Robert J. Brown v. John F. Darcy, A/K/A Jack Darcy, an Individual Ducommun, Inc., a Corporation , 783 F.2d 1389 ( 1986 )

CYNTHIA HOLCOMB HALL, Circuit Judge:

Polygraph evidence has been disfavored in the federal courts since the decision of the District of Columbia Circuit in Frye v. United States, 293 Fed. 1013 (D.C. Cir.1923).¹ Despite advances in polygraphy since the Frye decision, there are still significant questions regarding the reliability of polygraph examinations. Polygraph evidence also presents problems for our adversarial system because it has an overwhelmingly prejudicial effect when it is inaccurate, interferes with the jury’s authority to determine credibility, and imposes a burden on district courts to review the reliability of polygraph evidence in each case. Accordingly, we hold that polygraph evidence may not be admitted to establish the truth of statements made during the polygraph examination unless the parties have stipulated to the admissibility of the polygraph results before the examination is administered, and the court is satisfied that the examination has been administered in a manner which supports the reliability of the polygraph results.

BACKGROUND

Plaintiff-appellant Robert J. Brown appeals from a jury verdict in favor of defendants John F. Darcy and Ducommun, Inc. on his complaint for libel and slander. At the time this dispute arose Brown was a Corporate Vice President and Director of Sales and Marketing-West for Motorola, Inc. Brown was also one of seven members of Motorola’s Distributor Council, the body responsible for Motorola franchising decisions. Darcy was President of Kierulff Electronics, a wholly-owned subsidiary of Ducommun. Kierulff Electronics was a distributor of electronic components, including Motorola parts, and held Motorola franchises in eight cities throughout the United States. Kierulff Electronics was interested in obtaining Motorola franchises in eight additional locations. Motorola franchises are valuable to distributors of electronic components because franchisees are permitted to buy Motorola parts at a discount, and receive sales assistance from Motorola field personnel.

This dispute revolves around a May 15, 1980 meeting at the Radisson Hotel in Scottsdale, Arizona, among Brown, Darcy, and Donald Fullam, a Motorola executive with responsibilities similar to Brown’s and also a member of Motorola’s Distributor Council. According to Darcy, Brown and Fullam solicited a bribe from him at the meeting, asking him for $200,000 in return for Brown’s and Fullam’s efforts to lobby the Distributor Council to approve the additional franchises Kierulff Electronics was seeking. Brown maintains that the meeting was a legitimate discussion of Kierulff Electronics’ record in existing franchises and the possibility of obtaining further franchises. According to Brown, Darcy at one point inquired whether Brown and Fullam were soliciting a bribe in exchange for their help in getting franchises, but Brown claims that he did not take Darcy seriously because such action would be a clear violation of Motorola policy. Brown remembers Fullam responding to Darcy’s inquiry by describing a bribe that had been paid to another electronics company, but Brown claims that he directed the conversation back to Kierulff Electronics’ eligibility for new franchises based on its past performance.

Darcy returned to Los Angeles after the meeting, and consulted with Ducommun’s President, Wallace Booth, and Ducommun’s legal counsel, Arthur Schmultz. Acting on their advice, Darcy decided to wait and see what would happen rather than contact Motorola about the alleged solicitation.

In July 1980 Darcy was contacted by three different people in the electronics business who informed Darcy that Fullam was claiming that Darcy owed him a substantial sum of money. One of these individuals put Darcy in contact with Alfred Stein, Assistant General Manager of the Semiconductor Group at Motorola. Darcy recounted his version of the Radisson Hotel meeting for Stein, and later for Stein and Richard Weise, Motorola's General Counsel. During this second meeting Darcy provided a written statement and volunteered to take a polygraph test to support his accusations. The written statement summarized Darcy’s version of the Radisson Hotel meeting, including his claim that Brown and Fullam had solicited a payment in exchange for lobbying efforts on behalf of Kierulff Electronics. Darcy’s written statement also indicated that “[o]ver the course of the evening, [Darcy] consumed three glasses of 7-UP and Brown and Fullam accounted for the balance of a $33.00 bar bill,” and that while “walking from the cocktail lounge to the parking lot [Brown and Fullam] repeatedly made insulting remarks with respect to [Darcy’s] stupidity and lack of understanding as to what it took to be successful in business.” At trial Darcy testified that he had orally informed Stein and Weise that Fullam attributed Brown’s involvement in the solicitation scheme to Brown’s need to pay off substantial gambling debts.

