Statev.Porter

Supreme Court of ConnecticutMay 20, 1997
241 Conn. 57 (Conn. 1997)
241 Conn. 57698 A.2d 739

(SC 15363)

Syllabus

Convicted of the crime of arson in the first degree in connection with the destruction of his home by a fire of suspicious origin, the defendant appealed to the Appellate Court, which affirmed the judgment of the trial court, and the defendant, on the granting of certification, appealed to this court asserting that this state should reconsider the traditional rule that polygraph evidence is per se inadmissible at trial. Held: 1. The standard for admissibility of scientific evidence as set forth by the United States Supreme Court in Daubert v. Merrill Dow Pharmaceuticals, Inc. ( 509 U.S. 579) should govern the admissibility of scientific evidence in Connecticut; under that standard, before proffered scientific evidence may be admitted, the trial court must determine that it is reliable, i.e., that the underlying theory or methodology is scientifically valid, and that the evidence is relevant, i.e., that the underlying reasoning or methodology properly can be applied to the facts in the case. 2. The defendant could not prevail on his claim that Connecticut should abandon its traditional per se rule against the admission of polygraph evidence at trial; the prejudicial impact of that evidence being greatly in excess of its probative value, polygraph evidence should remain per se inadmissible in all trial court proceedings in which the rules of evidence apply and for all trial purposes in Connecticut courts. 3. The defendant was not deprived of his federal constitutional right to present a defense by the trial court's exclusion of his favorable polygraph results without at least having conducted an evidentiary hearing regarding the reliability of his polygraph examination; there is a very strong basis for excluding polygraph evidence, such an exclusion is not arbitrary in light of the demonstrated unreliability of polygraph tests, and an evidentiary hearing on the offer of proof would have been a "nugatory undertaking."

( One justice concurring and dissenting, one justice dissenting)

Argued September 24, 1996

Officially released May 20, 1997

Procedural History

Substitute information charging the defendant with two counts of the crime of arson in the first degree, brought to the Superior Court in the judicial district of New London, where the court, Parker, J., denied the defendant's motion to admit into evidence the results of his polygraph examination; thereafter, the matter was tried to the jury before Parker, J.; verdict and judgment of guilty of one count of arson in the first degree; subsequently, the court denied the defendant's motion for a new trial, and the defendant appealed to the Appellate Court, O'Connell, Landau and Hennessy, Js., which affirmed the judgment of the trial court, and the defendant, on the granting of certification, appealed to this court. Affirmed.

M. Hatcher Norris, with whom, on the brief, were Joel M. Ellis, Peter W. Soulsby and Lee A. Gold, legal intern, for the appellant (defendant).

Kevin T. Kane, state's attorney, with whom, were Lisa Herskowitz, deputy assistant state's attorney, and, on the brief, William R. Jay, certified legal intern, for the appellee (state).


Opinion


BORDEN, J.

The issues in this certified appeal are: (1) whether Connecticut should adopt as the standard for the admissibility of scientific evidence the standard set forth by the United States Supreme Court in Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579, 113 S.Ct. 2786, 125 L.Ed.2d 469 (1993); and (2) whether Connecticut should abandon its traditional per se rule that polygraph evidence is inadmissible at trial. The defendant, Christian E. Porter, appeals from the judgment of the Appellate Court affirming his conviction for arson in the first degree in violation of General Statutes § 53a-111 (a) (4). The defendant claims that: (1) the Appellate Court incorrectly concluded that the trial court properly denied his request for an evidentiary hearing regarding the admissibility of polygraph evidence; and (2) in light of the United States Supreme Court's decision in Daubert v. Merrell Dow Pharmaceuticals, Inc., supra, 579, this court should reconsider its test for determining the admissibility of scientific evidence, which is currently based on Frye v. United States, 293 F. 1013 (D.C. Cir. 1923), and should conclude that polygraph evidence is admissible under the Daubert test. We conclude that Daubert provides the proper threshold standard for the admissibility of scientific evidence in Connecticut. We also conclude, however, on the basis of our own independent examination of the extensive literature and case law regarding polygraph evidence, that polygraph evidence should remain per se inadmissible in Connecticut trials, and consequently that an evidentiary hearing was not necessary to evaluate the reliability of such evidence. Accordingly, we affirm the judgment of the Appellate Court.

The following facts and procedural history are undisputed. The defendant's home in Norwich was destroyed by a fire on July 20, 1992. The defendant was subsequently charged with two counts of arson in the first degree in violation of § 53a-111(a)(3) and (4).

General Statutes § 53a-111 provides in relevant part: "Arson in the first degree: Class A felony. (a) A person is guilty of arson in the first degree when, with intent to destroy or damage a building, as defined in section 53a-100, he starts a fire or causes an explosion, and . . . (3) such fire or explosion was caused for the purpose of collecting insurance proceeds for the resultant loss; or (4) at the scene of such fire or explosion a peace officer or firefighter is subjected to a substantial risk of bodily injury. . . ."

Before trial, the defendant retained Leighton Hammond, a polygrapher, to conduct a polygraph examination to determine whether the defendant was telling the truth when he claimed that he had no guilty knowledge of, and had not participated in, the burning of his home. The defendant did not give the state advance notification of the examination. The pertinent test questions asked of the defendant were: (1) "Did you set fire to your home?"; (2) "Did you tell even one lie, in your statement to the Norwich Police?"; and (3) "Do you know for sure, if any person deliberately set fire to your home?" In the opinion of Hammond, the defendant was telling the truth when he answered "no" to each of these questions.

The defendant then moved that the trial court admit the results of the polygraph examination. After a hearing, the trial court denied the defendant's motion, stating that it was not the place of a trial court to reconsider Connecticut's traditional per se ban on the admissibility of polygraph evidence.

Following a jury trial, at which the defendant did not testify, he was convicted of arson in the first degree in violation of § 53a-111 (a)(4). The defendant appealed from the judgment of conviction to the Appellate Court claiming, inter alia, that the trial court improperly had refused to admit the favorable results of his polygraph examination and further improperly had refused to allow him to make an evidentiary offer of proof on the polygraph results. State v. Porter, 39 Conn. App. 800, 801, 668 A.2d 725 (1995). The Appellate Court affirmed the trial court's judgment, concluding that "[t]he trial court, like this court, is bound by the Connecticut precedent which bars the admission of polygraph results. . . . Because an evidentiary hearing would have been a nugatory undertaking, the trial court was not required to grant the defendant's motion for an evidentiary offer of proof." (Citation omitted.) Id., 803. This certified appeal followed.

The dissent spends several pages recounting, and belittling, the state's evidence in this case, implicitly arguing that the supposed weakness of the state's case is relevant to the issue of the admissibility of the polygraph evidence. We fail to see, however, how the strength of the state's case, in and of itself, is relevant to the general legal question of the admissibility of polygraph evidence. Compare part II B 3 a and footnote 63 of this opinion. We are aware of no support for the proposition that the legal question of the admissibility of a particular type of evidence turns on the general strength of the opposing party's case.
In any event, we disagree with the dissent's characterization of the state's case. Indeed, the Appellate Court specifically rejected a sufficiency of the evidence claim raised by the defendant. State v. Porter, 39 Conn. App. 800, 803-805, 668 A.2d 725 (1995). The state presented testimony, which the jury was entitled to believe, that the damage to the defendant's house was caused by three independent and unconnected fires; that flammable liquid accelerants were present in all three locations; and that electricity did not cause any of the fires. Moreover, although not necessary in an arson prosecution, the state presented evidence of the defendant's financial motive for setting the fire. Specifically, the jury reasonably could have found that the defendant and his wife had falsified the amount and source of their income when applying for the mortgage loan to finance the purchase of the house, that the defendant's wife was unemployed and had stopped collecting unemployment benefits three months before the fire, and that the defendant and his wife had exhausted all of their savings by the time of the fire. From these facts, the jury reasonably could have concluded that the defendant's motive in setting the fire was not to collect the insurance proceeds himself, but to free himself from his obligations under the mortgage.

We granted the defendant's petition for certification to appeal from the Appellate Court, limited to the following issues: "Under the circumstances of this case: 1. Did the Appellate Court properly conclude that the trial court was correct in denying the defendant's request for an evidentiary hearing regarding the admissibility of the defendant's polygraph evidence?
"2. Should this court reconsider the applicability of the test for determining the admissibility of scientific evidence set forth in Frye v. United States, 293 F. 1013 (D.C. Cir. 1923), in light of the United States Supreme Court's decision in Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579, 113 S.Ct. 2786, 125 L.Ed.2d 469 (1993)?" State v. Porter, 236 Conn. 908, 670 A.2d 1308 (1996).

I The Daubert Standard

The defendant argues that Connecticut should adopt the federal test for the admissibility of scientific evidence, as set forth by the United States Supreme Court in Daubert v. Merrell Dow Pharmaceuticals, Inc., supra, 509 U.S. 579. We agree that, when read and applied correctly, Daubert provides the proper approach to the threshold admissibility of scientific evidence.

Indeed, the state agrees that " Daubert's focus on evidentiary reliability . . . is a positive step away from the rigid application of Frye as a precondition of admissibility."

A

We begin with a general examination of the Frye and Daubert standards for the admission of scientific evidence. The standard enunciated in Frye v. United States, supra, 293 F. 1013, was predominant in both state and federal courts for the seventy years from its formulation until the decision in Daubert in 1993. Frye itself was a lie detector case; indeed, it was the first appellate case in the United States to address the admissibility of lie detector examination results. In Frye, the defendant appealed from his murder conviction on the grounds that the trial court had improperly disallowed expert testimony that he had passed a "systolic blood pressure deception test," the precursor of the modern polygraph examination. See J. McCall, "Misconceptions and Reevaluation — Polygraph Admissibility After Rock and Daubert," 1996 U. Ill. L. Rev. 363, 367 (1996).

The systolic blood pressure deception test was simply a sphygmomanometer — a device that records blood pressure. The theory behind this device was that if a person lied, the resulting stress and fear of detection would cause that person's blood pressure to rise. J. Tarantino, Strategic Use of Scientific Evidence (1988) § 6.01, p. 205. The systolic blood pressure deception test was developed by William Marston, and was actually administered by Marston in Frye. Marston was also the creator of the "Wonder Woman" comic book character, who is well known for her truth-inducing magic lasso. R. Underwood, "Truth Verifiers: From the Hot Iron to the Lie Detector," 84 Ky. L.J. 597, 629 (1995-96).

In considering the defendant's claim in Frye, the Court of Appeals for the District of Columbia first determined that "general acceptance" in the scientific community was a precondition to the admissibility of any scientific evidence. Frye v. United States, supra, 293 F. 1014. "Just when a scientific principle or discovery crosses the line between the experimental and demonstrable stages is difficult to define. Somewhere in this twilight zone the evidential force of the principle must be recognized, and while courts will go a long way in admitting expert testimony deduced from a well-recognized scientific principle or discovery, the thing from which the deduction is made must be sufficiently established to have gained general acceptance in the particular field in which it belongs." Id. The court affirmed the trial court's exclusion of the systolic blood pressure evidence because the blood pressure device had "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." Id. Subsequently, Frye and its "general acceptance" standard were expressly adopted by a number of state and federal courts. See, e.g., United States v. Alexander, 526 F.2d 161 (8th Cir. 1975); Pulakis v. State, 476 P.2d 474 (Alaska 1970); see generally P. Giannelli, "The Admissibility of Novel Scientific Evidence: Frye v. United States, a Half-Century Later," 80 Colum. L. Rev. 1197, 1205 (1980).

In 1993, however, the federal standard for the admissibility of scientific evidence changed as a result of the decision of the United States Supreme Court in Daubert v. Merrell Dow Pharmaceuticals, Inc., supra, 509 U.S. 579. In Daubert, the court held that rule 702 of the Federal Rules of Evidence, which was enacted in 1975, had superseded the Frye test. Id., 587. The court concluded that "[n]othing in the text of [rule 702] establishes `general acceptance' as an absolute prerequisite to admissibility. . . . That austere standard, absent from, and incompatible with, the Federal Rules of Evidence, should not be applied in federal trials." (Citations omitted.) Id., 588-89.

Rule 702 of the Federal Rules of Evidence provides: "Testimony by Experts
"If scientific, technical, or other specialized knowledge will assist the trier of fact to understand the evidence or to determine a fact in issue, a witness qualified as an expert by knowledge, skill, experience, training, or education, may testify thereto in the form of an opinion or otherwise."

Instead, a federal trial court has a responsibility to determine, pursuant to rule 702, whether the proffered evidence will "assist the trier of fact." Id., 589. This entails a two part inquiry: "whether the reasoning or methodology underlying the [scientific theory or technique in question] is scientifically valid and . . . whether that reasoning or methodology properly can be applied to the facts in issue." Id., 592-93. In other words, before it may be admitted, the trial judge must find that the proffered scientific evidence is both reliable and relevant.

More specifically, the first requirement for scientific evidence to be admissible under rule 702 is that the subject of the testimony must be scientifically valid, meaning that it is scientific knowledge rooted "in the methods and procedures of science"; id., 590; and is "more than subjective belief or unsupported speculation." Id. This requirement "establishes a standard of evidentiary reliability"; id.; as, "[i]n a case involving scientific evidence, evidentiary reliability will be based upon scientific validity." (Emphasis in original.) Id., 590-91 n. 9.

The court emphasized that "[t]he focus, of course, must be solely on principles and methodology, not on the conclusions that they generate." Daubert v. Merrell Dow Pharmaceuticals, Inc., supra, 509 U.S. 595; see generally part I C 1 of this opinion.

The court listed four nonexclusive factors for federal judges to consider in determining whether a particular theory or technique is based on scientific knowledge: (1) whether it can be, and has been, tested; (2) whether the theory or technique has been subjected to peer review and publication; (3) the known or potential rate of error, including the existence and maintenance of standards controlling the technique's operation; and (4) whether the technique is, in fact, generally accepted in the relevant scientific community. Id., 593-94. The court emphasized, however, that the inquiry is "a flexible one. Its overarching subject is the scientific validity — and thus the evidentiary relevance and reliability — of the principles that underlie a proposed submission." Id., 594-95. Indeed, the court explicitly noted that other factors "may well have merit . . . [t]o the extent that they focus on the reliability of evidence as ensured by the scientific validity of its underlying principles. . . ." Id., 594-95 n. 12.

The second condition that scientific evidence must satisfy in order to be admissible under rule 702 is that it must "fit" the case in which it is presented. Id., 591. In other words, proposed scientific testimony must be demonstrably relevant to the facts of the particular case in which it is offered, and not simply be valid in the abstract. "The study of the phases of the moon, for example, may provide valid scientific `knowledge' about whether a certain night was dark, and if darkness is a fact in issue, the knowledge will assist the trier of fact. However (absent creditable grounds supporting such a link), evidence that the moon was full on a certain night will not assist the trier of fact in determining whether an individual was unusually likely to have behaved irrationally on that night." Id.

Finally, the court emphasized that even if a scientific theory or technique satisfied both of the previous criteria and thus would be admissible under a rule 702 analysis, it can still be excluded for failure to satisfy some other federal rule of evidence. Id., 595. Most important, it can still be excluded for failure to satisfy rule 403, which allows for the exclusion of relevant evidence "`if its probative value is substantially outweighed by the danger of unfair prejudice, confusion of the issues, or misleading the jury. . . .'" Id.

Rule 403 of the Federal Rules of Evidence provides: "Exclusion of Relevant Evidence on Grounds of Prejudice, Confusion, or Waste of Time
"Although relevant, evidence may be excluded if its probative value is substantially outweighed by the danger of unfair prejudice, confusion of the issues, or misleading the jury, or by considerations of undue delay, waste of time, or needless presentation of cumulative evidence."

Because Daubert was premised on an interpretation of a federal rule of evidence, its rejection of Frye is not binding authority on state courts. Nonetheless, subsequent to the Daubert decision, several states that had theretofore followed Frye reconsidered the issue and adopted the Daubert standard. See, e.g., State v. Alberico, 116 N.M. 156, 861 P.2d 192, 201-203 (1993) ("The Frye test . . . has been a part of New Mexico evidence law since 1952. . . . [T]he critical issue is whether the Frye test is a legitimate means for determining what is and what is not scientific knowledge. We hold that it is not and that the Frye test should be rejected as an independent controlling standard of admissibility. . . . Several factors could be considered by a trial court in assessing the validity of a particular scientific technique to determine if it is scientific knowledge. . . . See Daubert [v. Merrell Dow Pharmaceuticals, Inc., supra, 509 U.S. 593-95]. . . ." [Citations omitted; internal quotation marks omitted.]).

See also Newhart v. State, 669 N.E.2d 953, 955 (Ind. 1996); Cecil v. Commonwealth, 888 S.W.2d 669, 675 (Ky. 1994); State v. Foret, 628 So.2d 1116, 1123 (La. 1993); Commonwealth v. Lanigan, 419 Mass. 15, 25-26, 641 N.E.2d 1342 (1994); Taylor v. State, 889 P.2d 319, 328-39 (Okla.Cr. 1995); State v. Hofer, 512 N.W.2d 482, 484 (S.D. 1994); State v. Brooks, 162 Vt. 26, 30, 643 A.2d 226 (1994); Craddock v. Watson, 197 W. Va. 62, 475 S.E.2d 62, 66-67 (1996).
Several other states, which had already rejected Frye by the time Daubert was decided, have explicitly noted that their jurisprudence is already in conformance with the Daubert approach. See, e.g., Jones v. State, 314 Ark. 289, 294, 862 S.W.2d 242 (1993); Nelson v. State, 628 A.2d 69, 73 (Del. 1993); State v. Quattrocchi, 681 A.2d 879, 884 (R.I. 1996).

B

We now address the question of the proper standard for the threshold admissibility of scientific evidence in this state. We begin by noting that, at present, Connecticut nominally follows the Frye rule. See, e.g., State v. Sivri, 231 Conn. 115, 153-54, 646 A.2d 169 (1994) (restriction fragment length polymorphism procedure as part of DNA analysis "generally accepted in the scientific community"); State v. John, 210 Conn. 652, 678, 557 A.2d 93, cert. denied, 493 U.S. 824, 110 S.Ct. 84, 107 L.Ed.2d 50 (1989) (use of photographs by forensic entomologist to determine age of fly larvae found on body satisfied Frye); State v. McClary, 207 Conn. 233, 245-48, 541 A.2d 96 (1988) (testimony on "shaken baby syndrome" admissible under Frye); State v. Sherman, 38 Conn. App. 371, 409, 662 A.2d 767, cert. denied, 235 Conn. 905, 665 A.2d 905 (1995) ("Connecticut courts utilize the Frye standard in appraising the admissibility of evidence derived from innovative scientific techniques" [internal quotation marks omitted]).

See, e.g., J. Meaney, "From Frye to Daubert: Is a Pattern Unfolding?" 35 Jurimetrics J. 191, 194 (1995) (characterizing Connecticut as following Frye standard).

A closer examination of our precedent, however, reveals that on many occasions we have declined to apply Frye when considering expert scientific testimony. See, e.g., State v. Hasan, 205 Conn. 485, 494, 534 A.2d 877 (1987) (admissibility of expert testimony by podiatrist that sneakers that left prints at crime scene belonged to defendant "did not depend on general acceptance of his theories in the scientific community"). Indeed, we have specifically observed that "[i]f a duly qualified expert testifies that in his opinion [a scientific] test or device is reliable, but his opinion is not shared by the scientific community, either because it is novel and experimental or because they disagree with its bases, the court still has discretion to admit the evidence because competent proof is not rendered inadmissible merely by the fact that others take issue with it." (Emphasis added; internal quotation marks omitted.) State v. Ortiz, 198 Conn. 220, 227, 502 A.2d 400 (1985) (forensic odontologist's identification of person by bite mark on apple admissible).

It is clear that we have been moving toward a validity standard for a number of years. We believe that it is time to complete that process, and that "the Daubert [reliability] approach will provide structure and guidance to what has until now been a potentially confusing and sparsely defined area of legal analysis in our state jurisprudence." Taylor v. State, 889 P.2d 319, 329 (Okla.Cr. 1995). Accordingly, we conclude that the Daubert approach should govern the admissibility of scientific evidence in Connecticut.

