concurring in part and dissenting in part, in which RONEY, Chief Judge, HILL and CLARK, Circuit Judges, join:
I concur with the Court’s holding that polygraph evidence should be admissible in this Circuit when both parties stipulate m advance to the circumstances of the test and to the scope of its admissibility, subject to the understanding that such stipulations may be accepted or rejected by the trial judge at his discretion.1 I dissent, however, from the Court’s finding that the polygraph has gained acceptance in the scientific community as a reliable instrument for detecting lies, and from the Court’s holding that polygraph evidence is admissible under Fed.R.Evid. 608.
I. POLYGRAPH THEORY
A. Introduction
The Court’s reasoning begins with the proposition that polygraph technology has reached the point where its accuracy is generally accepted by the scientific community. In fact, the scientific community remains sharply divided on the reliability of the polygraph. U.S. Congress, Office of Technology Assessment, Scientific Validity of Polygraph Testing: A Research Review and Evaluation — A Technical Memorandum 43 (1983) [hereinafter OTA Memorandum]. Many theorists question the basic assumptions underlying the polygraph: that telling lies is stressful, and that this stress manifests itself in physiological responses which can be recorded on a polygraph. See Ney, Expressing Emotions and Controlling Feelings, in The Polygraph Test: Lies, Truth and Science 65 (A. Gale ed.1988) [hereinafter The Polygraph Test]; Employee Polygraph Protection Act: Hearing on HR. 208 Before the Education and Labor Comm., 100th Cong., 1st Sess. 51 (1987) (testimony of John F. Beary, III, M.D. on behalf of the American Medical Association) [hereinafter H.R. Hearing]. Moreover, Congress has sharply limited use of the polygraph in the private sector. Employee Polygraph Protection Act of 1988, P.L. 100-347, 102 Stat. *1538646 (codified at 29 U.S.C.A. § 2001 (West Supp.1989)).2
The polygraph device records the subject’s physiological activities (e.g., heart rate, blood pressure, respiration, and perspiration) as he is questioned by a polygraph examiner. Bull, What is the Lie Detection Test? in The Polygraph Test 11-12. There are two major types of polygraph examinations: the “control question test” and the “concealed information test.” The control question test is used most frequently in investigating specific incidents. The examiner compares the data corresponding to (a) questions relevant to the crime (b) “control” questions designed to upset the subject but not directly relevant to the crime, and (c) neutral questions. If the subject reacts more strongly to the relevant questions than to the control and neutral questions, then the examiner infers that the subject is lying. Id. at 13-17. There is much debate about the accuracy of control question tests in specific-incident investigations. Raskin, Does Science Support Polygraph Testing, in The Polygraph Test 98-99.
The concealed information test focuses on the fact that only the person involved in the crime could know the answers to certain questions. The examiner presents a series of multiple choice questions concerning the crime while the polygraph machine records the subject’s physiological activities. If the subject has relatively strong physiological reactions to the correct alternatives, then the examiner infers that the subject is attempting to conceal information about the crime. Id. at 102. The concealed information test assumes that information about the crime is protected, but in fact police often inform all suspects and even the media about the crime. Id.
B. The Polygraph Is Based On Questionable Assumptions
Lie detection is based on four assumptions: (1) that individuals cannot control their physiologies and behavior, (2) that specific emotions can be triggered by specific stimuli, (3) that there are specific relationships between the different aspects of behavior (such as what people say, how they behave, and how they respond physiologically), and (4) that there are no differences among people, so that most people will respond similarly.
The assumption that individuals cannot control their physiologies is subject to serious debate. Some theorists argue that individuals can learn to control their physiological responses and that by producing physiological responses at opportune times during the polygraph test these people could portray themselves as truthful when they are not. Ney, Expressing Emotions and Controlling Feelings at 67 (“Jet-fighter pilots learn to control their emotions (and therefore their physiology) in order to operate with maximum efficiency under extreme physical and psychological stress.”) These techniques for fooling the polygraph are called countermeasures. Gudjonsson, How to Defeat the Polygraph Tests in The Polygraph Test 126. Little research has been done on the effectiveness of countermeasures in reducing detection of lies, but the results of research that has been done, while conflicting, indicate that countermeasures can be effective. OTA Memorandum at 100-01; Gudjonsson, How to Defeat the Polygraph Tests at 135 (concluding that use of physical countermeasures (e.g., pressing toes to floor) is effective when the subject has been trained in countermeasures).3
*1539Another assumption underlying the polygraph is that specific emotions will be triggered by the act of lying. Some theorists, however, do not believe that emotions are automatically triggered by the presence of such specific stimuli. These theorists see a more indirect causal chain between stimuli and emotion: a person is presented with stimuli, then appraises it, and only then reacts with an emotion, which is based on the person's cognitive appraisal of the stimuli.4 According to this theory, people can adjust their thinking to “reappraise” the stressful stimuli and create a different emotional reaction than one might expect. Ney, Expressing Emotions and Controlling Feelings 68 (“tell the truth and think of something painful and the truth may appear on the polygraph as a lie”).5 Of course, there would be no way for an examiner to determine how the subject is appraising the stimuli in his mind.