. The polygraph examination² which Darcy agreed to take in support of his accusations against Brown was administered by Cy Gilson³ on September 9, 1980. The control questions employed by Gilson were criticized by some of the other polygraph experts.⁴ Gugas found that the control questions were inappropriate because they were not sufficiently related to the subject which was being investigated and concluded that Darcy’s truthfulness could not be evaluated because of this defect.⁵ The control questions were also disapproved of by Horvath because they were prefaced by the phrase “Did you honestly believe ... ?” This type of control question is ineffective because it tests what the examinee believed, not what actually happened.⁶ In spite of these problems Gilson concluded that Darcy was telling the truth about the Radisson Hotel meeting.

On September 10, 1980, investigators from Motorola informed Brown of the accusations by Darcy and told Brown that Darcy had passed a polygraph examination. Brown submitted a written statement contradicting Darcy’s accusations within two days, and agreed to submit to a polygraph test. The test was scheduled to be administered by Cy Gilson. Brown initially objected because Gilson had already opined that Darcy was telling the truth, but eventually consented to the examination by Gil-son on September 12, 1980. Gilson-found the Brown test results inconclusive. The Brown test was criticized by Horvath and Gugas because it was unlikely that Gilson could establish the rapport with Brown necessary to a successful polygraph examination.⁷ Gugas also stated that he would not have administered the polygraph examination to Brown on September 12 if he had known that Brown was suffering from emphysema and had experienced two relatively sleepless nights since being informed of Darcy’s accusations, because respiratory problems and high stress levels will affect the accuracy of a polygraph examination.⁸ At trial Horvath testified that no opinion could validly be drawn regarding Brown’s truthfulness or deceptiveness because the theft control questions were not relevant to the subject under investigation, too narrowly drawn, and may not have elicited the necessary untruthful response.

After Brown filed his motion to exclude the polygraph evidence, defendants arranged for a second examination of Darcy by Clarence Kirkland. Defendants did not inform the court or Brown of the examination until they knew that Kirkland had concluded that Darcy was telling the truth. The reliability of the Kirkland examination is questionable because it was not administered until December 1982, more than thirty months after the Radisson Hotel meeting. The lapse of time between the events in question and the polygraph examination increases the likelihood that the subject will convince himself that certain statements are true, and therefore not produce the physiological responses associated with lying.⁹

Before trial Brown moved to exclude all of the polygraph examinations. Alvin Bur-dick, the court-appointed expert,¹⁰ concluded that the Darcy examination administered by Gilson was reliable, and agreed with Gilson’s conclusion that Darcy was telling the truth. Burdick also concluded that Brown was attempting to beat the polygraph examination which Gilson found inconclusive, and supported Kirkland’s conclusion that Darcy was telling the truth during his second examination. After reviewing the conflicting expert opinions and hearing argument from counsel the district judge concluded:

It appears to me that, as far as the law is concerned, I have no question regarding reliability. Anything further, the factual nature of it, can be argued to the jury, much as you have argued it here to me today.

The jury returned a verdict for defendants, and the district judge denied Brown’s motion for new trial, indicating that he agreed with the jury’s verdict and that, in his opinion, the polygraph evidence “was almost irrelevant in terms of reaching the decision that was reached.” Brown filed a timely appeal.

ANALYSIS

On appeal Brown argues that the district court erred in admitting the polygraph evidence, and that the court should have instructed the jury that Darcy’s statements to Motorola regarding the bar bill, Brown’s alleged gambling debts, and Brown’s allegedly intimidating behavior could be considered as independent libels or slanders and therefore independent grounds for recovery.