The Court of Criminal Appeals of Oklahoma, in addressing the issue of whether to abandon Frye in favor of Daubert, described a legal landscape regarding the admissibility of scientific evidence very similar to that which currently exists in Connecticut. The court observed that "[a] review of our pertinent caselaw reveals that this Court has not consistently relied upon Frye when faced with questions involving the admissibility of expert testimony describing novel scientific evidence. In fact, since adopting Frye in 1951, we have specifically cited to it in only six reported cases. While a number of our cases have mentioned the `general acceptance' standard, several others appear to have utilized [a `helpfulness' standard] in analyzing the admission of expert testimony describing novel scientific evidence." Taylor v. State, supra, 889 P.2d 328-29. The court decided to adopt Daubert. Id.

1

We first discuss why a special standard for the admissibility of scientific evidence is required at all. Implicit in both Frye and Daubert is the notion that a trial judge should, by one method or another, serve as a "gatekeeper" and make a preliminary assessment of the validity of scientific testimony before allowing the fact finder even to consider it. A number of commentators, however, have suggested that the validity of proffered scientific evidence should go solely to its weight, not to its admissibility, and thus the fact finder should have a chance to consider all scientific evidence that is submitted. For example, the latest edition of Charles McCormick's treatise on evidence provides that in making admissibility determinations, "the traditional standards of relevancy and the need for expertise — and nothing more — should govern." 1 C. McCormick, Evidence (4th Ed. 1992) § 203, p. 874.

There is a concession in McCormick's treatise that, like all other relevant evidence, relevant scientific testimony can be excluded for the "familiar [reasons] of prejudicing or misleading the jury or consuming undue amounts of time." 1 C. McCormick, supra, § 203, p. 875.

Under this view, rule 702 of the Federal Rules of Evidence should not be interpreted as requiring judges independently to assess the validity of proffered scientific evidence, or to make admissibility decisions therefrom. Instead, "trial courts [should only] insure that expert witnesses are trustworthy — capable of explaining their reasoning in a manner that permits the fact-finder to assess the testimony effectively — and so knowledgeable in their subject area that the inferences they draw are helpful. Thus, a court faced with scientific proof offers should not do what the jury does — that is, decide whether the evidence can be believed and how much of it should be believed. Instead, the court should consider whether the fact-finder has enough information at its disposal to decide these issues for itself." R. Dreyfuss, "Is Science a Special Case? The Admissibility of Scientific Evidence After Daubert v. Merrell Dow," 73 Tex. L. Rev. 1779, 1801 (1995).

We disagree with the preceding analyses and conclude that the validity of the methodologies underlying proffered scientific evidence should be considered in determining the admissibility of such evidence, as well as in determining its weight. Accordingly, we also conclude that it is proper for a trial judge to serve a gatekeeper function.

a

In this regard, we first note the concern expressed by many authorities that juries will be overwhelmed by complex scientific evidence and will give such evidence more weight than it deserves. We acknowledge, however, that other commentators have specifically asserted that juries will not be overly impressed by such evidence. At present, empirical data regarding the impact of scientific testimony on juries is almost entirely lacking. As a consequence, whether jurors will unduly credit scientific evidence is quite uncertain. Indeed, direct empirical evidence regarding the impact of scientific evidence on judges is also lacking.

"`Lay jurors tend to give considerable weight to "scientific" evidence when presented by "experts" with impressive credentials.'" People v. Leahy, 8 Cal.4th 587, 595, 882 P.2d 321, 34 Cal.Rptr.2d 663 (1994). This is so even though "juries usually lack any reliable or consistent basis for evaluating the credibility of expert witness testimony." "Developments in the Law — Confronting the New Challenges of Scientific Evidence," 108 Harv. L. Rev. 1481, 1509 (1995). Accordingly, juries "may defer too much to the opinion of experts if the subject matter is beyond common knowledge and everyday experience and if the opinion is given with the air of authority that commonly accompanies an expert's testimony." (Internal quotation marks omitted.) State v. Cavaliere, 140 N.H. 108, 109, 663 A.2d 97 (1995); see also E.I. du Pont de Nemours Co. v. Robinson, 923 S.W.2d 549, 553 (Tex. 1995); P. Giannelli, supra, 80 Colum. L. Rev. 1239-40.

These authorities believe that modern juries are technologically sophisticated enough not to be overwhelmed by scientific evidence. See, e.g., R. Dreyfuss, supra, 73 Tex. L. Rev. 1798; J. Sanders, "From Science to Evidence: The Testimony on Causation in the Bendectin Cases," 46 Stan. L. Rev. 1, 3 (1993) (concluding that faulty presentation, and not inherent difficulty of subject matter, is largely responsible for jury difficulties with scientific evidence).

See D. Shuman, A. Champagne E. Whitaker, "Juror Assessments of the Believability of Expert Witnesses: A Literature Review," 36 Jurimetrics J. 371, 382 (1996) ("Although the picture painted by the evolving social science research on juror assessments of the believability of experts is far from complete or consistent, an unsettling image emerges. The typical juror forms impressions of experts stereotypically, based on the occupation of the experts, and superficially, based on the personal characteristics of the experts.").
The authors of the preceding article themselves subsequently surveyed 156 jurors from cases in which scientific evidence had been presented. D. Shuman, A. Champagne E. Whitaker, "Assessing the Believability of Expert Witnesses: Science in the Jurybox," 37 Jurimetrics J. 23 (1996). That study asked the jurors to rate the importance various factors had played in their assessment of expert scientific testimony. Id., 24-25. The study concluded that in assessing an expert's credibility, jurors properly placed the greatest emphasis on the expert's qualifications, familiarity with the case, quality of reasoning and impartiality. Id. 26, 27-28. The authors admit, however, that, because this study was based entirely on self-reporting by the jurors of what factors they had considered important, it was possible that jurors who in fact "were impressed by witnesses (for irrational reasons)" were simply justifying their actions by "say[ing], after the fact, that these witnesses [were] familiar with the facts and use[d] good reasoning." Id., 28 n. 20; see also P. Giannelli, supra, 80 Colum. L. Rev. 1240; cf. footnote 61 of this opinion.

But see L. Loevinger, "Science as Evidence," 35 Jurimetrics J. 153, 186 (1995) (arguing that judges are better able to cope with scientific evidence); J. Sanders, "From Science to Evidence: The Testimony on Causation in the Bendectin Cases," 46 Stan. L. Rev. 1, 84 (1993).

Although the effect of scientific evidence with regard to both judges and juries is uncertain, we note that, purely as a procedural matter, a judge is in a much better position than a juror to assess accurately the fundamental validity of such evidence. This is due to the different roles each serves at trial and the concomitant powers each has. "[T]he process of judicial decision-making, including the methods by which information is received and decisions are reviewed, is better suited than is the rather awkward process of juror decision-making for evaluation of evidence that is beyond at least the usual ken of laypeople, judges, and jurors alike." R. Friedman, "The Death and Transfiguration of Frye," 34 Jurimetrics J. 133, 144 (1994); see also B. Black, F. Ayala C. Saffran-Brinks, "Science and the Law in the Wake of Daubert: A New Search for Scientific Knowledge," 72 Tex. L. Rev. 715, 787 (1994) ("we agree with Daubert's preference for preliminary judicial screening only because judges are in a better position than juries to acquire and consider the kind of information that bears on the resolution of [scientific evidence] disputes").

For example, juries mainly have to rely on in-court testimony for their understanding of scientific evidence; that is, they are largely dependent upon the presentations of the parties and their experts. There is evidence, however, that expert presentations may often be misleading; J. Sanders, "From Science to Evidence: The Testimony on Causation in the Bendectin Cases," 46 Stan. L. Rev. 1, 3 (1993); and, at the same time, that cross-examination of experts may often be difficult and ineffective in bringing out flaws in the expert's reasoning. B. Black, F. Ayala C. Saffran-Brinks, supra, 72 Tex. L. Rev. 789; S. Gross, "Expert Evidence," 1991 Wis. L. Rev. 1113, 1172 (1991).

Judges, on the other hand, have the benefit of reviewing briefs and other documents. B. Black, F. Ayala C. Saffran-Brinks, supra, 72 Tex. L. Rev. 788. Indeed, a Connecticut trial judge has the power to request supplemental briefing on any issue that needs clarification, including issues surrounding proffered scientific evidence. Furthermore, Practice Book § 881 explicitly authorizes a judge presiding over a criminal case to appoint an independent expert when necessary. Such an appointment could certainly be made to assist the judge in evaluating proposed scientific testimony. Cf. Taylor v. State, supra, 889 P.2d 342 (Lumpkin, J., concurring).

Practice Book § 881 provides: "— Judicial Appointment of Expert Witnesses
"Whenever the judicial authority deems it necessary, on its own motion it may appoint any expert witnesses of its own selection. An expert witness shall not be appointed by the judicial authority unless the expert consents to act. A witness so appointed shall be informed of his or her duties by the judicial authority in writing, a copy of which shall be filed with the clerk, or the witness shall be informed of his or her duties at a conference in which the parties shall have an opportunity to participate. A witness so appointed shall advise the parties of his or her findings, if any, and may thereafter be called to testify by the judicial authority or by any party. A witness so appointed shall be subject to cross-examination by each party. The judicial authority may determine the reasonable compensation for such a witness and direct payment out of such funds as may be provided by law. This section shall not apply to appointments made pursuant to Gen. Stat., § 54-56d."
We do not suggest, moreover, that this rule of practice limits any inherent judicial power that a trial court may also have to appoint an expert in an appropriate civil case.

Moreover, "[a]n important element in adjudication . . . is the development of judicial expertise through repeated exposure to and familiarity with similar scientific issues." J. Kesan, "An Autopsy of Scientific Evidence in a Post- Daubert World," 84 Geo. L.J. 1985, 2038 (1996). Judges, on the one hand, are likely to gain familiarity with various common procedures, and with scientific techniques in general, by virtue of presiding over multiple cases involving such issues. "Over time, most judges will probably develop at least some facility for understanding science beyond the typical juror's level of understanding." B. Black, F. Ayala C. Saffran-Brinks, supra, 72 Tex. L. Rev. 788. On the other hand, "[t]aking the time required to educate jurors and to present them with similarly detailed information could easily overwhelm the other issues in a case." Id.

Given this background, we conclude that a gatekeeping role for trial judges in relation to scientific evidence is appropriate. Although the extent to which juries give scientific evidence undue deference is uncertain, the potential risk can be greatly reduced simply by allowing the judge, as the participant in the judicial process with both the greater access and ability to gather relevant information, to exclude wholly invalid scientific testimony altogether. Moreover, a trial judge who does admit scientific evidence will be in a better position, by virtue of the knowledge gained during the preliminary assessment, to conduct the trial and instruct the jury in such a way as to minimize the risk that jurors will give that evidence undue deference.

We emphasize that we do not assume that jurors lack the capacity to analyze and understand complex scientific evidence. Indeed, once such evidence is admitted, we ask jurors to do just that. Nonetheless, we conclude that it is incumbent upon trial judges to minimize, to the extent possible, the confusion and prejudice that scientific evidence may generate. Moreover, the very fact that proponents of scientific evidence will already have had to organize their evidence in a clear, comprehensible manner in order to pass the threshold validity assessment will, it is hoped, lead to clear, comprehensible presentations to the jury as well, thus aiding jurors in their ultimate assessments.

b

In addition, we believe it is proper for trial judges to serve as gatekeepers for scientific evidence because a relevance standard of admissibility inherently involves an assessment of the validity of the proffered evidence. More specifically, if scientific evidence has no grounding in scientific fact, but instead is based on conjecture and speculation, it cannot in any meaningful way be relevant to resolving a disputed issue.

McCormick himself, although a proponent of the relevance standard for scientific evidence, initially recognized this fact. In the first edition of his treatise on evidence, McCormick rejected the Frye test in favor of a relevancy standard, in much the same manner as that test is rejected in the fourth edition of his treatise. See C. McCormick, Evidence (1954) § 170, pp. 363-64.

He went on to conclude, however, that such a standard would entail "deflat[ing] the requirement [of general acceptance] to the normal standard which simply demands that the theory or device be accepted by a substantial body of scientific opinion. . . ." (Emphasis added.) Id., § 174, pp. 371-72. McCormick thus implicitly acknowledged that some indication of scientific validity — accomplished here by a showing of "substantial" acceptance — is necessary for scientific evidence even to be relevant. It is true that the previous passage was removed from subsequent editions of McCormick's treatise. See, e.g., C. McCormick, Evidence (2d Ed. 1972) § 207, pp. 506-507. Nonetheless, we believe that McCormick's original understanding was correct. See also P. Giannelli, supra, 80 Colum. L. Rev. 1233-34.

2

Having concluded that Connecticut judges should exercise a gatekeeper function with regard to scientific evidence, we now briefly explain our conclusion that the Frye "general acceptance" standard is not adequate for this role. We note that even before the decision in Daubert, the Frye rule was widely criticized. "[S]cholars have criticized the Frye approach for being unduly conservative and for abandoning the fundamental evidentiary [principle] of . . . probity." "Developments in the Law — Confronting the New Challenges of Scientific Evidence" ("Developments in the Law"), 108 Harv. L. Rev. 1481, 1486 (1995). These critics observe that scientific pioneers and dissenters are occasionally right. "`When photography was first introduced, it was seriously questioned whether pictures thus created could properly be introduced into evidence, but this method of proof, as well as by means of x-rays and the microscope, is now admitted without question.' [ People v. Jennings, 252 Ill. 534, 548, 96 N.E. 1077 (1911)]. As with most scientific phenomena, the passage of time can serve . . . to demonstrate the reliability and acceptance of a once speculative and unproved premise." State v. Bullard, 312 N.C. 129, 146, 322 S.E.2d 370 (1984). Excluding scientific views simply because they are not "generally accepted," without any further consideration of their validity, thus contravenes "the liberal nature . . . of modern evidentiary law." R. Friedman, supra, 34 Jurimetrics J. 133-34.

See generally 1 C. McCormick, Evidence (4th Ed. 1992) § 203, pp. 871-73 ("Especially in the last two decades . . . the Frye standard has been subjected to critical analysis, limitation, modification, and finally, outright rejection. . . . A drumbeat of criticism of the Frye test provides the background music to the movement away from the general acceptance test." [Internal quotation marks omitted.]).

We are persuaded by these criticisms of the Frye test. We conclude that an admissibility test for scientific evidence premised solely on its "general acceptance" is conceptually flawed and therefore must be rejected. In doing so, we follow in the footsteps of the many jurisdictions that, both before and after the Daubert decision, have chosen to reject the Frye standard as an absolute prerequisite to the admissibility of scientific evidence. See, e.g., State v. Pennington, 327 N.C. 89, 98, 393 S.E.2d 847 (1990) (admissibility of scientific evidence should be based on "the reliability of the scientific method rather than its popularity within a scientific community").

Even before Daubert, several courts rejected the Frye standard as not embodying an adequate gatekeeper function. Several federal courts expressly rejected the rigid "nose-counting" of Frye and instead based admissibility decisions on the "helpfulness" to the fact finder of proffered scientific evidence. DeLuca v. Merrell Dow Pharmaceuticals, Inc., 911 F.2d 941, 951, 955 (3d Cir. 1990); see also United States v. Jakobetz, 955 F.2d 786 (2d Cir.), cert. denied, 506 U.S. 834, 113 S.Ct. 104, 121 L.Ed.2d 63 (1992); Clinchfield R. Co. v. Lynch, 784 F.2d 545 (4th Cir. 1986); United States v. Downing, 753 F.2d 1224 (3d Cir. 1985); United States v. Mustafa, 22 M.J. 165 (C.M.A.), cert. denied, 479 U.S. 953, 107 S.Ct. 444, 93 L.Ed.2d 392 (1986). Many state courts also followed this approach. See, e.g., Barmeyer v. Montana Power Co., 202 Mont. 185, 191-92, 657 P.2d 594 (1983); State v. Brown, 297 Or. 404, 409, 416-17, 687 P.2d 751 (1984); State v. Dery, 545 A.2d 1014, 1017 (R.I. 1988).
In total, twenty states had already rejected Frye and had adopted a "helpfulness" or "relevance" test for admissibility of scientific evidence by the time Daubert was decided in 1993. See J. Meaney, "From Frye to Daubert: Is a Pattern Unfolding?" 35 Jurimetrics J. 191, 194-98 (1995) (compiling state judicial responses to Daubert on state-by-state basis).

The movement among state courts away from Frye and toward the Daubert approach has continued since the Daubert decision. See footnote 9; see also Harrison v. State, 644 N.E.2d 1243, 1252 (Ind. 1995); State v. Hofer, 512 N.W.2d 482, 484 (S.D. 1994); State v. Brooks, 162 Vt. 26, 30, 643 A.2d 226 (1994).

We recognize that, to date, most of the states that have rejected Frye and have adopted a Daubert approach; see, e.g., footnotes 20 and 21; have state evidentiary codes that contain an analogue to rule 702 of the Federal Rules of Evidence, which Connecticut lacks, and that this fact often plays some role in their decisions. But see Lattarulo v. State, 261 Ga. 124, 126, 401 S.E.2d 516 (1991) (even though Georgia lacks rule 702 analogue, test for admissibility of scientific evidence is "whether the procedure or technique `has reached a scientific stage of verifiable certainty'"); see generally J. Meaney, "From Frye to Daubert: Is a Pattern Unfolding?" 35 Jurimetrics J. 191, 199 (1995).
We also recognize that with the exception of Georgia and, now, Connecticut, every other state without a rule 702 analogue — Alabama, California, Illinois, Maryland and New York — currently operates under a Frye standard. We note, however, that three of these states — Alabama, Illinois and Maryland — simply have not addressed the issue. See, e.g., People v. Miller, 173 Ill.2d 167, 670 N.E.2d 721, 731 n. 3 (1996) (explicitly noting that Daubert issue has not yet been presented to court). Thus, only two of the states without a rule 702 analogue — New York and California — have expressly rejected Daubert and have retained Frye. See People v. Leahy, 8 Cal.4th 587, 595, 882 P.2d 321, 34 Cal.Rptr.2d 663 (1994); People v. Wesley, 83 N.Y.2d 417, 423, 633 N.E.2d 451, 611 N.Y.S.2d 97 (1994). Moreover, although California ostensibly still follows Frye, the California Supreme Court has implicitly recognized the imprudence of the strict "general acceptance" standard by holding that "`general acceptance' does not require unanimity, a consensus of opinion, or even majority support by the scientific community." People v. Leahy, supra, 601.
In any event, we conclude that our reasons for rejecting Frye and adopting a Daubert approach, as explained in this opinion, are valid independent of the existence of a codified rule 702 analogue.

3

We now turn to our reasons for concluding that the Daubert approach to the admissibility of scientific evidence is the proper approach for judges to follow in performing their gatekeeper role. In doing so, we are mindful of the fact that almost every other jurisdiction that has rejected Frye has also, either explicitly or effectively, adopted the Daubert multifactor "validity" approach to admissibility. We agree with, and join in, this movement.

See, e.g., footnotes 20 and 21.

Our reasons for adopting Daubert are based upon our understanding of that case. By its own terms, the opinion of the United States Supreme Court in Daubert v. Merrell Dow Pharmaceuticals, Inc., supra, 509 U.S. 579, sets forth a conceptual approach to the admissibility of scientific evidence, and not a functional test therefor. As discussed previously, the essential holding of the Supreme Court is the general principle that, as a threshold matter, and subject still to the rules of evidence generally, scientific evidence should be admitted in court only upon some showing of its scientific validity. Id., 592-93; see generally part I A of this opinion. Although the court provides "general observations" as to some factors that might be relevant to this determination, it explicitly states that "[m]any factors will bear on the inquiry, and we do not presume to set out a definitive checklist or test." (Emphasis added.) Daubert v. Merrell Dow Pharmaceuticals, Inc., supra, 593. " Daubert does not purport to answer all questions or to resolve all problems associated with scientific testimony." L. Loevinger, "Science as Evidence," 35 Jurimetrics J. 153, 179 (1995); see also "Developments in the Law," supra, 108 Harv. L. Rev. 1513 ( Daubert "offer[s] little practical guidance").