The third assumption underlying the polygraph is that there are set patterns of physiological responses that reflect dishonesty: changed blood pressure, heart rate, respiration, and perspiration. There is controversy over this proposition in the scientific community. Id. at 70; HR. Hearing at 51 (statement of John F. Beary, III, M.D.) (“there is no Pinocchio response. If you lie your nose does not grow a half inch longer or some other unique bodily response.”)
The fourth assumption underlying the lie detector is that people can be expected to respond to similar stimuli in similar ways. Some researchers maintain, however, that individuals do not respond to stress similarly and that no one index can be used to measure emotions in different individuals. Ney, Expressing Emotions and Controlling Feelings at 71-72; Gudjonsson, How to Defeat the Polygraph Tests 135.
C. Appellant’s Statistics Are Misleading
Piccinonna claims that “the relevant scientific community”6 estimates the accura*1540cy of the polygraph to be in the upper-eighty to mid-ninety percent range. Appellant’s En Banc Brief at 9. This figure is misleading and subject to serious dispute. The polygraph must do two things: correctly identify liars and correctly identify those who are telling the truth.7 Employee Polygraph Protection Act: Hearing on S. 185 Before the Senate Committee on Labor and Human Resources, 100th Cong., 1st Sess. (Appendix to statement of John F. Beary, III, M.D.) (1988) [Hereinafter “S. Hearing No single figure, therefore, can fully express the accuracy of the polygraph. The Office of Technology Assessment compiled the results of six pri- or reviews of polygraph research, ten field studies, and fourteen analog studies that the Office of Technology Assessment determined met minimum scientific standards. All of the studies used the control question technique in specific-incident criminal investigation settings. The results were as follows:
Six prior reviews of field studies:
-average accuracy ranged from 64 to 98 percent.
Ten individual field studies:
-correct guilty detections ranged from 70.6 to 98.6 percent and averaged 86.3 percent;
-correct innocent detections ranged from 12.5 to 94.1 percent and averaged 76 percent;
-false positive rate (innocent persons found deceptive) ranged from 0 to 75 percent and averaged 19.1 percent; and
-false negative rate (guilty persons found nondeceptive) ranged from 0 to 29.4 percent and averaged 10.2 percent.
Fourteen individual analog studies:
-correct guilty detections ranged from 35.4 to 100 percent and averaged 63.7 percent;
-correct innocent detections ranged from 32 to 91 percent and averaged 57.9 percent;
-false positives ranged from 2 to 50.7 percent and averaged 14.1 percent; and
-false negatives ranged from 0 to 28.7 percent and averaged 10.4 percent.
OTA Memorandum at 97. Note that because the question “Is the subject lying?” is a yes or no question, a random method of answering the question (e.g., a coin toss) would be correct 50% of the time. The Memorandum concluded,
The wide variability of results from both prior research reviews and [The Office of Technology Assessment’s] own review of individual studies makes it impossible to determine a specific overall quantitative measure of polygraph validity. The preponderance of research evidence does indicate that, when the control question technique is used in specific-incident criminal investigation, the polygraph detects deception at a rate better than chance, but with error rates that could be considered significant.
Id.
D. Extrinsic Factors Affect Accuracy
A number of extrinsic factors affect polygraph validity. Most important, because the examiner must formulate the questions, supplement the data with his own impression of the subject during the exam, and infer lies from a combination of the data and his impressions, the level of skill and training of the examiner will affect the reliability of the results. S.Rep. No. 284, 100th Cong., 2d Sess. 42, reprinted in 1988 U.S.Code Cong. & Admin. News 726, 729 [hereinafter Senate Report]; Barland, The Polygraph in the USA and Elsewhere in The Polygraph Test 82. Unfortunately, there are no uniform standards for the training of polygraph examiners in this country. Senate Report at 43, U.S.Code Cong. & Admin. News at 731; S. Hearing at 27 (statement of Mr. William J. Scheve, Jr., American Polygraph Association); see Barland, The *1541Polygraph in the USA and Elsewhere at 75 (the American Polygraph Association has accredited over 30 polygraph schools with courses ranging from seven to fourteen weeks).