I. POLYGRAPH EVIDENCE

We have consistently expressed an inhospitable view towards the admission of unstipulated polygraph evidence under the Federal Rules of Evidence. See, e.g., United States v. Givens, 767 F.2d 574, 585-86 (9th Cir.), cert. denied, — U.S.-, 106 S.Ct. 321, 88 L.Ed.2d 304 (1985); Demma, 523 F.2d at 987.¹¹ We have found no United States Court of Appeals decision which has affirmed the admission of unstipulated polygraph evidence under the Federal Rules of .Evidence, see Dowd v. Calabrese, 585 F.Supp. 430, 431 (D.D.C.1984); see generally Annot., 43 A.L.R.Fed. 68 (1979), or concluded that the refusal to admit polygraph evidence at trial was an abuse of discretion, see United States v. Williams, 737 F.2d 594, 611 (7th Cir.1984), cert. denied, — U.S. -, 105 S.Ct. 1354, 84 L.Ed.2d 377 (1985). At least five other Courts of Appeals have held that unstipulated polygraph evidence is per se inadmissible. Poole v. Perini, 659 F.2d 730, 735 (6th Cir.1981), cert. denied, 455 U.S. 910, 102 S.Ct. 1259, 71 L.Ed.2d 450 (1982); United States v. Alexander, 526 F.2d 161, 163-70 (8th Cir.1975); United States v. Cochran, 499 F.2d 380, 393 (5th Cir.1974), cert. denied, 419 U.S. 1124, 95 S.Ct. 810, 42 L.Ed.2d 825 (1975); United States v. Skeens, 494 F.2d 1050, 1053 (D.C.Cir.1974); Marks v. United States, 260 F.2d 377, 382 (10th Cir.1958), cert. denied, 358 U.S. 929, 79 S.Ct. 315, 3 L.Ed.2d 302 (1959). Given the questionable reliability of polygraph evidence and the great potential for prejudice from inaccurate polygraph evidence, we conclude that unstipulated polygraph evidence is inadmissible as technical or scientific evidence under Fed.R.Evid. 702 because it does not “assist the trier of fact to understand the evidence or to determine a fact in issue.”

The questionable accuracy of polygraph examinations is the most persuasive reason for excluding polygraph evidence. Estimates on the reliability with which polygraph examinations predict whether the ex-aminee is telling the truth range from seventy percent to ninety-five percent.¹² These reliability estimates are further complicated by the number of factors which affect the reliability of polygraph examinations in any particular case.

The reliability of polygraph evidence is widely debated. United States v. Marshall, 526 F.2d 1349, 1360 (9th Cir.1975), cert. denied, 426 U.S. 923, 96 S.Ct. 2631, 49 L.Ed.2d 376 (1976). Multiple variables may influence the results of a polygraph test, including the motivation of the subject, his physical and mental condition, the competence, integrity, and attitude of the operator, the wording of the relevant questions, the appropriateness of the control questions, and the interpretation of the resulting graph.

Givens, 767 F.2d at 585. See also Alexander, 526 F.2d at 165-66. In this case, for example, experts Horvath and Gugas doubted the validity of the control questions asked during the Darcy examination administered by Gilson, and the propriety of having Gilson examine Brown after he had rendered an opinion that Darcy was telling the truth.

The overwhelming potential for prejudice when incorrect polygraph evidence is introduced to a jury also supports our conclusion that polygraph evidence is per se inadmissible. As the Eighth Circuit has noted:

When polygraph evidence is offered in evidence at trial, it is likely to be shrouded with an aura of near infallibility, akin to the ancient oracle of Delphi. During the course of laying the evidentiary foundation at trial, the polygraphist will present his own assessment of the test’s reliability which will generally be well in excess of 90 percent____ Based upon the presentment of this particular form of scientific evidence, present-day jurors, despite their sophistication and increased educational levels and intellectual capacities, are still likely to give significant, if not conclusive, weight to a polygraphist’s opinion as to whether the defendant is being truthful or deceitful in his response to a question bearing on a dispositive issue____

Alexander, 526 F.2d at 168. See also Falsia, 724 F.2d at 1342 (noting polygraph’s “misleading appearance of accuracy”); Marshall, 526 F.2d at 1360. Even if the accuracy of polygraph examinations approaches the eighty percent or ninety percent claimed by the polygraph experts, this view of the polygraph as an absolute indicator of truth creates an overwhelming potential for prejudice when inaccurate results are introduced.¹³ See Marshall, 526 F.2d at 1360 (potential prejudice is one aspect district court may consider in refusing to admit polygraph evidence).

The introduction of polygraph evidence also infringes on the jury’s role in determining credibility. Our adversary system is built on the premise that the jury reviews the testimony and determines which version of events it believes. Allowing a polygraph expert to analyze responses to a series of questions and then testify that one side is telling the truth interferes with this function. See Alexander, 526 F.2d at 168; see also Dowd, 585 F.Supp. at 434. Polygraph evidence is different from other scientific evidence such as ballistics, fingerprints, or voice analysis, because it is an opinion regarding the ultimate issue before the jury, not just one issue in dispute. Alexander, 526 F.2d at 169. Providing the jury with an all or nothing evaluation of credibility and then telling the jury that this evaluation has an eighty percent to ninety percent chance of being accurate if the polygraph was properly administered interferes with, rather than enhances, the deliberative process.