Critics of Daubert emphasize this indefiniteness. Chief Justice Rehnquist, in his concurring and dissenting opinion, for example, although agreeing that "Rule 702 confides to the judge some gatekeeping responsibility"; Daubert v. Merrell Dow Pharmaceuticals, Inc., supra, 509 U.S. 600; commented that "[q]uestions arise simply from reading [the `general observations'] part of the Court's opinion, and countless more questions will surely arise when hundreds of district judges try to apply its teaching to particular offers of expert testimony." Id.; see also R. Dreyfuss, supra, 73 Tex. L. Rev. 1780 ( Daubert's "reach is indeterminate, and the few specific guidelines announced in the opinion do not begin to cover the issues that judges who follow it must consider").

We view Daubert's indefiniteness not as a flaw, but as a necessity. The term "scientific evidence" covers a large variety of subjects. The Federal Judicial Center, for example, included in its Reference Manual on Scientific Evidence essays on the broad topics of epidemiology, toxicology, survey research, DNA evidence, multiple regression, and estimation of economic losses in damages awards. See Reference Manual on Scientific Evidence, J. Moore, Federal Practice (1995). In addition to those subjects, courts have also treated as scientific evidence testimony on subjects ranging from the causation of a plaintiff's cataracts; O'Conner v. Commonwealth Edison Co., 13 F.3d 1090 (7th Cir. 1994); to an analysis of bloodstain patterns at a crime scene; State v. Goode, 341 N.C. 513, 461 S.E.2d 631 (1995); to, of course, polygraph evidence; State v. Beard, 194 W. Va. 740, 461 S.E.2d 486 (1995); and beyond.

Accordingly, we conclude that it is impossible to formulate a specific, clearly defined test that provides judges with a precise, complete list of factors to consider in evaluating the entire class of scientific evidence. No purely mechanical test based on a finite number of set considerations can, in and of itself, truly guide judges with regard to the admissibility of all of the varied and eclectic types of scientific evidence. Indeed, "[e]ach factor may shed some light on the scientific merits of the evidence, but none illuminates much of the total picture. Without a conceptual framework, using [mechanical] multiple-factor tests to evaluate science is like trying to light up a ball park with a few misaimed spotlights." B. Black, F. Ayala C. Saffran-Brinks, supra, 72 Tex. L. Rev. 735.

It is therefore no surprise that courts that do apply either Frye or Daubert in a purely mechanical manner — simply looking at whether certain, specified factors are met — do not reach consistent results across jurisdictions regarding the admissibility of particular types of scientific evidence. By reading the individual factors narrowly or broadly, courts can skew the analysis toward the end they desire. As a result, "courts purporting to apply the same . . . standard to essentially the same problem frequently reach conflicting conclusions." "Developments in the Law," supra, 108 Harv. L. Rev. 1494. Thus, for example, courts ostensibly applying Frye have both admitted and excluded voiceprint evidence, as have courts ostensibly applying multifactor validity checklists. See generally B. Black, F. Ayala C. Saffran-Brinks, supra, 72 Tex. L. Rev. 735-39; "Developments in the Law," supra, 1494-98.

For example, commentators have argued that, depending on the desired outcome, courts applying Frye have "differed on the issues of who constitutes the relevant community for acceptance, as well as what precisely that community must have approved in a given case." (Emphasis in original.) "Developments in the Law," supra, 108 Harv. L. Rev. 1486; see also P. Giannelli, supra, 80 Colum. L. Rev. 1208-15.

"Voiceprints are graphic depictions of sound produced by a device called a sound spectrograph. Their forensic use is based on the premise that each person's voice produces a unique `picture.' Voiceprints typically appear in court when there is a recording of an incriminating telephone call and the identity of the caller is at issue." B. Black, F. Ayala C. Saffran-Brinks, supra, 72 Tex. L. Rev. 735.

Indeed, because a generic, mechanical standard is often inapposite to an understanding of the validity of a particular type of scientific evidence, such a specific standard frequently deflates into no standard at all. This is because "[s]uperficially clear doctrinal standards that are in fact indeterminate nearly to the point of uselessness serve only to obstruct the rules that govern the system." "Developments in the Law," supra, 108 Harv. L. Rev. 1497. That is, mechanical factor tests emphasize specific considerations that are often irrelevant or unduly narrow while simultaneously obscuring the underlying principles upon which a court should be focusing. As a result, such tests have "little analytical value [and] serve primarily as labels to justify a court's instinctive reaction." B. Black, F. Ayala C. Saffran-Brinks, supra, 72 Tex. L. Rev. 735. Thus, the reality under most mechanical tests for the admissibility of scientific evidence is that, beneath the illusion of a rigid standard, the court in fact has essentially unfettered discretion.

We conclude that a test embodying a general, overarching approach to the threshold admissibility of scientific evidence is required. Although such a standard is more vague on its surface, it will in fact discourage untrammeled discretion by giving trial courts a workable principle to follow. Moreover, we believe that Daubert's focus on scientific validity properly directs trial judges to the core issue that they should address as gatekeepers of scientific evidence. For, as explained previously, scientific evidence is likely neither relevant nor helpful to the fact finder if it does not meet some minimum standard of validity. See generally part I B 1 of this opinion.

We emphasize, however, as did the court in Daubert v. Merrell Dow Pharmaceuticals, Inc., supra, 509 U.S. 595, that simply satisfying this minimum threshold standard does not automatically guarantee admission of scientific testimony. Such testimony, even if it satisfies Daubert, is still subject to the requirements imposed on all evidence by the rules of evidence generally. In particular, scientific testimony, like all testimony, can be excluded if its prejudicial impact outweighs its probative value. See part I A and D of this opinion.

C

We now examine, pursuant to our conception of the Daubert approach, the mechanics and scope of a Daubert assessment. Indeed, our reasons for adopting the Daubert approach at all are intertwined with, and dependent upon, our conception of how that approach should operate.

1

We begin by noting the distinction under the Daubert approach between the methodologies underlying an expert's scientific testimony and the expert opinion itself. As the court in Daubert noted, the focus of a validity assessment "must be solely on principles and methodology, not on the conclusions that they generate." Daubert v. Merrell Dow Pharmaceuticals, Inc., supra, 509 U.S. 595. So long as the methodology underlying a scientific opinion has the requisite validity, the testimony derived from that methodology meets the Daubert threshold for admissibility, even if the judge disagrees with the ultimate opinion arising from that methodology, and even if there are other methodologies that might lead to contrary conclusions. Thus, a judge should admit scientific testimony when "there are good grounds for [the] expert's conclusion, even if the judge thinks that there are better grounds for some alternative conclusion. . . ." In re Paoli R. Yard PCB Litigation, 35 F.3d 717, 744 (3d Cir. 1994).

For example, in Cella v. United States, 998 F.2d 418, 420 (7th Cir. 1993), the plaintiff claimed that his polymyositis — an inflammation of the muscles involving both the upper and lower extremities — was caused by certain events of physical and emotional trauma he suffered while working aboard a Navy vessel. The plaintiff presented the testimony of a single physician supporting his theory of liability, while the defendant presented several witnesses who testified that the cause of polymyositis is, in fact, unknown, and thus could not be attributed to the aforementioned traumas with any degree of certainty. Id., 421-22. After the court, in a bench trial, found in favor of the plaintiff, the defendant appealed, claiming, in part, that the testimony of the plaintiff's physician should not even have been admitted "in light of the testimony of the defendant's medical experts and the abundance of medical literature stating that the etiology of polymyositis is unknown." Id., 423.

The examples cited throughout part I C 1 of the opinion are merely illustrative. We do not necessarily endorse the conclusions reached in these cases.

After a thorough review of the bases upon which the plaintiff's physician had based his opinion, the court upheld the admission of his testimony. "Although [the plaintiff's physician's] conclusion differs from those of the defendant's medical experts, he has utilized an accepted methodology in reaching his conclusion — namely, analysis of medical literature and case study comparison with the individual characteristics of the patient's case to determine etiology." Id., 426. Thus, although there was little support in the literature for the physician's specific conclusion regarding the cause of the plaintiff's injury, the court found that he had employed a proper and thorough diagnostic methodology. Accordingly, it was not error to admit the physician's testimony, because "`[a]s long as the expert's methodology is well founded, the nature of the expert's conclusion is generally irrelevant, even if it is controversial or unique." Id. Once the methodology underlying an expert conclusion has been sufficiently established, the mere fact that controversy, or even substantial controversy, surrounds that conclusion goes only to the weight, and not to the admissibility, of such testimony.

We note that, although Cella v. United States, supra, 998 F.2d 418, nominally applied the Frye standard, the court's analysis is actually more consistent with the Daubert multifactor approach. See id., 424-27. Indeed, Cella demonstrates, yet again, that the label that courts apply when assessing the admissibility of scientific evidence is often completely divorced from the substance of the analysis that they undertake.

While the distinction between methodologies and conclusions is important, it is nonetheless true that the two are closely linked. It is plain that a judge's doubts about an expert's conclusions are often based upon doubts regarding the expert's methodology. "When a judge disagrees with the conclusions of an expert, it will generally be because he or she thinks that there is a mistake at some step in the investigative or reasoning process of that expert. If the judge thinks that the conclusions of some other expert are correct, it will likely be because the judge thinks that the methodology and reasoning process of the other expert are superior to those of the first expert." In re Paoli R. Yard PCB Litigation, supra, 35 F.3d 746.

Of course, even where a particular technique has been shown to satisfy Daubert, the proponent must also establish that the specific scientific testimony at issue is, in fact, derived from and based upon that methodology. The Supreme Court in Daubert referred to this concept as the "fit" requirement. Daubert v. Merrell Dow Pharmaceuticals, Inc., supra, 509 U.S. 591; see part I A of this opinion. "[A]lthough some conclusions can be reasonably inferred from the methodology employed, others cannot." J. Kesan, supra, 84 Geo. L.J. 2022. "When an expert's conclusions are not commensurate with the underlying methodology, they may be properly excluded under Daubert because they do not rely on scientific knowledge and are thus unhelpful to the jury." Id.; see, e.g., State v. Foret, 628 So.2d 1116, 1124-27 (La. 1993) (expert testimony that victim of alleged molestation had, in fact, been abused should have been excluded when testimony was based on psychological methodology designed and intended solely to assist in treatment, and not as diagnostic tool).

2

We now set forth some of the factors that various courts have considered in conducting a methodological analysis. We emphasize, however, that these factors are not exclusive. Some will not be relevant in particular cases; and some cases will call for considerations not discussed herein. "The factors a trial court will find helpful in determining whether the underlying theories and techniques of the proffered evidence are scientifically reliable will differ with each particular case." E.I. du Pont de Nemours Co. v. Robinson, 923 S.W.2d 549, 557 (Tex. 1995); see also Daubert v. Merrell Dow Pharmaceuticals, Inc., supra, 509 U.S. 593; In re Paoli R. Yard PCB Litigation, supra, 35 F.3d 742. Indeed, as discussed previously; see generally part I B 3 of this opinion; a mechanical list of mandatory factors would frustrate the entire concept underlying the Daubert approach.

Even under Daubert, courts should continue to consider whether a scientific principle has gained "general acceptance" in making admissibility determinations. Although "general acceptance" is no longer an absolute prerequisite to the admission of scientific evidence, it should, in fact, be an important factor in a trial judge's assessment. Indeed, "[w]e suspect that general acceptance in the relevant scientific community will continue to be the significant, and often the only, issue." Commonwealth v. Lanigan, 419 Mass. 15, 26, 641 N.E.2d 1342 (1994). Thus, "[a]lthough Frye may no longer be the standard for admissibility, general acceptance remains a part of the analysis, and in many cases its presence may alone be sufficient to admit the evidence." (Emphasis in original.) "Standards and Procedures for Determining the Admissibility of Expert Evidence After Daubert," 157 F.R.D. 571, 574 (1995). That is, if a trial court determines that a scientific methodology has gained general acceptance, then the Daubert inquiry will generally end and the conclusions derived from that methodology will generally be admissible. If a principle has not gained general acceptance, however, we emphasize that "a proponent of [the] scientific opinion . . . may [still] demonstrate the reliability or validity of the underlying scientific theory or process by some other means, that is, without establishing general acceptance." Commonwealth v. Lanigan, supra, 26; see also Daubert v. Merrell Dow Pharmaceuticals, Inc., supra, 509 U.S. 588-89.

We also note that some scientific principles have become so well established that an explicit Daubert analysis is not necessary for admission of evidence thereunder. By this, we do not mean to reestablish the Frye general acceptance test. We do acknowledge, however, as did the Supreme Court in Daubert, that a very few scientific principles "are so firmly established as to have attained the status of scientific law, such as the laws of thermodynamics, [and that such principles] properly are subject to judicial notice. . . ." Daubert v. Merrell Dow Pharmaceuticals, Inc., supra, 509 U.S. 592 n. 11. Evidence derived from such principles would clearly withstand a Daubert analysis, and thus may be admitted simply on a showing of relevance. For example, the Supreme Court of Montana recently noted that a Daubert analysis would not be necessary for ordinary fingerprint identification evidence to be admissible. State v. Cline, 275 Mont. 46, 55, 909 P.2d 1171 (1996). Of course, as with any other evidence, the weight that the fact finder should accord to such evidence may still be contested. Compare footnote 31.

Several other factors may properly play a role in a court's assessment of the validity of a scientific methodology. The remaining factors listed in Daubert whether that methodology has been tested and subjected to peer review, and the known or potential rate of error — are of course important. See Daubert v. Merrell Dow Pharmaceuticals, Inc., supra, 509 U.S. 592-94. Moreover, the prestige and background of the expert witness supporting the evidence can play a role in determining whether a novel technique employed by that individual is likely to have any scientific merit. State v. Goode, supra, 341 N.C. 527-28; State v. Brown, 297 Or. 404, 417, 687 P.2d 751 (1984); M. Gottesman, "Admissibility of Expert Testimony After Daubert: The `Prestige' Factor," 43 Emory L.J. 867, 878 (1994). The extent to which the scientific technique in question relies on subjective interpretations and judgments by the testifying expert, rather than on objectively verifiable criteria, can also be a factor. State v. Brown, supra, 417; E.I. du Pont de Nemours Co. v. Robinson, supra, 923 S.W.2d 557; "Developments in the Law," supra, 108 Harv. L. Rev. 1498. In the same vein, courts have looked at whether a testifying expert can present and explain the data and methodology underlying his or her scientific testimony in such a manner that the fact finder can reasonably and realistically draw its own conclusions therefrom. State v. Goode, supra, 527-28; "Developments in the Law," supra, 1502; see also State v. Hasan, supra, 205 Conn. 491. Several courts have also considered whether the scientific technique underlying the proffered expert testimony was developed and implemented solely to develop evidence for in-court use, or whether the technique has been developed or used for extrajudicial purposes. Daubert v. Merrell Dow Pharmaceuticals, Inc., 43 F.3d 1311, 1317 (9th Cir. 1995) (applying Supreme Court Daubert opinion on remand of case); State v. Goode, supra, 527-28; State v. Brown, supra, 417; E.I. du Pont de Nemours Co. v. Robinson, supra, 557.

We appreciate that many of these factors lack precision, but this indefiniteness is unavoidable. The actual operation of each factor, as is the determination of which factors should be considered at all, depends greatly on the specific context of each case in which each particular Daubert analysis is conducted. So long as trial judges remain focused on the underlying purpose behind the Daubert analysis — to establish whether a scientific methodology has sufficient validity to be helpful to the fact finder — we are confident that the previously mentioned uncertainties are not so overwhelming as to render Daubert functionally inoperative. Cf. R. Friedman, supra, 34 Jurimetrics J. 141 (ambiguities in Daubert factors "will probably be genuinely troublesome only if, contrary to the [Supreme] Court's stated intention, lower courts treat the criteria as checklist items, each of which must be satisfied for a proposition to be deemed scientific, rather than merely as factors to be weighed, along with others, into an overall balance").

3

We now turn to the threshold burden that a proponent of scientific testimony bears to establish that the testimony is admissible. "Once the party opposing the evidence objects, the proponent bears the burden of demonstrating its admissibility." E.I. du Pont de Nemours Co. v. Robinson, supra, 923 S.W.2d 557.

In addressing the showing that a proponent of scientific evidence must make, we are largely guided by the fundamental tenets of the law of evidence regarding admissibility. "Evidence is admissible when it tends to establish a fact in issue or to corroborate other direct evidence in the case." (Internal quotation marks omitted.) State v. McClendon, 199 Conn. 5, 8, 505 A.2d 685 (1986). "Evidence is not rendered inadmissible because it is not conclusive. All that is required is that evidence tend to support a relevant fact even to a slight degree, so long as it is not prejudicial or merely cumulative." (Emphasis in original.) C. Tait J. LaPlante, Connecticut Evidence (2d Ed. 1988) § 8.1.1, p. 226. Moreover, even evidence "susceptible to different interpretations" is admissible; State v. Gray, 221 Conn. 713, 725, 607 A.2d 391, cert. denied, 506 U.S. 872, 113 S.Ct. 207, 121 L.Ed.2d 148 (1992); "`[s]o long as the evidence may reasonably be construed in such a manner that it would be relevant. . . .'" Id.

These concepts are as applicable to scientific testimony as to other types of evidence. Thus, questions about the methodological validity of proffered scientific testimony will generally go to the weight of such evidence, not to its admissibility. Courts should exclude scientific evidence, however, when such concerns render the technique, and the resulting evidence, incapable of assisting the fact finder in a sufficiently meaningful way. Moreover, in light of the traditional policy regarding the admission of relevant evidence, "[a] judge frequently should find and expert's methodology helpful [and thus admissible] even when the judge thinks that the expert's technique has flaws sufficient to render the [expert's] conclusions inaccurate. He or she will often still believe that hearing the expert's testimony and assessing its flaws was an important part of assessing what conclusion was correct and may certainly still believe that a jury attempting to reach an accurate result should consider the evidence." In re Paoli R. Yard PCB Litigation, supra, 35 F.3d 744-45.

Once the validity of a scientific principle has been satisfactorily established, any remaining questions regarding the manner in which that technique was applied in a particular case is generally an issue of fact that goes to weight, and not admissibility. See, e.g., United States v. Porter, 618 A.2d 629, 636 (D.C.App. 1992) ("any failure by the scientists to adhere to the appropriate procedure is, of course, a proper subject of inquiry," but does not bar admission of evidence); Taylor v. State, supra, 889 P.2d 330 n. 46 ("allegations of error in the testing process affect the weight of scientific . . . evidence but not its admissibility"). Of course, even where questions about scientific evidence arise from concerns regarding application of a methodology, rather than the validity of the methodology itself, such evidence may still be excluded for failure to satisfy other of the rules of evidence.
Moreover, application of a methodology that is valid in the abstract can be so flawed that, in essence, a different, invalid methodology is being applied. Thus, although "[n]ot every error in the application of a particular methodology should warrant exclusion [a]n alleged error in the application of a reliable methodology should provide the basis for exclusion of the opinion [when] that error negates the basis for the reliability of the principle itself." United States v. Martinez, 3 F.3d 1191, 1198 (8th Cir. 1993).
For example, in In re Paoli R. Yard PCB Litigation, supra, 35 F.3d 717, several plaintiffs claimed that the defendant's use of polychlorinated biphenyl (PCB) in its railcar maintenance facility had caused various physical ailments. One witness for the plaintiffs was a physician who reviewed the medical history of all of the plaintiffs, but only personally examined two of them. Id., 764. After a thorough review of the physician's diagnostic technique, the Court of Appeals concluded that her basic methodology was valid. Id., 764-65. The court found, however, that this technique could not be properly applied at all in those cases where the physician did not personally examine a particular plaintiff, or with regard to particular symptoms of the plaintiffs she did examine when her inquiry into those symptoms was not sufficiently broad. Id., 764-70. Accordingly, the Court of Appeals concluded that the District Court had properly excluded much of the physician's testimony, even though the underlying methodology was valid, because the application of the methodology with regard to many of the plaintiffs was sufficiently flawed as to eliminate completely its reliability. Id.
It is important to note, however, the Court of Appeals' conclusion that, in those instances where the physician's inquiry into a patient's condition had been sufficient, it was an abuse of discretion for the trial court to excluded her testimony, Id., 765-67. Even though the physician was of "dubious expertise" id., 765; such concerns did not render her underlying methodology per se invalid, and once that validity had been established, concerns about her qualifications went to her credibility, not to the admissibility of her testimony.