A quality control system that reviews the examiners’ conclusions also affects the validity of polygraph results. The results of most federally administered polygraph exams are checked by quality control officers, who call for reexaminations if the data does not indicate that the examiner’s conclusion was correct. Barland, The Polygraph in the USA and Elsewhere 87. Few police examiners work within such a system, and almost no private examiners have quality control. Id. at 82.
The length of a polygraph exam will also affect the validity of the results. One advocate of the polygraph has stated that an expert polygraph exam would take a minimum of several hours to complete. Senate Report at 43, 1988 U.S.Code Cong, and Admin. News at 730-31.
II. POLYGRAPH TESTS SHOULD BE EXCLUDED UNDER THE FEDERAL RULES OF EVIDENCE
Under Federal Rule of Evidence 702, expert testimony is proper if the testimony would assist the trier of fact in analyzing the evidence. Fed.R.Evid. 702 advisory committee’s note (West 1989). Because the polygraph can predict whether a person is lying with accuracy that is only slightly greater than chance, it will be of little help to the trier of fact. Moreover, this slight helpfulness must be weighed against the dangers of unfair prejudice, confusion of the issues and waste of time. Fed.R.Evid. 403. The Ninth Circuit has found that polygraph evidence has an overwhelming potential for prejudicing the jury. Brown v. Darcy, 783 F.2d 1389, 1396 (9th Cir.1986) (citing United States v. Alexander, 526 F.2d 161, 168 (8th Cir.1975)); see also Gianelli, The Admissibility of Novel Scientific Evidence: Frye v. United States, a Half-Century Later, 80 Colum.L.Rev. 1197, 1237 (1980) (“The major danger of scientific evidence is its potential to mislead the jury; an aura of scientific infallibility may shroud the evidence and thus lead the jury to accept it without critical scrutiny.”) The Brown court determined that unstipulated polygraph evidence is inadmissible under both Rule 702 and Rule 403. Brown, 783 F.2d at 1396 n. 13. The polygraph presents itself as being very scientific. For instance, it is said to measure “galvanic skin response,” Appellant’s En Banc Brief at 10, which merely means that it measures how much a person perspires. Bull, What is the Lie-Detection Test? at 11. This scientific aura tends to cloud the fact that the machine’s accuracy at detecting lies is little better than chance. Brown, 783 F.2d at 1396 (quoting Alexander, 526 F.2d at 168); OTA Memorandum at 97.
The Ninth Circuit also found that admission of polygraph evidence had the potential of confusing the issues and wasting time. Id. at 1397; see Fed.R.Evid. 403. In the Brown case, for instance, the polygraph evidence consumed one fourth of the entire trial. Brown, 783 F.2d at 1397 (two full days of an eight-day trial). Because polygraph evidence is of little help to the trier of fact, and has great potential for prejudicing the trier of fact, confusing the issues and wasting time, it should be excluded under Federal Rule of Evidence 403.
The danger of prejudice, confusion of the issues and wasting time should also prevent courts from admitting polygraph evidence under Rule 608 for purposes of impeaching a witness. As the Court’s opinion correctly states, all offers of polygraph evidence should be analyzed in light of Rule 403. Cf. United States v. Miller, 874 F.2d 1255, 1261 (9th Cir.1989) (even when offered for a limited purpose, polygraph evidence must go through a Rule 403 analysis). To hold that polygraph evidence is admissible under Rule 608 would create too large an exception to the rule barring polygraph evidence generally, and polygraph test results would wind up being admitted into evidence in most cases. Moreover, there is nothing special about the Rule 608 impeachment procedure that lessens the dangers of prejudice and confusion of the issues. Cf. United States v. Toney, 615 F.2d 277 (5th Cir.1980) (“Rule 403 is. a general rule, ‘designed as a guide for the handling of situations for which no specific rules have been formulated.’ ”)
*1542III. CONCLUSION
The scientific community remains sharply divided over the issue of the validity of polygraph exams. Although presented as a rigorously “scientific” procedure, the polygraph test in fact relies upon a highly subjective, inexact correlation of physiological factors having only a debatable relationship to dishonesty as such. The device detects lies at a rate only somewhat better than chance. Polygraph evidence, therefore, should not be admissible under Rule 702 or under Rule 608 to impeach a witness.
In this case, the government did not stipulate to the admissibility of the defendant’s polygraph evidence and did not participate in selection of the examiner or the determination of the circumstances of the test. I would therefore AFFIRM the judgment below.