We also note the conservation of district court resources which will be achieved by creating a per se rule limiting the admission of unstipulated polygraph evidence. See Falsia, 724 F.2d at 1341 (time consumption involved in evaluating reliability of polygraph evidence is proper consideration for court in refusing to admit such evidence); Marshall, 526 F.2d at 1360 (same). In this case the district court reviewed the opinions of five polygraph experts and heard oral argument before ruling that the polygraph evidence was admissible. The jury then heard roughly two full days of testimony from polygraph experts in an eight day trial. Detailed case by case inquiries into the reliability of polygraph evidence such as this will not be necessary under the bright line rule that we adopt'today.

Darcy argues that the polygraph examinations of Brown and Darcy conducted by Gilson are operative facts in this dispute because they were administered before this action was filed, making the examinations admissible in this case even if polygraph evidence is generally inadmissible under the Federal Rules of Evidence. We disagree. When a polygraph examination is an operative fact evidence of the examination is admissible. See, e.g., Thome v. City of El Segundo, 726 F.2d 459, 469-71 & n. 11 (9th Cir.1983) (admitting polygraph questions into evidence in action by employee against employer and polygraph examiner for sexual discrimination in firing and in administration of the polygraph), cert. denied, — U.S.-, 105 S.Ct. 380, 83 L.Ed.2d 315 (1984); Smiddy v. Varney, 665 F.2d 261, 265 (9th Cir.1981) (polygraph evidence admitted when polygraph examination of defendant was cause of unlawful arrest of plaintiff), cert. denied, 459 U.S. 829, 103 S.Ct. 65, 74 L.Ed.2d 66 (1982). A polygraph examination, however, is not an operative fact simply because it is administered before the cause of action is filed. The key inquiry is for what purpose the polygraph is being introduced. If the polygraph evidence is being introduced because it is relevant that a polygraph was administered regardless of the results, or because the polygraph examination is the basis of the cause of action as in Thome or Smiddy, then the polygraph evidence may be admissible as an operative fact.¹⁴ If, on the other hand, the polygraph evidence is offered to establish that one party’s version of the events is the truth, the polygraph evidence is being introduced for its substantive value and is inadmissible absent a stipulation between the parties prior to administration of the polygraph.

The polygraph evidence is not admissible in this case as an operative fact. Darcy offered the polygraph evidence because it supports his version of the Radisson Hotel meeting, and ultimately his claim that his statements to Motorola are not libelous or slanderous because they are true.

Darcy also argues that, even if the introduction of the polygraph evidence was in error, any error was harmless. An error is harmless if “the jury’s verdict is more probably than not untainted by the error.” Haddad v. Lockheed California Corp., 720 F.2d 1454, 1459 (9th Cir.1983). In this case the polygraph evidence consumed approximately one fourth of the entire trial, and all of the experts testified that the polygraph is roughly ninety percent accurate when administered correctly. Under these circumstances introduction of the polygraph evidence was not harmless error.

II. LIBEL AND SLANDER CLAIMS

The district court declined to instruct the jury that Darcy’s statements regarding the bar bill, Brown’s abusive behavior, and Brown’s gambling debts could be considered as independent grounds for recovery. Whether a statement is reasonably susceptible of defamatory meaning is a question of state law, Forsher v. Bugliosi, 26 Cal.3d 792, 803, 163 Cal.Rptr. 628, 608 P.2d 716 (1980); Selleck v. Globe International, 166 Cal.App.3d 1123, 1132, 212 Cal.Rptr. 838 (1985), which we review de novo. In re McLihn, 739 F.2d 1395, 1400 (9th Cir.1984).

False statements which injure a person in his occupation can support a claim for libel or slander under California law. Cal. Civ.Code §§ 45, 46. “It is error for a court to rule that a publication cannot be defamatory on its face when by any reasonable interpretation the language is susceptible of a defamatory meaning.” Selleck, 166 Cal.App.3d at 1131, 212 Cal.Rptr. 838 (citing Cameron v. Wernick, 251 Cal.App.2d 890, 60 Cal.Rptr. 102 (1967)). In determining whether a statement is capable of defamatory meaning we place ourselves in the position of the reader or hearer, considering the statement as a whole and the context in which the statement is made. Selleck, 166 Cal.App.3d at 1131, 212 Cal.Rptr. 838 (citing MacLeod v. Tribune Publishing Co., 52 Cal.2d 536, 546-47, 343 P.2d 36 (1959)).