A trial judge should therefore deem scientific evidence inadmissible only when the methodology underlying such evidence is sufficiently invalid to render the evidence incapable of helping the fact finder determine a fact in dispute. We adopt the Daubert approach, however, specifically because we conclude that a sufficient showing of validity is necessary for scientific evidence to be helpful. See part I B 1 of this opinion. The interplay between these principles — a general policy in favor of admission of helpful evidence, and a specific policy of requiring a showing of a certain level of validity before scientific testimony can properly be presented to a fact finder — cannot be resolved by an absolute statement or rule. instead, a case-by-case analysis will be necessary.

D

It is important to remember that Daubert only provides a threshold inquiry into the admissibility of scientific evidence. Even evidence that has met the Daubert inquiry into its methodological validity, and thus has been shown to have some probative value, may be excluded for failure to satisfy other evidentiary rules. in particular, scientific evidence, like all evidence, is properly excluded if its prejudicial impact outweighs its probative value, even if it is otherwise admissible. See, e.g., Daubert, v. Merrell Dow Pharmaceuticals, Inc., supra, 509 U.S. 595; State v. Figueroa, 235 Conn. 145, 162, 665 A.2d 63 (1995); C. Tait J. LaPlante, supra, § 8.1.3, p. 227.

E

Finally, we address the concern that, by requiring trial courts to conduct a Daubert validity assessment, we are improperly requiring them to become amateur scientists, According to some critics, a significant problem with the Daubert approach is that "we sorely underestimate the complexity of may scientific controversies, particularly those involving complex quantitative analysis, when we presume that non-scientist judges can master the technical issues to the point that they should feel comfortable deciding what is or is not good science in a particular case. Scientists who have spent the greater portions of their professional lives wrestling with the complexities and mysteries of their disciplines must be amazed at the law's hubris in thinking that non-scientist judges can `get up to speed' on a scientific dispute and ultimately decide who has the better of the argument." P. Milich, "Controversial Science in the Courtroom: Daubert and the Law's Hubris," 43 Emory L.J. 913, 919 (1994); see also Daubert v. Merrell Dow Pharmaceuticals, Inc., supra, 509 U.S. 600-601 (Rehnquist, C.J., concurring in part and dissenting in part); J. Kesan, supra, 84 Geo. L.J. 2000.

Even under Frye, however, "general acceptance," and thus scientific validity, was often determined by reference to acceptance among judges, not scientists. "Contrary to the assertion . . . that the Frye test places the responsibility of determining scientific validity upon scientists, in practice too many courts reference reported case law to determine what is generally accepted in the scientific community." State v. Alberico, supra, 861 P.2d 203.

Indeed, this concern motivated the Court of Appeals for the Ninth Circuit, on remand in Daubert, to offer the following comment regarding the new Daubert standard: "[T]hough we are largely untrained in science and certainly no match for any of the witnesses whose testimony we are reviewing, it is our responsibility to determine whether those experts' proposed testimony amounts to `scientific knowledge,' constitutes `good science,' and was `derived by the scientific method.' . . . Our responsibility, then . . . is to resolve disputes among respected, well-credentialed scientists about matters squarely within their expertise, in areas where there is no scientific consensus as to what is and what is not `good science,' and occasionally to reject such expert testimony because it was not `derived by the scientific method.' Mindful of our position in the hierarchy of the federal judiciary, we take a deep breath and proceed with this heady task." Daubert v. Merrell Dow Pharmaceuticals, Inc., supra, 43 F.3d 1316.

We conclude that this concern is unfounded. Under Daubert, trial judges are not required to make a determination of the ultimate scientific validity of any scientific propositions. Instead, they need only make a much more limited inquiry: whether sufficient indicia of legitimacy exist to support the conclusion that evidence derived from the principle may be profitably considered by a fact finder at trial. It is true that answering even this question may require "a deeper and more detailed preliminary review of scientific claims than most courts have heretofore undertaken." B. Black, F. Ayala C. Saffran-Brinks, supra, 72 Tex. L. Rev. 721. Nonetheless, we do not expect this process to burden our courts unduly. "Though the details of science may be remote from common experience, non-scientists can understand the fundamental characteristics that separate valid science from pale imitations." Id., 720; see also E.I. du Pont de Nemours Co. v. Robinson, supra, 923 S.W.2d 557-58.

Moreover, to the extent that our adoption of Daubert does "[signal] that the time has come for courts and lawyers to learn the basic principles of science"; L. Loevinger, supra, 35 Jurimetrics J. 179; we see this as an unavoidable necessity. As science and technology have advanced and become increasingly prevalent in our society, the number of cases, both civil and criminal, in which scientific testimony plays a role has also grown. See, e.g., id., 172 n. 110. This number undoubtedly will continue to grow in the future. When such cases, with their attendant evidentiary issues, arise, it is indisputable that "[j]udges will . . . face difficult issues that are beyond their background and training. But they should not respond by evading their responsibility to reach a carefully reasoned decision or by using superficial labels `to justify [their] instinctive reaction.' Instead, judges should try to educate themselves about the proposed area of expertise. . . ." "Developments in the Law," supra, 108 Harv. L. Rev. 1517. Only by being knowledgeable, in at least a basic way, about the issues surrounding the scientific evidence before them, can judges discharge their duties properly. Accordingly, Daubert, at its most fundamental level, merely directs "trial judges consciously [to] do what is in reality a basic task of a trial judge — ensure the reliability and relevance of evidence without causing confusion, prejudice or mistake." "Standards and Procedures for Determining the Admissibility of Expert Evidence After Daubert," supra, 157 F.R.D. 577.

II Admissibility of Polygraph Evidence

We now turn to the defendant's claim that Connecticut should abandon its traditional per se rule against the admission of polygraph evidence at trial. "This court has repeatedly held that neither the results of a polygraph test nor the willingness of a witness to take such a test is admissible in Connecticut courts." State v. Esposito, 235 Conn. 802, 831, 670 A.2d 301 (1996). The defendant argues that this position is inconsistent with Daubert.

Most courts agree that evidence of both the results of a polygraph test and of a witness' willingness to take a polygraph test implicate the same concerns and should be treated alike. See State v. Webber, 260 Kan. 263, 918 P.2d 609, 620 (1996). Admittedly, even if the polygraph test itself lacks evidentiary value, an individual's willingness to undergo such a test might have some probative value so long as the subject believed in the test's efficacy. Such willingness may also indicate, however, that the witness knows about the test's weaknesses or is simply willing to take his or her chances. We conclude that a witness' willingness to take a test of questionable validity is itself of limited probative value. Accordingly, all references in this opinion to "polygraph evidence" refer to evidence of both the substantive results of polygraph tests and of a witness' willingness to take such tests.

Without deciding, we will assume, for the purposes of this opinion, that polygraph evidence satisfies the admissibility threshold established by Daubert. After reviewing the case law and the current, extensive literature on the polygraph test, however, we are convinced that the prejudicial impact of polygraph evidence greatly exceeds its probative value. Accordingly, we see no reason to abandon our well established rule of exclusion, and we conclude that polygraph evidence should remain per se inadmissible in all trial court proceedings in which the rules of evidence apply, and for all trial purposes, in Connecticut courts.

Indeed, concerns regarding the polygraph's prejudicial impact have historically been the primary basis upon which this court has excluded polygraph evidence. It is true that we have mentioned the Frye test in excluding polygraph evidence. See State v. Hasan, supra, 205 Conn. 489; State v. Miller, 202 Conn. 463, 484, 522 A.2d 463 (1987). A review of our cases excluding polygraph evidence, however, reveals that the polygraph's unreliability and prejudicial impact, and not reliance on the Frye standard, were the principal motivating factors behind the decisions. See, e.g., State v. Mitchell, 169 Conn. 161, 170, 362 A.2d 808 (1975) (polygraph evidence inadmissible because unreliable, and because "[c]redibility as an issue is committed to the sole determination of the trier of fact"); see also State v. Esposito, supra, 235 Conn. 831 (polygraph evidence inadmissible because of "[t]he questionable accuracy of polygraph tests"); State v. Duntz, 223 Conn. 207, 238, 613 A.2d 224 (1992) (polygraph evidence inadmissible "[d]ue to the questionable accuracy of the results of polygraph examinations"); State v. Plourde, 208 Conn. 455, 471, 545 A.2d 1071 (1988), cert. denied, 488 U.S. 1034, 109 S.Ct. 847, 102 L.Ed.2d 979 (1989) (polygraph evidence has "minimal probative value"); State v. Miller, supra, 486 (polygraph evidence of "`questionable accuracy'"); State v. Saia, 172 Conn. 37, 42, 372 A.2d 144 (1976) (polygraph evidence not accepted as reliable and accurate means of ascertaining truth).

Although most of the literature and case law address polygraph evidence admissibility in criminal cases, we note the finding by the United States Office of Technology Assessment, uncontroverted in any literature, that "meaningful scientific evidence of polygraph validity [exists] only in the area of criminal investigations." (Emphasis added.) United States Congress, Office of Technology Assessment, "Scientific Validity of Polygraph Testing: A Review and Evaluation — A Technical Memorandum," OTA-TM-H-15 (1983), reprinted in 12 Polygraph 198, 200 (1983). In light of this fact, and given our ultimate determination that polygraph evidence is of limited probative value even in the criminal context, we conclude that polygraph evidence should be excluded from civil as well as criminal trials. Cf. State v. Brown, supra, 297 Or. 445 (polygraph evidence inadmissible in any civil or criminal trial or any other proceeding subject to rules of evidence).

Some jurisdictions that bar the admission of polygraph evidence for the substantive truth of the matter asserted do allow it to corroborate or impeach a witness' testimony. See part II C of this opinion. We see no reason for this distinction. All of the valid criticisms of polygraph evidence; see part II B of this opinion; are equally applicable in either context. We therefore agree with the majority of jurisdictions that maintain a per se rule against the admissibility of polygraph evidence in any circumstance. See part II C of this opinion.

A

We first explain our methodology in assessing, in the absence of an evidentiary hearing before the trial court, the likely probative value of polygraph evidence. In conducting such an assessment, an appellate court "may take judicial notice of the existence of a body of scientific literature." Browning-Ferris Industries of South Jersey, Inc. v. Muszynski, 899 F.2d 151, 161 (2d Cir. 1990). To ensure consistency in the approach to scientific evidence, a court should examine the foundation evidence received, if any; the scientific literature; and other courts' analyses. State v. Streich, 163 Vt. 331, 343, 658 A.2d 38 (1995). Indeed, even when, as in this case, there has been no evidence introduced at the trial level, an appellate court may properly "analy[ze] . . . the issues . . . based [only] on consideration of the information gleaned from prior reported cases and published literature on the subject matter." State v. Foret, supra, 628 So.2d 1131. In the past, this court has taken notice of empirical, extralegal materials in determining that, as a matter of law, certain scientific evidence was, in the form in which it was offered, inadmissible at trial. See State v. Skipper, 228 Conn. 610, 622-23, 637 A.2d 1101 (1994) (literature search revealed fact, not elicited at trial, that DNA test for determining paternity relied on mathematical formula that assumed prior probability of paternity inconsistent with presumption of innocence in criminal case); cf. State v. Sivri, supra, 231 Conn. 158-61 (directing trial court to take notice of new scientific report, issued while case on appeal, in making determination of validity and admissibility of DNA population frequency calculations).

B

With this background in mind, we turn now to an assessment of the threshold validity, probative value, and prejudicial impact of polygraph evidence. In order to do so, it is necessary to understand some of the mechanics and theory behind the modern polygraph test.

1

Modern polygraph theory rests on two assumptions: (1) there is a regular relationship between deception and certain emotional states; and (2) there is a regular relationship between those emotional states and certain physiological changes in the body that can be measured and recorded. J. Tarantino, Strategic Use of Scientific Evidence (1988) § 6.01, p. 205. These physiological changes include fluctuations in heart rate and blood pressure, rate of breathing, and flow of electrical current through the body, and they are measured by a cardiosphygmograph, a pneumograph and a galvanometer, respectively. Id. These instruments, bundled together, form the basis of most modern polygraphs.

There is no question that a high quality polygraph is capable of accurately measuring the relevant physical characteristics. United States Congress, Office of Technology Assessment, "Scientific Validity of Polygraph Testing: A Review and Evaluation — A Technical Memorandum," OTA-TM-H-15 (1983) (OTA Memorandum), reprinted in 12 Polygraph 198, 201 (1983). Even polygraph advocates, however, acknowledge that "[n]o known physiological response or pattern of responses is unique to deception." D. Raskin, "The Polygraph in 1986: Scientific, Professional and Legal Issues Surrounding Application and Acceptance of Polygraph Evidence," 1986 Utah L. Rev. 29, 31 (1986). Indeed, "there is no reason to believe that lying produces distinctive physiological changes that characterize it and only it. . . . [T]here is no set of responses — physiological or otherwise — that humans omit only when lying or that they produce only when telling the truth. . . . No doubt when we tell a lie many of us experience an inner turmoil, but we experience similar turmoil when we are falsely accused of a crime, when we are anxious about having to defend ourselves against accusations, when we are questioned about sensitive topics — and, for that matter, when we are elated or otherwise emotionally stirred." (Citation omitted.) B. Kleinmuntz J. Szucko, "On the Fallibility of Lie Detection," 17 Law Society Rev. 85, 87 (1982). Thus, while a polygraph machine can accurately gauge a subject's physiological profile, it cannot, on its own, determine the nature of the underlying psychological profile. "The instrument cannot itself detect deception." OTA Memorandum, supra, reprinted in 12 Polygraph 196, statement of John Gibbons, Director of Office of Technology Assessment.

To paraphrase Judge Cardozo in People v. Zackowitz, 254 N.Y. 192, 195, 172 N.E. 466 (1930), a lie detector "records with graphic certainty the fluctuations of the pulse. There is no instrument yet invented that records with equal certainty the fluctuations of the mind."

The polygraph examiner, therefore, is responsible for transforming the output of a polygraph machine from physiological data into an assessment of truth or deception. See, e.g., P. Giannelli, "Forensic Science: Polygraph Evidence: Part I," 30 Crim. L. Bull. 262, 264 (1994). This mission actually involves two separate tasks. First, the examiner must design and implement a polygraph test in such a way that the physiological data produced is properly linked to a subject's deceptiveness, and not just to his nervousness or other unrelated emotional responses. Id., 263. Second, even if the data produced is linked to a subject's deception, the examiner must interpret the data, that is, grade the test, correctly. Id., 264.

The "control question test" is the polygraph method most commonly used in criminal cases to link physiological responses to deception. See, e.g., C. Honts M. Perry, "Polygraph Admissibility: Changes and Challenges," 16 Law Hum. Behav. 357, 360 (1992). The control question test is based on the theory that fear of detection causes psychological stress. Under that test, therefore, the "polygraph instrument is measuring the fear of detection rather than deception per se." (Emphasis added.) OTA Memorandum, supra, reprinted in 12 Polygraph 201.

The record does not specify the type of polygraph examination administered in this case. Such facts as are in the record, however, suggest that it was either a control question test or a relevant-irrelevant test. See footnote 43. Because the former test is both the more likely and the more favorable to the defendant's claim, we will assume that a control question test was administered in this case.

A number of theories other than fear of detection have been advanced to explain why a deceptive subject will undergo detectable physiological changes upon lying, including conditioned response theory, conflict theory, and threat-of-punishment theory. P. Giannelli, supra, 30 Crim. L. Bull. 263 n. 12. The fear of detection theory is simply the most commonly accepted.

In the control question test procedure, the polygrapher first conducts a pretest interview with the subject wherein the accuracy and reliability of the polygraph are emphasized. This is done to aggravate the deceptive subject's fear of detection while calming the innocent subject, which is crucial given that the test's efficacy is based entirely on the subject's emotional state. All exam questions are then reviewed with the subject, in order to minimize the impact of surprise on the test results and to ensure that the subject understands the questions. The actual control question test consists of a sequence of ten to twelve questions, repeated several times. There are three categories of questions: neutral; relevant; and control. All questions are formulated by the polygrapher conducting the examination based on a review of the facts of the case.

During the pretest interview, a "stimulation test" is often administered. "In this test, the examiner instructs the subject to choose a particular number or name from a preselected series. The polygraph examiner does not know which number or name the subject chooses. The polygraph examiner then instructs the subject to respond `no' every time he is asked whether he chose a specific number or name. The qualified examiner should be able to determine which response is deceptive by evaluating the physical responses detected by the polygraph. The examiner then confronts the subject with his findings, causing the subject to be `convinced' that the polygraph examination is effective." J. Tarantino, supra, § 6.11, p. 218. Occasionally, however, the correct response is determined by trickery, not polygraphy. J. Rat F. Inbau, Truth and Deception (2d Ed. 1977) pp. 42, 85 (stimulation test conducted using marked deck of cards).

A neutral question is entirely nonconfrontational and is designed to allow the polygrapher to get a baseline reading on the subject's physiological responses. A neutral question addresses a subject's name, age, address, or similar topic.

A relevant question is accusatory and directed specifically at the subject under investigation. "For example, in an assault investigation, a relevant question might be: `On May 1, 1986, did you strike Mr. Jones (the alleged victim) with any part of your body?'" J. Tarantino, supra, § 6.09, p. 215.

A control question concerns "an act of wrongdoing of the same general nature as the main incident under investigation," and is designed to be "one to which the subject, in all probability, will lie or to which his answer will be of dubious validity in his own mind." J. Rat F. Inbau, Truth and Deception (2d Ed. 1977) p. 28. Control questions "cover many years in the prior life of the subject and are deliberately vague. Almost anyone would have difficulty answering them truthfully with a simple `No.'" D. Raskin, supra, 1986 Utah L. Rev. 34. In an assault case, a control question might be: "Did you ever want to see anyone harmed?" J. Tarantino, supra, § 6.09, p. 215. Although few people honestly could deny these control questions categorically, they are "presented to the subject in a manner designed to lead him to believe that admissions would negatively influence the examiner's opinion and that strong reactions to those questions during the test would produce a deceptive result." D. Raskin, supra, 1986 Utah L. Rev. 34.

In other words, the control question test process requires that the examiner, during the pretest interview, manipulate the subject into both (1) lying on the control questions, out of fear that the examiner will otherwise react negatively to the subject's prior antisocial conduct, and (2) fearing that this same deception will taint the entire exam.

The theory behind the control question test is that "the truthful person will respond more to the control questions than to the relevant questions because they represent a greater threat to that person. For the same reason the deceptive person will respond more to the relevant questions than to the control questions." P. Giannelli, supra, 30 Crim. L. Bull. 266-67. Thus, in order for the test to work properly, both truthful and deceptive examinees must have particular mind sets during the exam. "The innocent examinee [must fear] that the polygraph examiner will pick up his deception [on the control question] and incorrectly conclude that he is also being deceptive about the relevant question." J. Tarantino, supra, § 6.09, p. 216. As a result, the innocent subject's physiological responses to the control question, stemming from this fear, will be greater than those to the relevant question, which the subject can answer honestly. A guilty subject, however, will be more worried about having his crime and deception exposed by the relevant question than he is about any control question issues. Accordingly, his physiological responses — prompted by his fear of detection — will be greater with regard to the relevant question than to the control question.

Under the control question test, the absolute measure of the subject's physiological responses to each question is unimportant. For example, the mere fact that a subject has a strong response to a relevant question can simply be indicative of nervousness and does not, by itself, indicate deception. Instead, the polygrapher looks to the relative strength of the responses to the control and relevant questions in order to determine truth or deception. The art of the polygrapher lies in composing control and relevant questions that elicit the appropriate relative responses from truthful and deceitful parties. See generally United States v. Kwong, 69 F.3d 663, 668 (2d Cir. 1995) (polygraph results excluded because examiner failed to formulate questions properly); J. McCall, supra, 1996 U. Ill. L. Rev. 378; D. Raskin, supra, 1986 Utah L. Rev. 47-49.

This process is what separates the control question test from the relevant-irrelevant polygraph test. In the latter test, there are no control questions, only relevant and irrelevant questions. A subject whose responses to the relevant questions are greater than those to the irrelevant questions is considered deceptive. Although still practiced by some polygraphers today; P. Giannelli, supra, 30 Crim. L. Bull. 266; the relevant-irrelevant test is almost universally rejected in the literature. D. Raskin, supra, 1986 Utah L. Rev. 33 ("A variety of factors might cause individuals to react more strongly to questions about crimes of which they are accused than to innocuous questions. . . . There is no clear and systematic way to interpret the outcome of a relevant-irrelevant test, and the result is subject to a great deal of error."); C. Honts M. Perry, supra, 16 Law Hum. Behav. 359 ("Almost all of the scientists involved in detection of deception research reject the notion that the relevant-irrelevant test could be a useful discriminator of truth and deception. . . . [M]ost individuals will respond physiologically to the relevant questions regardless of their guilt." [Citations omitted.]).