. If the parties wish to alter the applicability of Rules 403 and 702 in their case, they should be able to do so by advance stipulation, as long as they do not interfere with any third party’s interests or the adjudicatory role of the courts. See Wigmore on Evidence § 7a (P. Tillers rev. 1983). But see id. at 602 n. 35 (courts generally hold polygraph results inadmissible even where there is a stipulation). Because such a stipulation would alter the applicability of rules of evidence, however, the trial judge has the discretion to reject the parties’ proposed stipulation. The trial judge has broad discretion on questions of the admissibility of evidence and should not be reversed unless there is a clear abuse of discretion. United States v. Borders, 693 F.2d 1318, 1324 (11th Cir.1982); Scheib v. Williams-McWilliams Co., 628 F.2d 509, 511 (5th Cir.1980).
. The Employee Polygraph Protection Act prohibits the use of polygraphs in pre-employment screening and sharply curtails the permissible uses of the polygraph in specific-incident investigations. 29 U.S.C.A. §§ 2002, 2006 to 2007 (West Supp.1989).
. In order to fool the control question test, the subject must enhance his physiological reactions to neutral questions, and/or decrease his physiological reactions to relevant questions. Inducing physical pain or muscle tension during non-relevant questions can reduce the difference between physiological responses to relevant and neutral questions. One study found that pressing one’s toes against the floor during neutral questions reduced the detection of lies from 75% to 10%. Gudjonsson, How to Defeat the Polygraph tests, at 129 (citing Kubis, Studies in Lie Detection: Computer Feasibility Considerations (Technical Report 62-205, prepared for Air Force Systems Command) (1962)). A competing study concluded that such countermeasures caused no reduction in detection of lies. Id. (citing More, Polygraph Research and the University, 14 Law and Order 73-78 (1966)). *1539The Office of Technology Assessment reviewed the available research on this issue in 1983 and concluded that counter-measures can be effective and that further research in the area is necessary to prevent persons engaged in illicit activities from creating "false negatives” on polygraph exams and, in this way, clearing themselves of any suspicion. OTA Memorandum at 100-01 (‘The possible effects of countermeasures are particularly significant to the extent that the polygraph is used and relied on for national security purposes_ [T]hose individuals who the Federal Government would most want to detect (e.g., for national security violations) may well be the most motivated and perhaps the best trained to avoid detection.”)
. This is Lazarus’s cognitive appraisal theory of emotion. See Ney, Expressing Emotions and Controlling Feelings at 68 (citing Lazarus, Coyne, and Folkman, Cognition, Emotion and Motivation: The Doctoring of Humpty-Dumpty, in Approaches to Emotion (K. Scherer and P. Ekman eds.1984)).
. Even when the subject is not employing countermeasures, cognitive appraisal seems to affect the results of tests where the subject is accused of a nebulous crime or where the sole issue is criminal intent. In these cases, the issue is not as distinct as in cases where the subject is accused of a physical act. The issue calls for an interpretation, which may be subject to distortion or rationalization in the defendant’s mind. Barland, The Polygraph Test in the U.S.A. and Elsewhere in The Polygraph Test 83-84. In the instant case, the defendant is accused of knowingly telling a falsehood when he denied knowledge of an agreement among south Florida garbage companies. The defendant could have rationalized his answers to questions on such ambiguous issues, and avoided an emotional and a physiological response to the questions.
.Piccinonna claims that the "relevant scientific community” is "those who have done research on the techniques and/or have had training or experience in the techniques [of polygraph testing]_” Appellant's En Banc Brief at 9. The Office of Technology Assessment has stated, however, that "Basic polygraph research should consider the latest research from the fields of psychology, physiology, psychiatry, neuroscience, and medicine” in order to develop a stronger theoretical base for the polygraph. OTA Memorandum at 6. It is reasonable to argue, therefore, that experts from these fields are competent to comment on the validity of polygraph testing. Gianelli, The Admissibility of Novel Scientific Evidence: Frye v. United States, a Half-Century Later, 80 Colum.L.Rev. 1197, 1210 (1980) ("'The purpose of the Frye test is defeated by an approach which allows a court to ignore the informed opinions of a substantial segment of the scientific community which stands in opposition to the process in question.’ ” (quoting Reed v. State, 283 Md. 374, 399, 391 A.2d 364, 377 (1978))). Congress has recognized that the community of experts competent to testify on the polygraph reaches beyond polygraph examiners and their proponents. For example, Dr. John F. Beary III appeared on behalf of the American Medical Association before the House Education and Labor Committee and the *1540Senate Committee on Labor and Human Resources to oppose the use of polygraphs in the workplace. H.R. Hearing at 51; S. Hearing at 16.
. For example, a polygraph examiner who accused every subject of lying would be 100% accurate at detecting liars. His accuracy at detecting those who are truthful, however, would be unacceptably low.