Applying these rules we find that the district court erred in refusing to instruct the jury that Darcy’s written statements regarding the bar bill, Brown’s intimidating behavior, and Darcy’s oral statement regarding Brown’s gambling debts could be considered as a basis of recovery. Although these statements are less damaging to Brown than Darcy’s claim that Brown solicited a bribe, these statements could rationally be interpreted as injurious to Brown’s business reputation in the context in which they were made.¹⁵

REVERSED and REMANDED.

1 . The Frye court held:

We think the systolic blood pressure deception test [a predecessor of the modern polygraph] has not yet gained such standing and scientific recognition among physiological and psychological authorities as would justify the courts in admitting expert testimony deduced from the discovery, development, and experiments thus far made.

293 Fed. at 1014.

2 . Modern polygraph instruments measure blood pressure, pulse, respiration, and galvanic skin reflex. See Reid & Inbau, Truth and Deception, The Polygraph (“Lie-Detector'') Technique 5-6 (2d ed. 1977); Tarlow, Admissibility of Polygraph Evidence in 1975: An Aid in Determining Credibility In a Perjury-Plagued System, 26 Hastings L.J. 917, 921-22 (1975). Polygraphy is based on the theory that the stress of lying creates physiological responses which can be detected by monitoring these four factors. Tarlow, supra, at 921-22; Skolnick, Scientific Theory and Scientific Evidence: An Analysis of Lie-Detection, 70 Yale LJ. 694, 699-703 (1961). In a zone of comparison polygraph test, such as those used in this case, the subject is asked a "control question" designed to elicit an untruthful response. Reid & Inbau, supra at 28-30; Tarlow, supra at 921-22 & n. 35. This lie will give the examiner feedback which will be used for comparison when the subject is asked questions regarding the subject under investigation, known as “relevant questions." If the subject is telling the truth in response to the relevant questions, the physiological response should be less significant than the response measured when the control question is answered. Reid & Inbau, supra at 29.

3 . Gilson, a polygraph examiner for the Arizona Department of Public Safety, testified that he has conducted approximately 6,000 polygraph examinations since he was trained at the Bachelor School of Lie Detection in 1971.

4 . Brown presented documentary evidence and trial testimony by Chris Gugas, a polygraph examiner for thirty years and past President of the American Polygraph Examiners Association, and Frank S. Horvath, an associate professor of criminology at Michigan State University licensed in polygraphy since 1965 and also currently serving as research director at John E. Reid and Associates. Darcy presented expert testimony through Clarence Kirkland, who was trained in polygraphy at Texas A & M in 1978 and has served as an instructor at the Los Angeles Institute of Polygraphy, and Cy Gilson. The district court also appointed its own expert, Alvin Burdick, an examiner since 1971 and formerly employed as such by the Los Angeles Police Department and the United States Postal Service.

5 . Specifically Gugas stated:

The theft control questions actually asked are not close to the present matter, and these questions did not elicit the type of significant reaction needed from Mr. Darcy because it is possible that Mr. Darcy was in no position to steal anything from his employers. Regardless of these two theft control questions, they did not give enough response to score Darcy’s charts as truthful. The importance of a good control question cannot be overemphasized, since it is the heart of the scoring technique and visual observation by the examiner.

[T]he lack of sufficient valid control questions propounded to Mr. Darcy by Mr. Gilson renders highly suspect any conclusion expressed by Mr. Gilson as to either the truthfulness or falsity of Mr. Darcy’s responses to the relevant questions

6 . See Reid & Inbau, supra note 2, at 26 (relevant questions should be directed at factual information; they should not evoke a response based on the subject’s opinion).

7 . Horvath noted that one examiner may choose to examine more than one subject in an investigation, but explained that the examiner should not render an opinion on truthfulness until after both subjects have been examined. Mr. Gugas summarized the problem as follows:

In administering a polygraph examination, the examiner must take steps to make sure that the examinee has confidence in the examiner and in the process of the examination. Prior to the examination, the examiner works with the examinee to formulate the questions and to assure the examinee that no other questions will be asked. At all times, the examiner attempts to obtain and maintain the confidence of the examinee in him. If the examinee lacks confidence in the integrity and ability of the examiner, the results of an examination of that examinee are highly suspect. Where an examinee is told that the polygraph examiner has already formed the opinion that a person contradicting the examinee is being truthful, the examinee is likely to lose all confidence in the examiner, thus rendering any examination results to be [sic] unreliable.