A control question exam ordinarily pairs relevant and control questions with some neutral questions interspersed. For example, a typical progression would be:

"1. (Neutral) Do you understand that I will ask only the questions we have discussed?

"2. (Pseudo-Relevant) Regarding whether you took that ring, do you intend to answer all of the questions truthfully?

"3. (Neutral) Do you live in the United States?

"4. (Control) During the first twenty-four years of your life, did you ever take something that did not belong to you?

"5. (Relevant) Did you take a ring from the Behavioral Sciences Building on July 1, 1985?

"6. (Neutral) Is your name Joanne?

"7. (Control) Between the ages of ten and twenty-four, did you ever do anything dishonest or illegal?

"8. (Relevant) Did you take that diamond ring from a desk in the Behavioral Sciences Building on July 1?

"9. (Neutral) Were you born in the month of February?

"10. (Control) Before 1984 did you ever lie to get out of trouble or to cause a problem for someone else?

"11. (Relevant) Were you in any way involved in the theft of that diamond ring from the Behavioral Sciences Building last July?" D. Raskin, supra, 1986 Utah L. Rev. 36. The entire sequence is normally gone through three times, after which the examiner scores the result to attempt to reach a determination of truthfulness or deception.

The most common technique for scoring polygraph charts is pure numerical grading. In the most prevalent numerical system, the polygrapher assigns a numerical value along the range of -3 to + 3 to each pair of relevant and control questions. A score of + 3 indicates a much stronger reaction to the control question than to the relevant question and, therefore, truthfulness; a score of -3 indicates a much stronger reaction to the relevant question and, therefore, deception; and a score of 0 indicates that there was no significant difference in response. The examiner considers only the polygraph chart in assigning these scores; no consideration is given to any subjective impressions regarding the subject's truthfulness that the examiner develops over the course of the exam. The scores for all question pairs in all three sequences are then totaled. If the sum is + 6 or greater, the subject is classified as truthful; if the sum is -6 or lower, the subject is classified as deceptive; scores of -5 to + 5 are deemed inconclusive. Computers are sometimes used to give more precise numerical scores to polygraph charts.

We note that there is disagreement in the literature even as to what method of scoring is proper. Most authorities agree that numerical scoring is the only valid approach. Nonetheless, many examiners still use "global scoring." See, e.g., D. Lykken, "The Validity of Tests: Caveat Emptor," 27 Jurimetrics J. 263, 264-65 (1987); D. Raskin, supra, 1986 Utah L. Rev. 37. Under this approach, the examiner looks at the subject's relative reactions to the control and relevant questions, but also considers various "clinical impressions," such as the subject's demeanor, in arriving at a conclusion of truthfulness or deception. D. Raskin, supra, 1986 Utah L. Rev. 37. A few authors, albeit generally polygraph detractors, suggest that the global method is, in fact, more accurate. See, e.g., D. Carroll, "How Accurate Is Polygraph Lie Detection?," in The Polygraph Test (A. Gale ed., 1988) pp. 19, 27.

If an analysis of the first three charts produces inconclusive results, the examiner will often repeat the question sequence twice more. After that, however, further repetitions are generally considered meritless, as the subject will have become habituated to the test questions and, therefore, will no longer have sufficiently strong emotional responses for polygraph purposes. Id., 37-40.

The other main type of polygraph examination used in criminal matters is the guilty knowledge test. The guilty knowledge test "does not attempt to determine whether the [subject] is lying but, rather, whether he or she possesses guilty knowledge, that is, whether the [subject] recognizes the correct answers, from among several equally plausible but incorrect alternatives, to certain questions relating to a crime. For example, escaping through an alley a bank robber drops and leaves behind his hat. A likely suspect is later apprehended and, while attached to the polygraph, he is interrogated as follows:
"1. `The robber in this case dropped something while escaping. If you are that robber, you will know what he dropped. Was it: a weapon? a face mask? a sack of money? his hat? his car keys?' . . .
"Unlike the control question test, the accuracy of the guilty knowledge test does not depend upon the nature or degree of the subject's emotional concern. The physiological variables employed are not intended to measure emotional response but, rather, to signal the cognitive processes involved in the recognition of the correct alternative." D. Lykken, "The Case Against Polygraph Testing," in The Polygraph Test (A. Gale ed., 1988) pp. 111, 121-23.
"The guilty knowledge test assumes that the guilty subject will have a greater physical response to the `significant alternative' than would a subject without any guilty knowledge." J. Tarantino, supra, § 6.13, p. 219. Advocates claim that the primary advantage of the guilty knowledge test is that recognition can be more directly measured by physiological data than can truth or deception. D. Lykken, "The Case Against Polygraph Testing," supra, pp. 121-23.
For the guilty knowledge test to work, however, there must be "concealed knowledge" that only the guilty party would know and recognize. This requirement greatly limits the number of cases in which the test can be utilized. P. Giannelli, supra, 30 Crim. L. Bull. 266. In any event, although the guilty knowledge test does have its advocates; D. Lykken, "The Case Against Polygraph Testing," supra, pp. 121-24; B. Kleinmuntz J. Szucko, supra, 17 Law Society Rev. 98; the guilty knowledge test's validity is as hotly debated as that of the control question test. C. Honts M. Perry, supra, 16 Law Hum. Behav. 359; D. Raskin, supra, 1986 Utah L. Rev. 31-32 n. 12 (many guilty people can pass guilty knowledge test). Because the validity of the guilty knowledge test is so uncertain, and because all of the prejudicial effects of allowing control question test evidence apply to guilty knowledge test evidence as well; see part II B 2 b of this opinion; we conclude that guilty knowledge test evidence must also be excluded from use in our courts.

2

We now examine the validity of the results produced by the polygraph test. The word "validity" has two meanings in the polygraph context: for the purposes of this discussion, they will be labeled "accuracy" and "predictive value." Courts generally do not specify to which concept they are referring when they address polygraph issues. Maintaining this distinction is essential, however, if one is to evaluate fairly the validity of the polygraph test.

Although courts generally use the word "reliability" when discussing the polygraph test; see, e.g., United States v. Crumby, 895 F. Sup. 1354, 1358 (D.Ariz. 1995) ("[t]he Court must consider whether polygraph evidence is reliable under Daubert" [emphasis added]); the concept to which the courts are referring is actually the test's "validity." In the polygraph context, reliability and validity have specialized meanings. Reliability refers only to reproducibility of results, or consistency, while validity relates to the test's actual ability to do what it claims to do, namely, detect deception. 1 C. McCormick, supra, § 206, p. 909 n. 28. Reliability is important, but the polygraph debate really centers around the test's validity. See generally S. Blinkhorn, "Lie Detection As a Psychometric Procedure," in The Polygraph Test (A. Gale ed., 1988) pp. 29, 31-35.

a

The "accuracy" of the polygraph test itself has two components: sensitivity and specificity. The polygraph's sensitivity is its ability to tell that a guilty person is, in fact, lying. If the polygraph test had a 90 percent sensitivity, then it would correctly label a deceptive subject as being deceptive 90 percent of the time. Thus, the test would incorrectly label a deceptive subject as being truthful 10 percent of the time; this mislabeling is called a "false negative" error. The polygraph's specificity is its ability to tell that an innocent person is, in fact, being truthful. If the polygraph test had an 80 percent specificity, then it would label a truthful subject as being truthful 80 percent of the time. The test would thus incorrectly label a truthful subject as being deceptive 20 percent of the time; this mislabeling is called a "false positive" error. It is generally agreed in the literature, by both advocates and critics, that polygraphs have greater sensitivity than specificity; that is, that false positives outnumber false negatives. See, e.g., C. Honts M. Perry, supra, 16 Law Hum. Behav. 362; see also United States v. Galbreth, 908 F. Sup. 877, 885 (D.N.M. 1995).

There is wide disagreement, however, as to what the sensitivity and specificity values actually are for a well run polygraph exam. See generally 1 C. McCormick, supra, § 206, pp. 907-17. Dozens of studies of polygraph accuracy have been conducted. Id. They fall into two basic types, namely, laboratory simulations of crimes and field studies based on data from polygraph examinations in actual criminal cases. P. Giannelli, supra, 30 Crim. L. Bull. 270-73. The variance in expert opinion regarding polygraph accuracy arises from disagreements as to which methods and which studies within each method are methodologically valid. C. Honts M. Perry, supra, 16 Law Hum. Behav. 360.

"The most accepted type of laboratory study simulates a real crime in which subjects are randomly assigned to guilty and innocent treatment conditions. . . . Guilty subjects enact a realistic crime, and innocent subjects are merely told about the nature of the crime and do not enact it. All subjects are motivated to produce a truthful outcome, usually by a substantial cash bonus for passing the test." (Citations omitted.) D. Raskin, "Does Science Support Polygraph Testing?," in The Polygraph Test (A. Gale ed., 1988) pp. 96, 99.

"The best available method for field research uses cases in which suspects were administered polygraph tests after which their guilt or innocence was established when the guilty person confessed. Other polygraph examiners are then asked to make diagnoses based solely on the polygraph charts from those tests without knowledge of the guilt or innocence of the subjects or the opinions of the original examiners. The decisions from these blind analyses are then compared to the confession criterion to estimate the accuracy of the polygraph tests." D. Raskin, supra, 1986 Utah L. Rev. 44.

Polygraph supporters base their accuracy estimates on both laboratory simulation and field studies. These advocates acknowledge that field studies are theoretically preferable for establishing the polygraph test's field accuracy, but they conclude that serious methodological difficulties inherent in such studies, such as establishing the actual guilt or innocence of the study subjects, make most of these studies unreliable. They think, however, that laboratory studies, when designed to approximate field conditions and when carefully conducted, can provide useful and valid data. See generally J. Kircher, S. Horowitz D. Raskin, "Meta-analysis of Mock Crime Studies of the Control Question Polygraph Technique," 12 Law Hum. Behav. 79, 80 (1988); see also C. Honts M. Perry, supra, 16 Law Hum. Behav. 361. David Raskin, perhaps the foremost polygraph advocate in the United States, recently reviewed the literature on polygraph studies and concluded that eight laboratory studies and four field studies of the control question test polygraph technique were methodologically valid. D. Raskin, "The Scientific Status of Research on Polygraph Techniques," in West Companion to Scientific Evidence 2 (Faigman et al. eds., forthcoming 1996), cited in C. Honts B. Quick, "The Polygraph in 1995: Progress in Science and the Law," 71 N.D. L. Rev. 987, 995, 1018-19 (1995). The laboratory studies that Raskin cites, taken together, indicate that the polygraph test has an 89 percent sensitivity rate and a 91 percent specificity rate; the field studies give an 87 percent sensitivity and a 59 percent specificity. Id. Other studies indicate higher levels of accuracy. See generally P. Giannelli, supra, 30 Crim. L. Bull. 271-73; C. McCormick, supra, § 206, pp. 909-10. The United States Department of Defense, although acknowledging that more research needs to be done, concluded after a thorough review of the literature that there was no "data suggesting that the various polygraph techniques and applications . . . have high false positive or high false negative error rates." United States Department of Defense, The Accuracy and Utility of Polygraph Testing (1984) p. 63.

It is interesting that the recent laboratory studies cited here by Raskin as valid give lower aggregate rates for polygraph sensitivity and specificity than did those studies deemed reliable by him in an article he wrote ten years ago. D. Raskin, supra, 1986 Utah L. Rev. 43 (concluding that polygraph has 97 percent sensitivity and 92 percent specificity). It would appear that as testing has advanced, the validity of the polygraph has become more questionable.

Pursuant to standard practice in calculating specificity and sensitivity, we exclude all of the inconclusive outcomes in the raw data from our calculations, because inconclusive results are not conclusions. See, e.g., C. Honts B. Quick, supra, N.D. L. Rev. 996 n. 65.

Critics, however, view the existing body of polygraph studies quite differently. First, although polygraph detractors agree with the advocates that most field studies are invalid due to methodological concerns, they disagree as to which tests are valid. David Lykken, a prominent polygraph critic, has concluded from the field tests he deems valid that the polygraph has a sensitivity of 84 percent and a specificity of only 53 percent. D. Lykken, "The Validity of Tests: Caveat Emptor," 27 Jurimetrics J. 263, 264 (1987). Another critic has concluded that reliable field studies indicate that there is "little or no case" for using the polygraph, and that "polygraph lie detection adds nothing positive to conventional approaches to interrogation and assessment." D. Carroll, "How Accurate Is Polygraph Lie Detection?," in The Polygraph Test (A. Gale ed., 1988) pp. 19, 28. After its own thorough review of the polygraph field studies, the United States Office of Technology Assessment concluded that "the cumulative research evidence suggest that . . . the polygraph test detects deception better than chance, but with significant error rates." OTA Memorandum, supra, reprinted in 12 Polygraph 200.

Lykken admittedly wrote this assessment before any of the recent field studies on which Raskin relies were published. Nonetheless, there is inarguably a fundamental schism in the type of field study that Raskin and Lykken will regard as valid. For example, the three studies that Lykken deems "scientifically credible"; D. Lykken, supra, 27 Jurimetrics J. 264; are specifically attacked by Raskin as not satisfying even "basic methodological requirements." D. Raskin J. Kircher, "The Validity of Lykken's Criticisms: Fact or Fancy?," 27 Jurimetrics J. 271, 272 (1987).

Moreover, polygraph critics argue that laboratory simulation studies are almost completely invalid. They point out that, although the accuracy of the control question test turns entirely on the subject having the "right" emotional responses, the emotional stimuli in the laboratory are completely different from those in the field. D. Carroll, "How Accurate Is Polygraph Lie Detection?," supra, p. 24. "In the mock crime paradigm . . . it is likely that volunteer subjects regard the experience as a kind of interesting game. Those persons instructed to commit the mock crime and to lie during the test no doubt feel a certain excitement, but not the guilt or fear of exposure that a real thief feels when tested for the police. Volunteers assigned to the innocent group have no reason at all to fear the relevant questions; they are not suspected of any wrongdoing and they will not be punished or defamed even if the test goes awry. On the other hand, the control questions used in laboratory studies . . . unlike the relevant questions, do refer to real-world events and, presumably, have the same embarrassing or disturbing effect on volunteer subjects that they have on criminal suspects. This is probably the reason why mock crime studies typically show a much lower rate of false positive errors than do studies of actual criminal interrogation in the field. Innocent suspects often fail police-administered tests . . . because they find the relevant questions more disturbing than the control questions, since they know they are in real jeopardy in respect to the accusations contained in the relevant questions while the controls involve no comparable risk. For the volunteer laboratory subject, this balance is reversed." D. Lykken, "The Case Against Polygraph Testing," in The Polygraph Test (A. Gale ed., 1988) pp. 111, 114-15. Raskin has admitted that these concerns with laboratory simulations are significant. J. Kircher, S. Horowitz D. Raskin, supra, 12 Law Hum. Behav. 80, 88-89.

Lykken describes a laboratory investigation that "succeeded in eliciting genuine concern in both the `guilty' and `innocent' examinees. [Two examiners] had forty-eight prison inmates tested in a mock crime situation. It was understood that each participant's bonus of $20 would be withheld if more than ten of the forty-eight subjects failed the polygraph test. Moreover, the names of inmates failing the test would be posted in the prison for all to see. The intent was to make both `guilty' and `innocent' subjects believe that, if they failed the test, they might be blamed by their fellow inmates for the loss of the $20, a considerable sum by prison standards. That this manipulation was successful is suggested by the fact that several inmates expressed their concern about the consequences of failing and a few actually declined to participate for that reason. Two skilled and experienced examiners administered control question tests and all charts were independently scored by both examiners. . . . [O]nly thirty of the forty-eight subjects were correctly classified. Excluding inconclusive tests, there were 13 percent false negative errors and 44 percent of the `innocent' inmates were misclassified as deceptive." D. Lykken, "The Case Against Polygraph Testing," supra, p. 115.

Even if one accepts Raskin's field study estimates of accuracy over those of the polygraph critics, polygraph evidence is of questionable validity. Raskin's 87 percent sensitivity indicates a 13 percent false negative rate. In other words, 13 percent of those who are in fact deceptive will be labeled as truthful. Moreover, Raskin's 59 percent specificity indicates a 41 percent false positive rate. In other words, 41 percent of subjects who are, in fact, truthful will be labeled as deceptive.

b

In the previous section, we demonstrated that the basic accuracy of the polygraph test is still open to considerable debate. The actual probative value of polygraph evidence as a signifier of guilt or innocence, moreover, is even more questionable. This is because sensitivity, for example, only tells how likely a polygraph is to label accurately a person as deceptive given that the person really is lying. At trial, however, we would not yet know that a subject is deceptive — indeed, making that determination may be the entire point of the trial. Knowing how accurately the polygraph test labels deceptive people as deceptive is not, therefore, directly helpful. We are instead interested in a related, but distinct, question: how likely is it that a person really is lying given that the polygraph labels the subject as deceptive? This is called the "predictive value positive." Similarly, at trial we are not directly interested in the polygraph test's specificity, but rather in its "predictive value negative": how likely is it that a subject really is truthful given that the polygraph labels the subject as not deceptive?

Predictive value positive and predictive value negative depend on the sensitivity and specificity of the polygraph test, but also turn on the "base rate" of deceptiveness among the people tested by the polygraph. Unfortunately, no reliable measure of this base rate currently exists if, indeed, one is possible at all. Raskin has claimed, on the basis of an analysis of a United States Secret Service study and on the basis of his own empirical experience, that only about 40 to 60 percent of criminal defendants who are willing to submit to polygraph tests are actually guilty. D. Raskin, supra, 1986 Utah L. Rev. 59-60. If a base rate of about 50 percent were correct, then, using Raskin's own field derived figures of 87 percent sensitivity and 59 percent specificity, the predictive value positive of the polygraph test would only be 68 percent and the predictive value negative would be 82 percent. That is, even if we were to agree with all of Raskin's figures, we should only be 68 percent confident that a subject really is lying if the subject fails a polygraph exam, and only 82 percent confident that the subject is being truthful if the subject passes. Therefore, although the probative value of the polygraph test may be greater than that of a coin toss, it is not significantly greater, especially for failed tests.

The term "base rate" refers to the prevalence of a condition among the relevant tested population. In the context of the polygraph test, the base rate is the percentage of people who submit to a polygraph exam who are, in fact, deceptive on the exam. If, out of every 100 people who take a polygraph test, we could empirically demonstrate that fifty are, in fact, giving deceptive responses, then the base rate of deception would be 50 percent.
The base rate is important because it can greatly accentuate the impact of the false positive and false negative rates arising from any given specificity and sensitivity values. If one assumes base rates progressively higher than 50 percent, then, by definition, the number of deceptive examinees increases and the number of honest examinees decreases. A logical consequence is that, even holding specificity and sensitivity rates constant, as the base rate increases the number of false negatives (the labeling of deceptive subjects as truthful) also rises and the number of false positives (the labeling of truthful subjects as deceptive) falls, because only deceptive subjects produce false negatives and only truthful subjects produce false positives. Likewise, if one were to assume base rates progressively lower than 50 percent, then, even holding sensitivity and specificity constant, as the base rate falls the number of false positives will necessarily rise and the number of false negatives will fall.
For example, a very low base rate would dramatically emphasize the problem of false positives, even if sensitivity and specificity were both relatively high. Suppose that the polygraph has a sensitivity of 90 percent (and thus a false negative rate of 10 percent) and a specificity of 80 percent (and thus a false positive rate of 20 percent), and that the base rate of deception is 10 percent. If 100 subjects are tested, then the 10 percent base rate signifies that ten subjects are deceptive and ninety are truthful. Given the specificity of 80 percent, seventy-two of the ninety truthful subjects will be labeled accurately as truthful (80 percent of ninety is seventy-two); the remaining eighteen truthful subjects will be mislabeled as deceptive due to the 20 percent false positive rate. Similarly, given the sensitivity of 90 percent, nine of the ten deceptive subjects will be labeled accurately as deceptive (90 percent of ten is nine); the remaining deceptive subject will be mislabeled as truthful due to the 10 percent false negative rate. These results can be summarized as follows:
Innocent Guilty


Pass 72 1 Fail 18 9

"A hundred people are tested: 81 percent are correctly classified; 90 percent of the guilty fail; 80 percent of the innocent pass. And yet of these who fail, only one in three is guilty." S. Blinkhorn, "Lie Detection As a Psychometric Procedure," in The Polygraph Test (A. Gale ed., 1988) pp. 29, 34.

