It could be if you depended on that one reaction alone. But here you have it on the control question and you don't have it on the cruel question.

Mr. Moss. What other reactions could it be typical of?

Mr. INBAU. The individual suppression at that particular time?
Mr. Moss. Yes.

Mr. INBAU. It could be an accidental occurrence.

Mr. Moss. Could it be symptomatic of some other physiological disturbance?

Mr. INBAU. If it is, you are going to have something of that nature throughout the record or a repetition of it throughout the record. You are not going to have it isolated in these particular questions and have it duplicated on a series of tests.

In other words, on three tests, you will get this suppression in respiration at the control question. You don't get it at

Mr. Moss. Are you saying that it is impossible for it to occur on a series of questions?

Mr. INBAU. I wouldn't say anything is impossible here, but it has been our experience that it just doesn't work out that way. I don't know anything else we can rely on.

Mr. Moss. You have taken it on the basis of a response to a question without relating it to the overall physiological function of the person before you.

Mr. LINDBERG. I might respond that the same problem that you are suggesting, of course, would require considerable more definition than you are allowing us at this point and considerable more explanation. But the random occurrences that you are suggesting are also the problems which the other sciences, as I think you would agree they are sciences, cardiology, neurology, and these other medical sciences, are confronted with that same element of probability. They are

dealing with everyday life and death matters. I might also suggest or ask of you, Mr. Chairman, if you would give us some direction in answering your question as to what science you would consider adequate, what medical practice is adequate for us to more or less parallel in our recording of observational physiological data. I presume you have one in mind.

Mr. Moss. No, I haven't. I am going to attempt before we complete the hearings on this subject to seek, however, opinions from medical specialists and from specialists in the field of psychiatry, and if possible, arrange an appropriate panel where this matter can be discussed. I know I go in for electrocardiograms at least twice a year and the operator who gives them to me is a very competent operator. But I would not want to rely on her interpretation of them, her diagnosis of my then current condition.

I look to my doctor who has had many years and is alert to all of the things that could bring about those little peaks and valleys in that reading.

Normally, he is not going to make a judgment just on that. He is going to have many other tests taken. Then his diagnosis is probably going to be made with considerable humbleness on his part, recognizing the large latitude existing for error. But here we are also dealing with what could be economic life or death for an individual.

Mr. LINDBERG. Or life and death possibly.

Mr. Moss. Or life and death. Yet, we haven't the same highly qualified diagnosis. You are using the same basic material. You are having the body report on its functions, and you are diagnosing them but the diagnosis is made in most instances by a person with less training by far than that nurse who gives me my test.

Mr. LINDBERG. I would suggest that it would be very illuminating in order to place the polygraph field in its proper perspective as a science to subject these psychiatrists that you are going to call and the medical people to the same type of cross-examination that you have engaged in here for the purpose of comparing the two.

Mr. Moss. We would. Most certainly, we would.

Mr. LINDBERG. I would be interested to see if there is a significant difference. Because our people at the Northwestern Biomedical Engineering Staff have been unable to find any immediate corrections in our procedures that would bring more of what you desire into our field.

We are constantly working on these problems. They are not answered. One of the limitations is obviously money. We are private individuals engaged in private enterprise. I would hope that one of the results of this committee meeting might be some real good research in these interdisciplinary sciences.

Mr. Moss. We would hope it would lead to improvement, too. Again, I have not with intent, Mr. Reuss, taken the floor from you. You have the floor and I think you should be permitted to continue your questions.

Mr. REUSS. If I could just go back to the galvanic skin detector which Mr. Inbau and I think your colleagues have said is worthless or practically worthless as a sole test. I still don't quite understand how this thing is supposed to work. Who figured out that a given pat

tern of responses stigmatizes the liar as opposed to the truth-teller, or are you all prepared to say that it is absolutely worthless?

Mr. LINDBERG. Are you talking about the galvanic skin reflex?
Mr. REUSS. Yes.

Mr. LINDBERG. There are examinations or tests that can be given which in a very substantial number of instances will detect certain types of deception. I might suggest one, where the subject answers no to a series of questions as to whether or not he chose a particular card or a number. When he is instructed to answer no to all of the cards, even when the number that he chose is called, frequently, and even in our laboratory conditions, you will find a rise similar to the one that was seen on this chart here. But, then, when you immediately proceed into the specific issue framework, we have not found in the technique that we use, that there is sufficient consistency for the reliability needed for a scientific assessment, as we conceive of it.

Mr. BACKSTER. Sir, is it possible to comment on that?