8 . See Highleyman, The Deceptive Certainty of the "Lie Detector" 10 Hastings L.J. 47, 60 (1958) (respiratory disorders and extreme emotional tension can affect realiability of polygraph results).

9 . See Reid & Inbau, supra note 2, at 228. See also United States v. Demma, 523 F.2d 981, 987 (9th Cir.1975) (en banc) ("probative force of the [polygraph] evidence diminished by the lapse of time between the occurrence of the events and the taking of the test”).

10 . See supra note 4.

11 . We have reviewed district court decisions refusing to admit polygraph evidence on at least eighteen occasions and never found an abuse of discretion. Givens, 767 F.2d at 585-86; United States v. Johnson, 735 F.2d 1200, 1203 (9th Cir.1984); United States v. Falsia, 724 F.2d 1339, 1341-42 (9th Cir.1983); United States v. Ferris, 719 F.2d 1405, 1408 (9th Cir.1983); United States v. Eden, 659 F.2d 1376, 1382 (9th Cir.1981) (trial court did not err in refusing to compel presentation of exculpatory polygraph evidence before grand jury), cert. denied, 455 U.S. 949, 102 S.Ct. 1450, 71 L.Ed.2d 663 (1982); United States v. Glover, 596 F.2d 857, 867 (9th Cir.), cert. denied, 444 U.S. 860, 100 S.Ct. 124, 62 L.Ed.2d 81 (1979); United States v. McIntyre, 582 F.2d 1221, 1226 (9th Cir.1978); United States v. Radlick, 581 F.2d 225, 229 (9th Cir.1978); United States v. Benveniste, 564 F.2d 335, 338-39 (9th Cir.1977); United States v. Flores, 540 F.2d 432, 436-437 (9th Cir.1976); United States v. Marshall, 526 F.2d 1349, 1360 (9th Cir.), cert. denied, 426 U.S. 923, 96 S.Ct. 2631, 49 L.Ed.2d 376 (1976); Demma, 523 F.2d at 987; United States v. Watts, 502 F.2d 726, 728 (9th Cir.1974); United States v. Bagsby, 489 F.2d 725, 726-27 (9th Cir.1973); United States v. Alvarez, 472 F.2d 111, 113 (9th Cir.), cert. denied, 412 U.S. 921, 93 S.Ct. 2742, 37 L.Ed.2d 148 (1973); United States v. Jenkins, 470 F.2d 1061, 1064 (9th Cir.1972), cert. denied, 411 U.S. 920, 93 S.Ct. 1544, 36 L.Ed.2d 313 (1973); United States v. Salazar-Gaeta, 447 F.2d 468, 469 (9th Cir.1971); United States v. Sadrzadeh, 440 F.2d 389, 390 (9th Cir.), cert. denied, 404 U.S. 850, 92 S.Ct. 84, 30 L.Ed.2d 88 (1971). In Marshall we specifically concluded that “[w]ith the polygraph’s misleading reputation as a ‘truth teller,’ the widespread debate concerning its reliability, the critical requirement of a competent examiner and the judicial problems of self-incrimination and hearsay, a trial court will rarely abuse its discretion by refusing to admit the evidence, even for a limited purpose and under limited conditions.” 526 F.2d at 1360.

In the only decision in which we have affirmed the admission of polygraph evidence offered to establish the truth of the statements made during the examination the parties stipulated to admissibility before the polygraph was administered. Herman v. Eagle Star Ins. Co., 396 F.2d 427 (9th Cir.1968). On two other occasions we noted the presence of polygraph evidence in the record when the issue of admissibility was not before us, United States v. Estrada-Lucas, 651 F.2d 1261, 1263-65 (9th Cir.1980); United States v. Murray, 492 F.2d 178, 189 (9th Cir.1973), cert. denied, 419 U.S. 942, 95 S.Ct. 210, 42 L.Ed.2d 166 (1974), and on one occasion we indicated that admission of polygraph evidence on remand would not necessarily be an abuse of discretion, Benveniste, 564 F.2d at 339 n. 3.