Predictive value positive (PVP) and predictive value negative (PVN) are determined by an equation involving the polygraph test's sensitivity and specificity, and the base rate of deception among the tested population. Mathematically, the relationship among these concepts is expressed as follows:


PVP = ------------------------ ( + — 1 ) + (1 — )

and: (1 — ) PVN = --------------------- (1 — ) + (1 — )

where: = the base rate of deception among people who choose to take the polygraph exam;
= the polygraph test's sensitivity; and
= the polygraph test's specificity.
See D. Kaye, "The Validity of Tests: Caveant Omnes," 27 Jurimetrics J. 349 (1987).

The figures of 68 percent predictive value positive and 82 percent predictive value negative noted in the text of this opinion are the result of the application of the equation stated in footnote 54, using a base rate of 50 percent, a sensitivity rate of 87 percent, and a specificity rate of 59 percent.

Even if we were to use only Raskin's laboratory derived values for sensitivity and specificity in calculating predictive value positive and predictive value negative, the polygraph test would still be of questionable worth. Assuming an 89 percent sensitivity, a 91 percent specificity and a base rate of 50 percent, the predictive value positive of the polygraph test is only 91 percent and the predictive value negative is only 89 percent. In other words, approximately one out of every ten polygraph examinations would still mislabel the subject.
We emphasize, moreover, that Raskin himself has never advocated assessing the validity of the polygraph examination solely by the use of laboratory data, without reference to field studies. Indeed, for the reasons set forth by the polygraph critics, we are persuaded that it would be reversible error for a trial court to follow such a procedure. Furthermore, to the extent that the previously mentioned figures may appear impressive to some, it is important to remember that, even if one does believe that use of laboratory derived values for sensitivity and specificity is appropriate, the assumed base rate of 50 percent underlying the calculations has essentially no empirical support.

Furthermore, the 50 percent base rate that Raskin posits is far from universally accepted. "[T]he figures for [the base rate] that [Raskin] pull[s] out of the hat should not be taken as firm." D. Kaye, "The Validity of Tests: Caveant Omnes," 27 Jurimetrics J. 349, 357 (1987). Lykken posits, albeit with as equally sparse evidence as Raskin, that the base rate of guilt among people volunteering for a polygraph exam is 80 percent. D. Lykken, supra, 27 Jurimetrics J. 268. Using this base rate, the polygraph test's predictive value positive is 89 percent and its predictive value negative is 53 percent. Lykken's base rate, therefore, makes a failed test more probative than it is under Raskin's base rate, but makes a passed test much less probative.

The specific predictive value positive and predictive value negative figures generated by a particular set of assumptions, however, is not the significant point for the legal determination of whether to admit polygraph evidence. The point is that, given the complete absence of reliable data on base rates, we have no way of assessing the probative value of the polygraph test. Under one set of assumptions, a failed test has some significance, while a passed test does not; under another, the situation is reversed. The figures are further muddied when one recalls that the sensitivity and specificity of the polygraph are also hotly debated.

c

Countermeasures are also a concern with regard to polygraph validity. A countermeasure is any technique used by a deceptive subject to induce a false negative result and thereby pass the test. For a countermeasure to work on the control question test, all it must do is "change the direction of the differential reactivity between the relevant and control questions. . . ." G. Gudjonsson, "How to Defeat the Polygraph Tests," in The Polygraph Test (A. Gale ed., 1988) pp. 126, 127.

It may be true that "subjects without special training in countermeasures are unable to beat the polygraph test, even if they have been provided with extensive information and suggestions on how they might succeed. . . ." (Citation omitted; emphasis added.) D. Raskin, "Hofmann, Hypnosis, and the Polygraph," 3 Utah B.J. 7, 8 (1990); G. Gudjonsson, "How to Defeat the Polygraph Tests," supra, p. 135. Yet as one polygraph supporter, Charles Honts, concedes, "studies have indicated that [expert-conducted] training in specific point countermeasures designed to increase [physiological responses to control questions] is effective in producing a substantial number of false negative outcomes. . . ." C. Honts B. Quick, supra, 71 N.D. L. Rev. 1001. Specifically, "[s]ubjects in these studies were informed about the nature of the control question test and were trained to recognize control and relevant questions. Countermeasure subjects were then instructed to employ a countermeasure (e.g., bite their tongue, press their toes to the floor, or count backward by seven) during the control question zones of a control question test. In one study, none of the guilty subjects who received this brief training was correctly detected. . . . Across all of the studies more than 50 % of the decisions on countermeasures subjects were incorrect." (Citation omitted.) C. Honts M. Perry, supra, 16 Law Hum. Behav. 374. Although we share Honts' hope that "the required [expert] training is . . . difficult to obtain"; C. Honts B. Quick, supra, 71 N.D. L. Rev. 1001; we question whether such is the case and whether it would remain the case if polygraph examination of witnesses became common, especially given the apparent brevity and simplicity of the training in question.

But see D. Lykken, supra, 27 Jurimetrics J. 267 ("As it happens, there is a simple, easily learned technique with which a guilty person can `beat' the control question test. In one informal prison study, twenty-seven inmates accused of violating prison rules were given some fifteen minutes of instruction in this method [by a fellow inmate, based on information provided by Lykken] before reporting for a test concerned with the alleged infraction. Although all twenty-seven privately admitted their guilt, twenty-four of them managed to pass the polygraph.").

3

With the foregoing information in mind, we will assume, without deciding, that polygraph evidence satisfies Daubert. Although the subjective nature and highly questionable predictive value of the polygraph test weigh heavily against admission, we assume that polygraph evidence may have enough demonstrated validity to pass the Daubert threshold for admissibility.

We conclude, however, that admission of the polygraph test would be highly detrimental to the operation of Connecticut courts, both procedurally and substantively. Moreover, as illustrated in part II B 2 b of this opinion, the probative value of polygraph evidence is very low, even if it satisfies Daubert. Accordingly, we also conclude that any limited evidentiary value that polygraph evidence does have is substantially outweighed by its prejudicial effects. We therefore reaffirm our per se rule against the use of polygraph evidence in Connecticut courts. See State v. Brown, supra, 297 Or. 438, 439-42 (polygraph evidence properly not admissible at trial, even though it "may possess some probative value and may, in some cases, be helpful to the trier of fact," because probative value substantially outweighed by prejudicial impact); see also United States v. Black, 831 F. Sup. 120, 122-23 (E.D.N.Y. 1993) ("That the Frye test was displaced by the Rules of Evidence does not mean . . . that the Rules themselves place no limits on the admissibility of purportedly scientific evidence. Nor is the trial judge disabled from screening such evidence. . . . [N]othing in Daubert . . . disturb[s] the settled precedent that polygraph evidence is . . . [not] admissible. . . ." [Citations omitted; internal quotation marks omitted.]).

a

The most significant, and fundamental, problem with allowing polygraph evidence in court is that it would invade the fact-finding province of the jury. The jury has traditionally been the sole arbiter of witness credibility. State v. James, 237 Conn. 390, 438, 678 A.2d 1338 (1996); State v. Person, 236 Conn. 342, 347, 673 A.2d 463 (1996); see State v. Mitchell, 169 Conn. 161, 170, 362 A.2d 808 (1975) (polygraph evidence excluded because "[c]redibility as an issue is committed to the sole determination of the trier of fact"). Indeed, an underlying premise of our legal system is that the jury is capable in this regard. Accordingly, we generally disallow expert testimony as to witness credibility when the subject matter of the testimony "is within the knowledge of jurors and expert testimony [therefore] would not assist them. . . ." State v. Kemp, 199 Conn. 473, 477, 507 A.2d 1387 (1986); State v. Boscarino, 204 Conn. 714, 733, 529 A.2d 1260 (1987) (expert testimony on reliability of eyewitness identifications properly excluded).

Cf. Perkins v. State, 902 S.W.2d 88, 94 (Tex.App. 1995) ("Even though serious doubts remain about the reliability of polygraph evidence, its unreliability is not the primary reason for its exclusion under our holding. Instead, we find that such evidence should be excluded because it impermissibly decides the issues of credibility and guilt for the trier of fact and supplants the jury's function.").

The defendant suggested at oral argument that polygraph evidence would not impinge on the province of the jury at all because it would not directly address witness credibility. He proposes limiting testimony of the polygrapher to whether a subject's physiological responses are consistent with those of most people when telling the truth or lying, rather than allowing the polygrapher to testify as to an opinion of the subject's actual truthfulness or deception. This is certainly a more honest approach, but we nonetheless recognize that, in the context of trial testimony, the message conveyed to the jury by both styles of testimony would be the same. Furthermore, altering the technical format of the evidence does not alter the general balance between its probative value and prejudicial impact.

A determination of whether a witness is telling the truth is well within the province of all jurors' understanding and abilities. Indeed, one polygraph skeptic, after reviewing the literature, concluded that any observant person can assess witness truthfulness as well as a trained polygrapher: "[I]t would appear that an observer, regarding the [polygraph examinee's] general behavior . . . does just as well as an experienced polygraph examiner. . . ." D. Carroll, "How Accurate Is Polygraph Lie Detection?," supra, p. 28. Most scientific testimony is facially different from polygraph evidence, however, in that it involves matters that jurors cannot assess on their own. A juror cannot, for example, determine whether a blood stain contains DNA consistent with that of a particular individual without expert assistance and testimony.

By contrast, we allow expert testimony as to the general behavior patterns of battered women and sexual assault victims, even though such testimony also implicates witness credibility, specifically because such matters are not within the understanding and personal knowledge of the average juror. See State v. Borrelli, 227 Conn. 153, 173-74, 629 A.2d 1105 (1993) (expert may testify regarding battered women's syndrome because such testimony addressed issues "that in all likelihood were beyond the jury's experience and knowledge"); State v. Spigarolo, 210 Conn. 359, 378, 556 A.2d 112, cert. denied, 493 U.S. 933, 110 S.Ct. 322, 107 L.Ed.2d 312 (1989) (expert testimony regarding typical behavior patterns of children who are sexual assault victims "is admissible because the consequences of [this] unique trauma . . . are matters beyond the understanding of the average person").

Very few studies have been done on the influence that polygraph evidence has over juries, and the cumulative results of those studies are inconclusive. In view of the importance of maintaining the role of the jury, this uncertainty alone justifies the continued exclusion of polygraph evidence. Moreover, polygraph evidence so directly abrogates the jury's function that its admission is offensive to our tradition of trial by jury. Indeed, the specter of polygraph evidence demonstrates why the traditional role of the jury in assessing credibility is so important and should be guarded so assiduously.

The earliest reported study of the impact of polygraph evidence on a jury was in connection with a robbery prosecution in People v. Kenny, 167 Misc. 51, 3 N.Y.S.2d 348 (1938). Ten jurors responded to a mailed questionnaire regarding their reaction to the polygraph evidence submitted in that case. Five accepted the testimony of the lie detector expert "`without question,'" and six felt the lie detector testimony was conclusive proof of the defendant's innocence. Id.; S. Carlson, M. Pasano J. Jannuzzo, "The Effect of Lie Detector Evidence on Jury Deliberations: An Empirical Study," 5 J. Police Sci. Admin. 148, 149 (1977).
Of course jurors, and society as a whole, have become more technologically sophisticated over the last sixty years. Nonetheless, more recent studies, although inconclusive, do not dispel the fear that juries will give undue weight to polygraph results. Each of eight jurors interviewed after the conclusion of a federal drug case, for example, indicated that, in a close case, the polygraph tests standing by themselves would be sufficient to raise reasonable doubt. F. Barnett, "How Does a Jury View Polygraph Examination Results?," 2 Polygraph 275, 276 (1973). In another study, one quarter of the jurors in a mock trial competition indicated that they would change their votes if contrary polygraph evidence that was stated to be 95 percent accurate were presented to them, while another one quarter stated they were uncertain what they would do. S. Carlson, M. Pasano J. Jannuzzo, supra, 5 J. Police Sci. Admin. 152; see also A. Markwart B. Lynch, "The Effect of Polygraph Evidence on Mock Jury Decision-Making," 7 J. Police Sci. Admin. 324 (1979).
The most scientific study on this topic to date is reported in A. Cavoukian R. Heslegrave, "The Admissibility of Polygraph Evidence in Court: Some Empirical Findings," 4 Law Hum. Behav. 117 (1980). In one experiment discussed therein, 52 percent of mock jurors in a control group found a defendant "guilty" based on a written fact pattern, whereas only 28 percent found the defendant "guilty" when exculpatory polygraph evidence with a stated accuracy of 80 percent was added. Id., 121, 123. A cautionary statement by the judge in the fact pattern regarding the limitations of polygraph reliability somewhat mitigated the effect of polygraph evidence. Id., 123, 128. A second study demonstrated, however, that when an antipolygraph expert was added to the fact pattern, who testified that polygraph results should be treated "with skepticism," the effect of the polygraph evidence was almost completely eliminated.
Given the limited number and scope of the extant studies, and the equivocal conclusions thereof, definitive contentions that jurors will blindly follow polygraph evidence; see, e.g., United States v. Alexander, 526 F.2d 161, 169 (8th Cir. 1975); or that they will not; C. Honts M. Perry, supra, 16 Law Hum. Behav. 366; seem equally unfounded.

A polygrapher can ascertain a witness' deception or truth only indirectly, through physical manifestations thereof. It is unfair, however, to label a person as truthful or a liar based solely on such indirect, secondary indicia as a polygraph provides. Although a juror also considers secondary indicia, like demeanor, in assessing a witness' credibility, the juror has access to the rest of the evidence in the case as well, giving the juror several points of reference on which to base an assessment. In this way, a witness is not wrongly condemned, or elevated, solely because his or her body does not act in the "right" way.

This argument applies equally to the judge's role as fact finder at a bench trial.

Moreover, we afford criminal defendants the right to trial by a panel of several jurors partly out of the recognition that, although one person may be misled when a witness gives the "incorrect" physical signals, the cumulative impressions of the group are likely to lead to the truth. It violates the premise of this entire system to allow a single person — the polygrapher — to label a witness as honest or as dishonest based solely on the same type of indirect evidence that we generally maintain takes an entire jury to evaluate.

In this regard, we do not dispute that polygraphers may often reach a correct conclusion regarding a subject's guilt or innocence. We conclude, however, that this fact, in and of itself, is irrelevant. As illustrated in part II B 2 of this opinion, the ability of the polygraph technique to tell whether a subject is lying or telling the truth is still highly questionable. Thus, one cannot say with any degree of certainty that a polygrapher's ultimate conclusion about a subject's veracity is in fact based upon the polygraph machine — that is, based upon science. It is just as likely, if not more likely, that a polygrapher's conclusion will be based either on chance or on his or her general impressions of the subject's credibility. An assessment of witness credibility based simply on chance or on intuition is not, however, admissible at trial. See generally C. Tait J. LaPlante, supra, §§ 7.18 through 7.28 (credibility of witness can only be bolstered by certain methods, and only after credibility has been impeached). Indeed, forming impressions and intuitions regarding witnesses is the quintessential jury function; moreover, to the extent possible, luck should be excluded from the assessment process altogether.

It is interesting to note that critics of the Frye rule and critics of polygraph exclusion in general have on occasion asserted Frye's factual innocence as part of their arguments. "The premise appears to be that if, in truth, Frye was innocent of the crime charged, then the rule in the case is the culprit, for it permitted an innocent man to be convicted when science was ready and able to exonerate him. In the absence of the Frye court's parochial, constraining attitude toward the efficacy of scientific evidence, or so the proponents of this position seem to say, an innocent man would not have been unjustly convicted of murder and punished for it." J. Starrs, "`A Still-Life Watercolor': Frye v. United States," 27 J. Forensic Sci. 684, 687 (1982); see also footnotes 2 and 80. Thus, authors have at various times asserted that someone else confessed to the murder for which Frye was convicted; see, e.g., W. Wicker, "The Polygraphic Truth Test and the Law of Evidence," 22 Tenn. L. Rev. 711, 715 (1953); and that Frye was thereafter pardoned. See, e.g., P. Giannelli, supra, 80 Colum. L. Rev. 1204 n. 42.
Subsequent scholarship has demonstrated, however, that both the third party confession and the pardon arising therefrom are nothing more than "folklore," and that, in fact, Frye served eighteen years in jail before being paroled. See generally J. Starrs, supra, 27 J. Forensic Sci. 690, 692. Moreover, even if Frye had, in fact, been innocent, we conclude, for the reasons discussed in the text of this opinion, that the court in Frye was nonetheless correct to exclude the sphygmomanometer evidence.

Accordingly, the fact that a polygrapher's ultimate assessment happens to be right does not mean that it should be admitted at trial. If that opinion was not arrived at objectively through scientific principles and proof, then it is immaterial whether it happens to be correct because it was not based on any information that the jury itself could not have gleaned through its own observations. Because one cannot say with a high degree of certainty that a polygrapher's conclusion is based firmly in objective, scientific truth, and because one therefore cannot say that polygraph evidence provides a better informed assessment of a witness' credibility than the personal observations of each juror, allowing polygraph testimony would be a direct invasion of the province of the jury.

b

Furthermore, admission of polygraph test results at trial would likely produce regular, and immensely time consuming, "battles of the experts." Cf. Commonwealth v. Mendes, 406 Mass. 201, 202-203, 547 N.E.2d 35 (1989) (even after fifteen years of experience with polygraph evidence, evidentiary hearing consumed four days of court time). Admittedly, such a battle is common whenever scientific evidence is offered in court, and many types of scientific evidence are nonetheless routinely admitted at trials. In those instances, however, the probative value of the evidence is such that the delay attendant to the admission of the evidence is warranted. Because polygraph evidence is of such dubious probative value, however, the prejudicial impact of the likely delays weighs more heavily against admission.

Indeed, polygraph evidence is especially likely to cause disruptive conflicts between experts. Even polygraph advocates admit that "a substantial proportion of those who conduct tests in the public and private sectors lack adequate training and competence." D. Raskin, supra, 1986 Utah L. Rev. 66-67; see also C. Honts B. Quick, supra, 71 N.D. L. Rev. 998-99. Although expert qualification is always an issue with regard to scientific evidence, it is of particular import in the polygraph case.

With most types of scientific evidence, and especially with psychological tests, there are certain set procedures that are generally deemed to give accurate results. S. Blinkhorn, "Lie Detection As a Psychometric Procedure," in The Polygraph Test (A. Gale ed., 1988) pp. 29, 30-31. The expert debate in the courtroom simply focuses on whether these generally accepted, fixed procedures were adequately observed. This process ensures a certain level of "quality control."

There is, however, little standardization of polygraph test procedures. Id. With the polygraph test, each administration of a control question test is necessarily different, with appropriate control and relevant questions chosen and formulated by each individual examiner. As a result, a battle of the experts over polygraph evidence cannot simply address whether "the" test was properly administered. In the case of the control question test, for example, the experts will have to debate both whether the standardized control question test format was properly observed, and whether the particular, nonstandardized questions involved, as chosen by the examiner, were theoretically sound. The latter point will require an in-depth examination of the polygrapher's credentials. This additional layer of controversy will cause the qualification and questioning of experts to consume even more courtroom time than it ordinarily does.

Asking questions such as those used in the present case, involving such phrases as "do you know for sure," renders answers "chimerical" at best. United States v. Kwong, supra, 69 F.3d 668. The formulation of such questions adds to their prejudicial impact.

Furthermore, the very nature of the polygraph test makes extensive expert battles inevitable. Many experts dispute that the polygraph has any value whatsoever. By contrast, much other scientific evidence is premised on tests and procedures that are generally accepted at least in the abstract, if not as performed in a particular case. For example, the "basic genetic theory underlying DNA profiling [restriction fragment length polymorphism analysis] essentially is undisputed." J. McKenna, J. Cecil P. Coukos, "Reference Guide on Forensic DNA Evidence," in Reference Manual on Scientific Evidence, J. Moore, Federal Practice (1995) pp. 273, 285. Because many experts doubt the entire physiological and emotional basis of the polygraph test, many courtroom skirmishes will be fought over the scientific validity of the entire polygraph procedure.