Mr. REUSS. Let me proceed, and I do have a question of you, Mr. Inbau. Getting back to this question yesterday of the accuracy of the assertion in the 1953 work by Mr. Inbau and Mr. John Reid, which was the subject of considerable questioning by me yesterday, doubt about the validity of those statistics was also expressed by three ple who wrote an article in the November, December 1962 issue of the Harvard Business Review. (See exhibit 2E, p. 147.)

peo

Mr. Backster, you wrote an article in the 1963 issue of Law and Order magazine, did you not, which had something to say about that Harvard Business Review article?

Mr. BACKSTER. Yes, I did.

Mr. Moss. Do you have any more questions of Mr. Reid or Mr. Lindberg?

Mr. REUSS. I have one of Mr. Lindberg. Going back to the business of the two-way mirror which was rather infrequently used in your five polygraph chambers, do you ever use that to give prospective clients or police officers or company officials or governmental officers who might be contemplating the employment of your agency as a lie detector unit-do you ever use that double mirror to give them an opportunity to observe polygraph tests in action?

Mr. LINDBERG. Yes, so long as they have no interest in the individual case itself.

Mr. REUSS. When they do observe this, this is done without the knowledge of the subject that they are being observed by these people? Mr. LINDBERG. Yes.

Mr. REUSS. Let me conclude, then, before Mr. Lindberg and Mr. Reid go well, thank you very much.

Mr. Moss. I want to thank you gentlemen for appearing here and I regret that we have had so many delays. I hope you catch your train.

Mr. LINDBERG. Thank you very much, sir.

Mr. JOHN REID. Thank you, sir.

Mr. REUSS. Now, Mr. Backster, I was going to read to you from your February 1963 article in Law and Order magazine. Do you have a copy of that in front of you? (See exhibit 2F, p. 154.)

Mr. BACKSTER. Yes, I do.

Mr. REUSS. I will just read the relevant sentences and you check me so that I don't misread any of them. The title of your article goes, "Don't Trust the Lie Detector Versus Do Trust the Lie Detector." Then you say:

The first of the above titles "Don't Trust the Lie Detector" is actually that of a feature article which appeared in the November-December issue of the Harvard Business Review. This shocking injustice to the truly scientific brand of lie detection was authored by Richard A. Sternbach, Lawrence A. Gustafson, and Ronald L. Colier. An insert paragraph states that two of the authors are "professional psychologists who have been doing extensive research with polygraph techniques, the third is a graduate student at Massachusetts Institute of Technology who has studied commercial practice."

One thing you can be certain of is that none of them are authorities, or even well-known practitioners, in the field of scientific lie detection. For these individuals to take it upon themselves to be so critical of a field they know so little about is like this author writing a critical article on the Rorschach inkblot test or the Bernreuter personality inventory, both of these being important tests in the field of psychology.

Despite our justification to challenge these individual's right to present such a totally destructive criticism of a specialty they mistakenly may feel within their realm-the sad part of it is that we, within the polygraph field, have provided them with the ammunition to fire at us.

Then, let us go to page 2, second full paragraph:

In their subsection entitled "Accuracy and Error" in the Harvard Business Review article, the authors severely criticize the vague, incomplete, and erroneous published polygraph statistics as well as the methods by which they were calculated. We have now been called to task for inadequate statistics released during past years-and what defense do we have? I know of none, and furthermore, if we attempt to defend such statistics of the past we are playing the psychologist's game with his rules and on his home grounds.

The statistics which the Harvard Business Review article authors were criticizing were the statistics of Mr. Reid and Mr. Inbau, are they not?

Mr. BACKSTER. I am not sure if they localize themselves to them but I can check.

Mr. REUSS. Will you take a look at the article and having refreshed your recollection, answer my question?

Mr. BACKSTER. Yes, sir; they quite closely stick to the book “Lie Detection and Criminal Interrogation."

Mr. REUSS. Is it your opinion from the 1963 work I was describing which asserts that among the some 4,200 polygraph cases listed there was only 1-percent inaccuracy-that those statistics are vague, incomplete, and erroneous; is that correct?

Mr. BACKSTER. Sir, I was told this by a statistician whom I consulted to see if there was foundation for the criticism on their part. I know nothing of statistics. I am not an expert in that. I was told that there were some grounds to criticize.

Mr. REUSS. And on the basis of your conference with that statistician you referred to the 1953 assertion and the statistics back of it as vague, incomplete, and erroneous, did you not?

Mr. BACKSTER. Yes, sir; I did.

Mr. REUSS. Thank you; I have no further questions.

Mr. Moss. Mr. Reid.

Mr. OGDEN REID. I have just one or two questions of Mr. Backster. What are the facts, Mr. Backster with regard to preemployment screening in the industrial field? What is the percentage of individ

31-647-64——8