12 . See Inbau & Reid, Lie Detection and Criminal Interrogation 111-12 (3d ed. 1953) (estimating 95% accuracy in 4,280 criminal investigations over five years); Holmes, The Degree of Objectivity in Chart Interpretation, reprinted in II Academy Lectures in Lie Detection, 62, 62-70 (1958) (75% accuracy estimate); Hunter & Ash, The Accuracy and Consistency of Polygraph Examiners’ Diagnosis, 1 J. Pol.Sci. & Admin. 370, 372 (1973) (86% accuracy achieved by seven examiners in controlled study); Sternbach, Gustafson, & Colier, Don't Trust the Lie Detector, 127, 130 Harv.Bus.Rev., (Nov.-Dee. 1962) (70% accuracy estimate). Accuracy is lower for individual questions than for overall conclusions regarding truthfulness. See Horvath & Reid, The Reliability of Polygraph Examiner Diagnosis of Truth and Deception, 62 J.Crim.L., C., & P.S. 276, 278-80 (1971) (overall accuracy 88% in controlled study by ten examiners; per question accuracy of 79%); Hunter & Ash, supra, at 372 (overall accuracy 86%, per question accuracy of 82%). Inexperienced examiners are less accurate in their evaluations than experienced examiners, see Horvath & Reid, supra, at 279 (examiners with less than six months experience were 12% less accurate in overall predictions of truth than experienced examiners in controlled study), and the same examiner may reach different conclusions regarding truthfulness when reviewing the same examination results on two different occasions, see Hunter & Ash, supra, at 372 (examiners achieved only 85% consistency in overall determinations of truth and deception and 80% consistency in rating individual questions with lapse of at least three months between interpretations).

These accuracy figures may be inflated. The Inbau & Reid study claiming 95% accuracy has been criticized because the determinations of accuracy was based on a comparison between polygraph results and trial outcomes or confessions. Sternbach, Gustafson & Colier, supra, at 130. Such an approach assumes the accuracy of polygraph determinations which coincide with trial outcomes. Only forty-eight percent of the cases in the Inbau & Reid study were actually verified by confession or other corroborative evidence. Id. The controlled studies inflate accuracy figures because they protect against errors in the formulation of questions and physical problems of the examinee among other variables. Even Reid & Inbau, the primary proponents of polygraphy in the United States, admit these difficulties:

In many case investigations involving the use of the Polygraph technique, the truth or falsity of the examiner's findings is never factually established by subsequent events or disclosures. Proof is often lacking, therefore, as to whether the examiner in any given case was right or wrong. Nor can the accuracy of the Polygraph technique be determined in a psychology laboratory setting or by the use of fictitious crimes under other testing circumstances. This limitation prevails for the simple reason that it is practically impossible to simulate conditions comparable to those involved in actual case situations.

Reid & Inbau, supra note 2, at 304.

13 . Our conclusion that unstipulated polygraph evidence is per se inadmissible is the same

whether we examine the usefulness of the polygraph evidence to the jury under Fed.R.Evid. 702, or undertake the more traditional relevance and prejudice analysis of Fed.R.Evid. 401 and Fed.R.Evid. 403. The questionable reliability of polygraph evidence undermines its relevance, and the potential prejudice, time consumption and confusion caused by the introduction of polygraph evidence outweigh its probative value. See Falsía, 724 F.2d at 1342.

14 . Even when polygraph evidence is offered for these limited purposes the trial court must determine that the probative value of the polygraph evidence outweighs the potential prejudice and time consumption involved in presenting such evidence. See supra note 13.

15 . Darcy argues that the district court did not err in refusing to submit the above issues to the jury because they were not included in the first amended complaint or the pretrial conference order. See Fed.R.Civ.P. 16(e). The first amended complaint and the pretrial conference order address Darcy’s written statement and Darcy’s oral statements regarding Brown’s allegedly intimidating behavior. Although the complaint does not refer to Darcy's oral statements regarding Brown’s alleged gambling debts, the pretrial conference order does list the falsity of Darcy’s oral representations as a fact at issue, and includes the question of whether Motorola officers became concerned that Brown’s gambling debts would interfere with his job performance. On remand the district court may wish to allow amendment of the complaint and the pretrial conference order to coincide with the evidence presented at the first trial. See Fed.R.Civ.P. 15. See also Foman v. Davis, 371 U.S. 178, 182, 83 S.Ct. 227, 230, 9 L.Ed.2d 222 (1982) (leave to amend should be freely given).