Many experts assert that polygraph practitioners have completely failed to establish that the physical data produced by the polygraph machine are linked to deception, as opposed to other emotions, and thus that the physiological responses actually recorded by the polygraph have any regular correlation with a deceptive state of mind, as opposed to general anxiety. See, e.g., T. Ney, "Expressing Your Emotions and Controlling Feelings," in The Polygraph Test (A. Gale ed., 1988) p. 65. Moreover, there has been no investigation of whether individuals in different social, racial and ethnic subgroups have different emotional responses when confronted with a polygraph test. S. Blinkhorn, "Lie Detection As a Psychometric Procedure," supra, p. 39.

Polygraph results are ripe for in-court expert controversy, moreover, because, as previously discussed, the polygraph measures deception only indirectly, by gauging a subject's physiological state. The examiner must formulate questions that will properly link actual deception to physiological responses and relate the responses to those questions back to truth or deception. Other types of scientific evidence directly address an objective fact. Ballistics evidence, for example, directly addresses whether a particular bullet could have come from a particular gun, based on directly relevant physical indicators. With the polygraph test, however, the physiological test results can be linked to truth or deception only through the polygrapher's interpretation. This nexus between polygraph chart and polygraph significance is easily attacked by competing experts, and is likely to consume significant court time. Cf. M. Abbell, "Polygraph Evidence: The Case Against Admissibility in Federal Criminal Trials," 15 Am. Crim. L. Rev. 29, 41 (1977) ("even under optimal conditions, the results of polygraph examinations lack the reliability of the typical objective forensic tests which are admissible in evidence" due to uniquely subjective manner in which polygraph exams are interpreted).

c

Finally, there is a risk that, if polygraph evidence were admissible, juries would come to expect it with regard to all witnesses, a possibility that implicates both of the aforementioned concerns. Indeed, one wonders whether juries would draw adverse inferences against the witness whenever such evidence was not presented, or whether a Secondino instruction would be warranted. These admittedly minor issues would be less important if polygraph evidence had substantial probative value for the fact finder. As matters stand, however, this Pandora's box of subsidiary concerns further weighs against allowing the use of polygraph evidence at trial.

Secondino v. New Haven Gas Co., 147 Conn. 672, 676, 165 A.2d 598 (1960), and its progeny entitle a party to an adverse inference instruction if an opposing party fails to produce available evidence that, if favorable to that party, would naturally be produced.

C

Our decision to maintain our per se rule of exclusion with regard to polygraph evidence is consistent with the conclusion of the majority of courts that have considered this issue. State appellate courts, for whom Daubert is not mandatory authority, largely agree with our assessment that the prejudicial impact of polygraph evidence outweighs its probative value. As a result, approximately one half of the states have an absolute rule barring admission of polygraph evidence in criminal cases. See, e.g., People v. Sanchez, 169 Ill.2d 472, 662 N.E.2d 1199, 1210 (1996). Several of these courts have specifically held that the Daubert standard does not require admission of polygraph evidence, in light of the polygraph's questionable reliability and prejudicial impact. See, e.g., State v. Beard, supra, 194 W. Va. 746-47.

See also Pulakis v. State, 476 P.2d 474, 479 (Alaska 1970); People v. Anderson, 637 P.2d 354, 361 (Colo. 1981); Smith v. United States, 389 A.2d 1356, 1359 (D.C.), cert. denied sub nom. Washington v. United States, 439 U.S. 1048, 99 S.Ct. 726, 58 L.Ed.2d 707 (1978); State v. Okumura, 78 Haw. 383, 397, 894 P.2d 80 (1995); Morton v. Commonwealth, 817 S.W.2d 218, 221-22 (Ky. 1991); State v. Harnish, 560 A.2d 5, 8 (Me. 1989); State v. Hawkins, 326 Md. 270, 275, 604 A.2d 489 (1992); State v. Anderson, 379 N.W.2d 70, 79 (Minn. 1985), cert. denied, 476 U.S. 1141, 106 S.Ct. 2248, 90 L.Ed.2d 694 (1986); State v. Biddle, 599 S.W.2d 182, 191 (Mo. 1980); State v. Staat, 248 Mont. 291, 292, 811 P.2d 1261 (1991); State v. Steinmark, 195 Neb. 545, 548, 239 N.W.2d 495 (1976); Petition of Grimm, 138 N.H. 42, 54-55, 635 A.2d 456 (1993); Commonwealth v. Brockington, 500 Pa. 216, 220, 455 A.2d 627 (1983); State v. Hart, 911 S.W.2d 371, 377-78 (Tenn.App. 1995); Romero v. State, 493 S.W.2d 206, 213 (Tex.Crim.App. 1973); State v. Hamlin, 146 Vt. 97, 108-110, 499 A.2d 45 (1985); Robinson v. Commonwealth, 231 Va. 142, 156, 341 S.E.2d 159 (1986); State v. Beard, supra, 194 W. Va. 747; cf. People v. Angelo, 88 N.Y.2d 217, 223, 666 N.E.2d 1333, 644 N.Y.S.2d 460 (1996) (polygraph evidence properly excluded where there continues to be no showing that such evidence is generally accepted as reliable by scientific community).

The majority of the remaining states that have considered the issue admit polygraph evidence at trial only when its admission is stipulated to in advance by all parties. See, e.g., State v. Webber, 260 Kan. 263, 918 P.2d 609, 620 (1996). Even the jurisdictions that allow polygraph evidence by stipulation, however, generally do not assert that the evidence is probative, or that it gains validity by means of the stipulation. Instead, the allowance is simply based on the parties' right to waive evidentiary objections. See, e.g., Delap v. State, 440 So.2d 1242, 1247 (Fla. 1983). We are unpersuaded by this rationale. In our view, the limited reliability of polygraph evidence, taken together with its significant potential for prejudicial effect, compel the conclusion that such evidence should remain inadmissible even pursuant to a stipulation.

"Stipulation" in this context refers to a practice whereby both parties agree, before a subject takes a polygraph exam, that the results thereof will be admissible, but that the adversely affected party retains the right to cross-examine the polygraph witness and otherwise to attempt to impeach the polygraph evidence.

See also Ex parte Clements, 447 So.2d 695, 698 (Ala. 1984); State v. Valdez, 91 Ariz. 274, 283, 371 P.2d 894 (1962); Holcomb v. State, 268 Ark. 138, 140, 594 S.W.2d 22 (1980); People v. Fudge, 7 Cal.4d 1075, 1122, 875 P.2d 36, 31 Cal.Rptr.2d 321 (1994), cert. denied, 514 U.S. 1021, 115 S.Ct. 1367, 131 L.Ed.2d 223 (1995) (based on Cal. Evid. Code § 351.1 [Deering 1986], barring admission of any evidence relating to polygraph test in all criminal prosecutions); Melvin v. State, 606 A.2d 69, 71 (Del. 1992); Delap v. State, 440 So.2d 1242, 1247 (Fla. 1983), cert. denied, 467 U.S. 1264, 104 S.Ct. 3559, 82 L.Ed.2d 860 (1984); Fargason v. State, 266 Ga. 463, 467 S.E.2d 551, 553 (1996); State v. Fain, 116 Idaho 82, 86-87, 774 P.2d 252, cert. denied, 493 U.S. 917, 110 S.Ct. 277, 107 L.Ed.2d 258 (1989); Harris v. State, 481 N.E.2d 382, 384 (Ind. 1985); State v. Losee, 354 N.W.2d 239, 242 (Iowa 1984); Corbett v. State, 94 Nev. 643, 644, 584 P.2d 704 (1978); State v. McDavitt, 62 N.J. 36, 46, 297 A.2d 849 (1972); State v. Souel, 53 Ohio St.2d 123, 129-34, 372 N.E.2d 1318 (1978); State v. Eldredge, 773 P.2d 29, 37 (Utah), cert. denied, 493 U.S. 814, 110 S.Ct. 62, 107 L.Ed.2d 29 (1989); State v. Renfro, 96 Wash.2d 902, 905-907, 639 P.2d 737, cert. denied, 459 U.S. 842, 103 S.Ct. 94, 74 L.Ed.2d 86 (1982); Schmunk v. State, 714 P.2d 724, 731 (Wyo. 1986).

If, however, after a subject undergoes a polygraph examination, the adversely affected party believes that the examination did, in fact, result in an accurate factual determination, that party is of course free to stipulate to the results of the examination and, therefore, to the truth of those facts. In such a situation, however, it is difficult to believe that the case would actually be tried.

Of the states that do allow polygraph evidence without a stipulation, most allow it only in proceedings other than at trial. See, e.g., State v. Catanese, 368 So.2d 975, 981-83 (La. 1979) (admissible in posttrial proceedings). Only a very few states actually allow polygraph evidence that has not been stipulated to at the trial itself. See, e.g., Conner v. State, 632 So.2d 1239, 1257-59 (Miss. 1993) (evidence of willingness to take polygraph test admissible to rehabilitate impeached witness); State v. Sanders, 117 N.M. 452, 872 P.2d 870, 877 (1994) (polygraph evidence authorized by local rule of evidence under supervision of trial court); State v. Wright, 322 S.C. 256, 471 S.E.2d 700, 701 (1996) (polygraph admissibility at discretion of trial court, but should usually be excluded due to prejudicial effects). These states have concluded that, at least under certain circumstances, polygraph evidence is sufficiently more probative than prejudicial to warrant admission. For the reasons discussed in part II B of this opinion, however, we disagree, and instead conclude in agreement with the majority of states that such a balance in favor of admissibility is not, in fact, ever reached.

See also State v. Jones, 110 Ariz. 546, 551, 521 P.2d 978 (no error when polygraph results considered by sentencing judge), cert. denied, 419 U.S. 1004, 95 S.Ct. 324, 42 L.Ed.2d 280 (1974), overruled in part on other grounds, State v. Conn, 137 Ariz. 148, 151, 669 P.2d 581 (1983); People v. Barbara, 400 Mich. 352, 359, 255 N.W.2d 171 (1977) (admissible in postconviction hearing for new trial); State v. Bartholomew, 101 Wash.2d 631, 646, 683 P.2d 1079 (1984) (admissible at sentencing phase of capital cases).

In 1975, the New Mexico Supreme Court rejected its per se rule of polygraph exclusion, in the absence of a stipulation, as unnecessarily "mechanistic in nature." State v. Dorsey, 88 N.M. 184, 185, 539 P.2d 204 (1975). In 1983, that court adopted rule of evidence § 11-707, N.M. Stat. Ann. § 11-707 (Michie 1994), in an attempt to standardize the requirements for polygraph evidence admission and thereby minimize the prejudicial effects thereof. The prerequisites for admissibility established by § 11-707, including minimum polygrapher qualifications under N.M. Stat. Ann. § 11-707 (B) (Michie 1994), and the procedure for administering and scoring the exam that must be followed under N.M. Stat. Ann. § 11-707 (C) (Michie 1994), probably somewhat mitigate the prejudicial impact of polygraph evidence. Nonetheless, many of our concerns regarding such evidence, including its fundamental inaccuracy and improper invasion of the province of the jury, remain unaddressed, while others, such as the potential expert battles over examiner qualifications and test procedures, although addressed, seem likely to remain to a greater or lesser extent even under the statute.

In this regard, it is particularly instructive to note that several courts, after experimenting with polygraph admissibility for several years, rejected its admissibility and reinstated the traditional rule of inadmissibility. In each case, the court realized that its earlier assessment, namely, that the probative value of polygraph evidence outweighed its prejudicial impact, was mistaken. For example, in 1974 the Massachusetts Supreme Judicial Court decided to allow polygraph evidence for the limited purpose of corroborating or impeaching a defendant's trial testimony. Commonwealth v. A Juvenile, 365 Mass. 421, 425-26, 313 N.E.2d 120 (1974). The court concluded that "polygraph testing has advanced to the point where it could prove to be of significant value to the criminal trial process if its admissibility initially is limited to carefully defined circumstances designed to protect the proper and effective administration of criminal justice." Id., 425.

In 1989, however, the court reconsidered its position. Commonwealth v. Mendes, supra, 406 Mass. 201. After noting that the polygraph evidentiary hearing in Mendes had "consumed four days" of trial court time; id., 202; the court concluded that polygraph evidence continued to suffer from serious shortcomings, including the subjective nature of the polygraph method; the uncertain validity of polygraph evidence; the danger of proliferating a "battle of experts"; the danger of confusing the jury; the danger of usurping the jury's role; and the burden placed on trial courts in conducting evidentiary hearings. Id., 211. The court concluded that "[f]ifteen years has been more than enough time for examination and evaluation [of polygraph evidence]. . . . Further hope or expectation . . . is no longer warranted. . . . Accordingly, supported by the overwhelming authority throughout the country, we announce that polygraphic evidence, with or without pretest stipulation, is inadmissible in criminal trials in this Commonwealth. . . ." Id., 212.

Similarly, after a seven year experiment with admitting polygraph evidence on stipulation, the Wisconsin Supreme Court reinstated its per se exclusionary rule, concluding that "the burden on the trial court to assess the reliability of stipulated polygraph evidence [outweighs] any probative value the evidence may have." State v. Dean, 103 Wis.2d 228, 279, 307 N.W.2d 628 (1981). In 1974, the Wisconsin Supreme Court had decided to allow polygraph evidence to corroborate or impeach a defendant's testimony when: (1) all parties stipulated to admissibility; (2) the trial court had the discretion to reject such evidence notwithstanding the stipulation; (3) the party against whom such evidence was offered was allowed an opportunity for thorough cross-examination of the polygrapher; and (4) the jury was instructed that such evidence was only to be considered for corroborative purposes, and that they were free to determine what weight such evidence should be given. State v. Stanislawski, 62 Wis.2d 730, 742-43, 216 N.W.2d 8 (1974). As applied over the next several years, however, the Stanislawski rule created substantial practical problems in Wisconsin's trial courts. For example, it became unclear exactly what type of hearing, if any, the second Stanislawski prong required, and what standard appellate courts should use in reviewing a trial court's determination of polygraph reliability. State v. Dean, supra, 270. Under the third prong, it was unclear whether the party adversely affected by polygraph evidence could call its own experts to testify that polygraphs were unreliable. Id., 274-75. On account of these and similar problems, the Wisconsin Supreme Court in 1981 held polygraph evidence inadmissible in criminal trials. Id., 279. "Our analysis of and our experience with the Stanislawski rule lead us . . . to conclude that the Stanislawski conditions are not operating satisfactorily to enhance the reliability of the polygraph evidence and to protect the integrity of the trial process as they were intended to do." Id.

North Carolina and Oklahoma also allowed stipulated polygraph results for a brief time, then retreated to a per se rule of inadmissibility. State v. Grier, 307 N.C. 628, 643, 300 S.E.2d 351 (1983) ("[W]e are forced to conclude that the administration of justice simply cannot, and should not, tolerate the incredible burdens involved in the process of ensuring that a polygraph examination has been properly administered. If a trial court were to adequately police the reliability of stipulated results, the time required to explore the innumerable factors which could affect the accuracy of a particular test would be incalculable."); Fulton v. State, 541 P.2d 871, 872 (Okla. 1975).

In all, then, four states that at one time allowed polygraph evidence have subsequently, in light of their experiences, rethought and rejected this policy. Although not a decisive justification for our decision to retain our per se exclusionary rule, this pattern leads us to conclude that our concerns are, in fact, well founded.

Federal appellate courts generally grant trial judges more leeway to admit polygraph evidence than do their state counterparts, especially in the wake of Daubert. Indeed, the majority of federal courts of appeals do not have a per se rule that polygraph evidence is inadmissible at trial. Nonetheless, most maintain that, although admission is within the discretion of the trial court, such evidence should, as a general policy, be excluded under rule 403 of the Federal Rules of Evidence.

Only the Courts of Appeals for the Fourth and the District of Columbia Circuits currently maintain a per se rule of polygraph inadmissibility. See United States v. A S Council Oil Co., 947 F.2d 1128, 1133-34 and n. 4 (4th Cir. 1991) (affirming per se exclusion but noting that "[c]ircuits that have not yet permitted evidence of polygraph results for any purpose are now in the decided minority"); United States v. Skeens, 494 F.2d 1050, 1053 (D.C. Cir. 1974).
Moreover, only the Eighth Circuit Court of Appeals limits polygraph evidence at trial to instances where its admissibility has been stipulated to by both parties. See Houston v. Lockhart, 982 F.2d 1246, 1251-52 (8th Cir. 1993); but see United States v. Williams, 95 F.3d 723, 729-30 (8th Cir. 1996) (implying that polygraph evidence that has not been stipulated to may be admissible pursuant to Daubert if trial court finds that such evidence satisfies rule 403 of Federal Rules of Evidence).

See, e.g., United States v. Cordoba, 104 F.3d 225, 228 (9th Cir. 1997) (although Daubert precludes per se rule of polygraph inadmissibility, "[t]he inherent problematic nature of such evidence remains," and polygraph evidence is still subject to rule 403); United States v. Sherlin, 67 F.3d 1208, 1216 (6th Cir. 1995) (even if polygraph evidence is admissible under Daubert, "`unilaterally obtained polygraph evidence is almost never admissible under Evidence Rule 403'"); United States v. Posado, 57 F.3d 428, 435-36 (5th Cir. 1995) (although Daubert provides the proper threshold for admissibility of polygraph evidence, rule 403 must be satisfied as well); see also United States v. Scheffer, 44 M.J. 442, 446-47 (C.M.A. 1996). The Second Circuit Court of Appeals, while admitting it had not yet been presented with a record adequate to explore fully the polygraph issue, also recently affirmed a district judge's rejection of polygraph evidence on rule 403 grounds, even in light of Daubert. United States v. Kwong, supra, 69 F.3d 668-69. The First Circuit Court of Appeals recently reaffirmed that "polygraph examinations are generally inadmissible" as well. United States v. Black, 78 F.3d 1, 7 (1st Cir. 1996).

The Court of Appeals for the Eleventh Circuit has gone further than any other federal appellate court toward allowing polygraph evidence at trial in the absence of a stipulation. See United States v. Piccinonna, 885 F.2d 1529 (11th Cir. 1989). In Piccinonna, the court summarily concluded that, while polygraphy is still "a developing and inexact science," it is a sufficiently "useful and reliable scientific tool" to justify its admission to impeach or corroborate the testimony of a witness at trial, even without a stipulation. Id., 1535-36. The court did emphasize, however, that "[n]either of these two modifications to the per se exclusionary rule should be construed to preempt or limit in any way the trial court's discretion to exclude polygraph expert testimony on other grounds under the Federal Rules of Evidence." Id., 1536.

District courts in the Eleventh Circuit were thus given broader license to admit polygraph evidence than were district courts in any other circuit. It is noteworthy, therefore, that they have almost unanimously declined to do so. In fact, only one reported district court opinion from the Eleventh Circuit has held polygraph evidence admissible at trial. United States v. Padilla, 908 F. Sup. 923 (S.D.Fla. 1995). The court in Padilla conceded that "the polygraph examination at issue here does not have substantial probative value," but concluded nonetheless that polygraph evidence "will not disrupt the trial enough to conclude that its prejudicial effect substantially outweighs its probative value." (Emphasis in original.) Id., 929. Of particular note, however, is the court's acknowledgment that, but for the existence of Piccinonna, it probably would have excluded the polygraph evidence. Id., 927 n. 2. In other words, the court allowed the polygraph evidence to fulfill what it viewed as the mandate of Piccinonna, and not because its independent analysis was that such admission was wise or legally sound. Thus, even when presented with an opportunity to admit polygraph evidence, most district courts are decidedly reluctant to do so.

See United States v. One Parcel of Real Estate, 804 F. Sup. 319, 322 (S.D.Fla. 1992) (polygraph evidence excluded after proponent failed to lay adequate foundation for expert testimony as required by rule 702 of Federal Rules of Evidence); Elortegui v. United States, 743 F. Sup. 828, 830 n. 4 (S.D.Fla. 1990) (noting in dicta that use of polygraph to corroborate defendant's testimony "clearly could [be] excluded . . . under rule 403 on the grounds that it [is] cumulative"); United States v. Piccinonna, 729 F. Sup. 1336, 1338 (S.D.Fla. 1990) (on remand, trial court found that polygraph evidence failed to satisfy rule 608 [a] of Federal Rules of Evidence).

The most substantial of the few federal opinions permitting polygraph evidence at trial comes from the District Court of New Mexico. United States v. Galbreth, supra, 908 F. Sup. 877. The Tenth Circuit Court of Appeals had only addressed the question of polygraph admissibility before Daubert had been released; see United States v. Soundingsides, 820 F.2d 1232, 1241-42 (10th Cir. 1987); so the court in Galbreth felt free to formulate its own standard. The court accepted that Daubert provided the proper threshold standard; id., 878; and then relied largely on testimony by Raskin to conclude that polygraph evidence satisfied Daubert and rule 403 of the Federal Rules of Evidence. Id., 895. Although the court in Galbreth did address many of the concerns that have motivated us to retain our per se rule of exclusion, it did so by recounting only the most propolygraph studies and information. Id., 885-93. We believe that a more balanced review of the polygraph literature, such as we have conducted in the present case, reveals substantially more uncertainty regarding the effectiveness and prejudicial impact of the polygraph test than the court in Galbreth acknowledged.
The defendant also relies heavily on United States v. Crumby, 895 F. Sup. 1354 (D.Ariz. 1995). The court in Crumby held that, in light of Daubert, a reevaluation of the admissibility of polygraph evidence was required. Id., 1358. The court concluded, moreover, that with the proper safeguards, polygraph evidence should be permitted regarding a testifying defendant's credibility. Id., 1363.
Putting aside the question of the relevance of Crumby to the present case, in which the defendant did not testify, we find Crumby unpersuasive to the extent that it specifically argues that the probative value of polygraph evidence outweighs its prejudicial effect. First, based largely on the testimony of Raskin, the court concluded that the control question test has great probative value. Id., 1361. It is unclear from the opinion, however, whether the court's validity assessment referred to the polygraph's sensitivity and specificity, or to the more relevant predictive positive and predictive negative values. Regardless, our review of the literature leads us to a contrary conclusion. See generally part II B of this opinion. Second, the court in Crumby disputed the severity of the prejudicial impact from polygraph evidence. United States v. Crumby, supra, 895 F. Sup. 1361. The court in Crumby rejected the argument that the evidence would consume too much time at trial, opining that as polygraph evidence becomes more established, standards will develop, and so time spent reviewing this evidence will decrease. Id., 1362. Experience in Massachusetts and Wisconsin, among other states, indicates otherwise, as noted previously. The court also flatly denied that juries will be unduly influenced by polygraph evidence. Id. Although the court may be right on this point, the matter is not as clear as the opinion suggests.

III

The defendant's final claim is that the right to compulsory process, as guaranteed under the sixth amendment to the United States constitution and as applied to the states through the fourteenth amendment to the United States constitution, and also as accorded by article first, § 8, of the Connecticut constitution, guarantees him the right to present a defense, and that the exclusion of favorable polygraph results violates that right. Consequently, the defendant asserts that, at the very least, he was entitled to an evidentiary hearing regarding the reliability of the polygraph examination. We disagree.

Because the defendant has failed to provide any independent analysis under the state constitution, we limit our review to the federal constitution. State v. Dyson, 238 Conn. 784, 794, 680 A.2d 1306 (1996); State v. Barnes, 232 Conn. 740, 744 n. 4, 657 A.2d 611 (1995).

Although an accused does, of course, have a federal constitutional right to present a defense, "the right to present relevant testimony is not without limitation." Rock v. Arkansas, 483 U.S. 44, 55, 107 S.Ct. 2704, 97 L.Ed.2d 37 (1987); Montana v. Egelhoff, 518 U.S. 37, 42, 116 S.Ct. 2013, 135 L.Ed.2d 361 (1996) ("[t]he proposition that the Due Process Clause guarantees the right to introduce all relevant evidence is simply indefensible"); Batick v. Seymour, 186 Conn. 632, 637, 443 A.2d 471 (1982). "Rules for the admission and exclusion of evidence should be found offensive to notions of fundamental fairness embodied in the United States Constitution only when, (1) without a rational basis, they disadvantage the defendant more severely than they do the State, or (2) [they] arbitrarily exclude reliable defensive evidence without achieving a superior social benefit." Perkins v. State, 902 S.W.2d 88, 94 (Tex.App. 1995); see Rock v. Arkansas, supra, 53-56. As already discussed at length in this opinion, there is a very strong basis for excluding polygraph evidence, and such exclusion is not arbitrary given the polygraph test's demonstrated lack of probative value.

Some authorities have argued that Rock provides a "strong constitutional impetus" for polygraph admission. J. McCall, supra, 1996 U. Ill. L. Rev. 407. We disagree. In Rock, the Supreme Court struck down Arkansas' per se ban on hypnotically refreshed testimony, concluding that such testimony could be reliable under certain conditions, and thus that a per se rule barring admission was unconstitutional. Rock v. Arkansas, supra, 483 U.S. 59-60, 62. Rock arose, however, in the context of a manslaughter case in which the defendant wished to testify on her own behalf. In reaching its conclusion, the court emphasized that Arkansas' rule had impermissibly infringed on the right "of a defendant to testify"; id., 62; that the "rule had a significant adverse effect on [the defendant's] ability to testify"; id., 57; and that no other state had a per se rule barring the hypnotically refreshed "testimony of a defendant." (Emphasis in original.) Id. By excluding polygraph evidence, we in no way restrict any defendant's right or ability to testify.

See also Jackson v. Garrison, 677 F.2d 371, 373 (4th Cir. 1981) ("[R]estrictions governing the admission of polygraph evidence are matter[s] of state law and procedure not involving federal constitutional issues. . . . The exclusion of polygraph evidence did not negate the fundamental fairness of the petitioner's trial or violate a specific constitutional right." [Citation omitted; internal quotation marks omitted.]).

Moreover, we agree with the Appellate Court that an evidentiary hearing in the present case would have been a "nugatory undertaking." State v. Porter, supra, 39 Conn. App. 803. We have conducted our own thorough review of the literature and case law, and are convinced of the continued wisdom of our long established rule of polygraph inadmissibility. We acknowledge that several different polygraph testing techniques currently exist, and that lie detection technology continues to evolve. If, at some future date, substantial evidence indicates that some polygraph or other lie detection technique has reached a sufficiently high level of validity that the probative value of such evidence potentially outweighs its prejudicial impact, we may be forced to revisit this issue. Until then, however, we see no purpose in requiring the trial courts of this state to undertake evidentiary hearings every time a defendant proffers polygraph evidence.

The dissent asserts that denying the defendant an evidentiary hearing on his polygraph evidence deprives him "of his constitutional right to present a defense, [and] also his right to be heard at a meaningful time and in a meaningful manner — both rights which . . . are at the core of due process under the federal and state constitutions." The flaw in this argument is its underlying premise that the defendant could make a showing, if a hearing were granted, that his polygraph evidence is admissible. For, as a matter of logic, there can be no constitutional right to an evidentiary hearing regarding evidence that, as a matter of law, is inadmissible. Cf. State v. Coleman, 41 Conn. App. 255, 262-64, 675 A.2d 887, cert. granted, 237 Conn. 931, 677 A.2d 1372 (1996). Given our conclusion that, as a matter of law, the prejudicial impact of polygraph evidence outweighs any possible probative value, we fail to see why the defendant is entitled to a hearing on this topic, or what function such a hearing could serve. In this connection, we simply disagree with the majority opinion in United States v. Scheffer, 44 M.J. 442 (C.M.A. 1996), on which the dissent herein relies, and agree with the dissent in Scheffer that the sixth amendment right to confrontation is not violated by a per se evidentiary rule barring the admissibility of polygraph evidence.
The dissent sets forth two scenarios to support its underlying position that, despite the polygraph's questionable validity, we should abandon our per se rule against admitting polygraph evidence at trial. First, the dissent claims that minority criminal defendants are so disadvantaged by the problems attendant to cross-racial identifications that, essentially, it is only fair to allow these defendants to present exculpatory polygraph results. This court has specifically rejected the notion of special treatment for defendants in cross-racial identification situations, however, holding that the mere fact that a defendant is of a different race than a witness does not entitle the defendant to a special instruction on eyewitness identification at trial. State v. Cerilli, 222 Conn. 556, 571-72, 610 A.2d 1130 (1992). Similarly, the mere fact that a witness and a defendant may be of different races does not warrant admitting evidence that, whatever the races of the trial participants, is of very questionable probative value and is highly prejudicial.
The dissent also asserts that our per se exclusionary rule may lead to the death penalty being imposed on an innocent defendant. Due to the prejudicial nature of polygraph evidence and its lack of probative value, we have concluded that polygraph evidence is inadmissible in any judicial proceeding which is subject to the rules of evidence. It would make little sense to subject such proceedings — which are part of the trial process, even when they are conducted separately — to the minimal validity and great prejudice of polygraph evidence that we determine is inappropriate for the trial itself. As the dissent is well aware, however, the rules of evidence simply do not apply to the presentation of mitigating factors during the penalty phase of a capital case. See General Statutes § 53a-46a (c); see also State v. Ross, 230 Conn. 183, 268, 646 A.2d 1318 (1994), cert. denied, 513 U.S. 1165, 115 S.Ct. 1133, 130 L.Ed.2d 1095 (1995) ("[o]n its face, [§ 53a-46a (c)] authorizes a judge presiding over a penalty hearing to exclude mitigating evidence only on the basis of a lack of relevancy" [emphasis in original]). Consequently, the per se rule that we articulate today is not a bar to the introduction of polygraph evidence during the penalty phase of a capital case pursuant to § 53a-46a (c). Cf. State v. Bartholomew, 101 Wash.2d 631, 646, 683 P.2d 1079 (1984) (polygraph evidence admissible by defense at sentencing phase of capital case). Thus, with respect to court proceedings to which the rules of evidence do not apply, such as the sentencing phase of a capital case, or sentencing proceedings generally, we adopt no blanket rule regarding the admissibility of polygraph evidence, and leave that question to a case in which there has been a claim that the trial court has abused its discretion in either admitting or precluding such evidence.
Moreover, the dissenter's inference that we have "attempt[ed] to allay [his] concerns by suggesting that polygraph evidence is admissible" in the penalty phase of a capital case is misguided. We are not engaged in the process of "allaying" the dissenter's "concerns," nor do we endorse the judicial psychoanalysis by which he draws such an inference.

See part II of this opinion.

For example, we recognize that a modification of the control question test, the "directed lie control question test," has been developed over the past few years. Indeed, the results from the latter test have been admitted in at least one case. See United States v. Galbreth, supra, 908 F. Sup. 877. "The basic rationale of the [directed lie control question test] is similar to that of the [control question test]. Subjects are asked direct accusatory relevant questions. Physiological responses to the relevant questions are evaluated against the physiological response to comparison questions. However, in the [directed lie control question test] the comparison questions are known lies." C. Honts B. Quick, supra, 71 N.D. L. Rev. 1001-1002. "[T]he subject [is] asked a question like, `Before 1994, did you ever tell even one lie in your entire life?'. . . The subject [is then] instructed to answer this question on the test with a lie, and [is also] instructed to: think of a specific incident of lying and to pay close attention to his or her physiological responses. . . . Subjects are told that it is important that they respond appropriately when lying and that if they fail to do so the test will be inconclusive. . . ." (Citations omitted.) Id., 1002 n. 93. "The rationale of the [directed lie control question test] predicts that guilty subjects will respond with greater physiological responses to the relevant questions while innocent subjects will respond with greater responses to the directed lie questions." Id., 1002.
The directed lie control question test has been praised both because it allows for standardization of the pretest interviews and control questions; id.; and because it is less intrusive than the traditional control question test. See United States v. Galbreth, supra, 908 F. Sup. 885. Moreover, such limited studies as have been conducted indicate that the directed lie control question test might produce somewhat fewer false positives than the traditional control question test. See, e.g., C. Honts D. Raskin, "A Field Study of the Validity of the Directed Lie Control Question," 16 J. Police Sci. Admin. 56, 60 (1988). Even if we concede, however, that the directed lie control question test is "slightly more accurate" than the control question test; C. Honts B. Quick, supra, 71 N.D. L. Rev. 1002; all of the prejudicial effects of control question test evidence persist under the directed lie control question test, thereby justifying its exclusion.


The court today retains a per se rule that bars from evidence the results of a polygraph test under any circumstances. This per se rule infringes on the defendant's constitutional right to present a defense. The significance — and indeed the absurdity — of this rule barring polygraph evidence under any circumstances is demonstrated by hypothesizing the following factual scenario. A defendant is accused of a capital felony subject to the death penalty and the only issue is one of identification — that is, whether he was incorrectly identified by a witness as the perpetrator of the crime. The defendant submits to a properly administered polygraph examination that indicates that he is telling the truth when he says that he did not kill the victim. He also submits to a polygraph examination administered by the state that confirms the truthfulness of his statement that he was not the perpetrator. Notwithstanding the results of these two polygraph tests, the state continues with its prosecution of the defendant because of the strong identification testimony of its sole eyewitness. The thought of executing a person found guilty under these circumstances would shock anyone's conscience, whether the polygraph is 91 percent or merely 59 percent accurate. "The quintessential miscarriage of justice is the execution of a person who is entirely innocent." Schlup v. Delo, 513 U.S. 298, 324-25, 115 S.Ct. 851, 130 L.Ed.2d 808 (1995). This per se rule of the court collides with the "`fundamental value determination of our society,' given voice in Justice Harlan's concurrence in Winship, that `it is far worse to convict an innocent man than to let a guilty man go free.'" Francis v. Franklin, 471 U.S. 307, 313, 105 S.Ct. 1965, 85 L.Ed.2d 344 (1985), quoting In re Winship, 397 U.S. 358, 372, 90 S.Ct. 1068, 25 L.Ed.2d 368 (1970) (Harlan, J., concurring). "[C]oncern about the injustice that results from the conviction of an innocent person has long been at the core of our criminal justice system." Schlup v. Delo, supra, 325.

The majority, in footnote 80, in response to my concerns about the execution of an innocent person, attempts to allay those concerns by suggesting that polygraph evidence is admissible before the jury in the penalty phase of a capital case. I am unable to understand how polygraph evidence could be sufficiently reliable for the jury to consider when determining whether a person should be executed or whether a person should spend the rest of his or her life confined in prison, but not sufficiently reliable for the purpose of determining guilt or innocence. The logic of the majority eludes me.

See part IV of this opinion.

Indeed, if the defendant in the present case was an African-American, as in the companion case, and the case involved cross-racial identification, not to allow the jury to consider the properly administered polygraph test results when considering the reliability of the identification of the accused is simply unacceptable, whether the accused's life is at stake or whether he faces the loss of his liberty. Furthermore, it is well known that the frailties of eyewitness identification are not limited to cross-racial identification.

See State v. Hunter, 241 Conn. 165, 694 A.2d 1317 (1997).

"It is well documented that cross-racial identification is less reliable than identification of one person by another of the same race. [C]onsiderable evidence indicates that people are poorer at identifying members of another race than of their own. . . . Moreover, counterintuitively, the ability to perceive the physical characteristics of a person from another racial group apparently does not improve significantly upon increased contact with other members of that race. Because many crimes are cross-racial, these factors may play an important role in reducing the accuracy of eyewitness perception. Note, `Did Your Eyes Deceive You? Expert Psychological Testimony on the Unreliability of Eyewitness Identification,' 29 Stan. L. Rev. 969, 982 (1977). Elizabeth F. Loftus, in her classic treatise entitled `Eyewitness Testimony,' wrote that [i]t seems to be a fact — it has been observed so many times — that people are better at recognizing faces of persons of their own race than a different race. E. Loftus, Eyewitness Testimony (1979) pp. 136-37. She concluded, on the basis of an examination of studies made on the subject, that [p]eople have greater difficulty in recognizing faces of another race than faces of their own race. Id., 139. State v. Cerilli, 222 Conn. 556, 588-89, 610 A.2d 1130 (1992) ( Berdon, J., dissenting).
"Concern for the problems of cross-racial identification is well documented in our case law as well as by other social scientists. We are painfully aware of miscarriages of justice caused by wrongful identification. Those experienced in criminal trial work or familiar with the administration of justice understand that one of the great problems of proof is posed by eyewitness identification, especially in cross-racial identification. . . . Judge Bazelon, dissenting in United States v. Butler, 636 F.2d 727 (D.C. Cir. 1980), cert. denied, 451 U.S. 1019, 101 S.Ct. 3010, 69 L.Ed.2d 392 [1981], observed that many experts have concluded that convictions based solely on one eyewitness identification represent conceivably the greatest single threat to the achievement of our ideal that no innocent men shall be punished. . . . Brown v. Davis, 752 F.2d 1142, 1146 (6th Cir. 1985); see also Kampshoff v. Smith, 698 F.2d 581 (2d Cir. 1983). Another important factor to be considered in assessing the reliability of an identification is whether the witness and the person identified are of the same or different races. In general, there is a much greater possibility of error where the races are different than where they are the same. Where they are different, there is more likelihood of error where the suspect belongs to a minority group and the witness to a majority group than there is in the opposite situation. Almost fifty years ago, it was said to be: well known that, other things being equal, individuals of a given race are distinguishable from each other in proportion to our familiarity, to our contact with the race as a whole. Thus to the uninitiated American, all [Asians] look alike, while to the [Asian] all white men look alike. P. Wall, Eye-Witness Identification in Criminal Cases (2d Ed.) p. 122; A. Yarmey, The Psychology of Eyewitness Testimony (1979) pp. 130-31." (Emphasis in original; internal quotation marks omitted.) State v. Reddick, 224 Conn. 445, 471-72 n. 1, 619 A.2d 453 (1993) ( Berdon, J., dissenting.)

In this case, the defendant sought a hearing before the trial court in order to demonstrate the validity of the polygraph test results he sought to introduce. The trial court, relying upon our failure to adopt Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579, 113 S.Ct. 2786, 125 L.Ed.2d 469 (1993), for the admission of scientific evidence and assuming that we would continue to adhere to the antiquated rule of Frye v. United States, 293 F. 1013 (D.C. Cir. 1923), rejected the defendant's request for an evidentiary hearing. The trial court in the present case stated that it was "only a trial level court" and that an "appellate court [must] direct a trial court like this court to have [a] hearing." Nevertheless, the defendant sought before the trial court the very procedure that we subsequently indicated should be followed in a recent opinion authored by Justice Borden in State v. Esposito, 235 Conn. 802, 832-33, 670 A.2d 301 (1996). In Esposito, we sent the strong message that in order "[t]o create a record sufficient to allow this court to consider altering the long-standing Connecticut law barring polygraph evidence, the defendant bore the burden of demonstrating that traditional reasons for the rule are no longer applicable. Even under the Daubert rule, the evidence must be shown to be relevant and reliable to be admissible. . . . [ T] he defendant bore the burden of creating a factual record before the trial court that the polygraph test possesses sufficient reliability to justify its introduction as scientific evidence." (Citation omitted; emphasis added.) Id.

In groping for support for its position, the majority, quoting the Appellate Court out of context, claims at the end of its opinion that the Appellate Court had also concluded that a hearing on the admissibility of the polygraph evidence would be a "nugatory undertaking." What the Appellate Court held was that because of this court's precedent holding that polygraph evidence is inadmissible, a hearing on the validity of the polygraph examination would serve no purpose. See State v. Porter, 39 Conn. App. 800, 803, 668 A.2d 725 (1995) ("The trial court, like this court, is bound by the Connecticut precedent which bars the admission of polygraph results. . . . Because an evidentiary hearing would have been a nugatory undertaking, the trial court was not required to grant the defendant's motion for an evidentiary offer of proof." [Citation omitted.]).

The majority of this court, instead of predicating its decision on a factual record, today decides as a matter of law that the polygraph, under any circumstances, does not have sufficient validity for the results of the test to be admissible into evidence. There are two procedural problems underlying the majority's conclusion that have constitutional implications. First, the defendant was never given an opportunity to either develop the factual basis for his claim, or, equally important, an opportunity to argue the validity of an appropriately administered polygraph. This court, on its own, contrary to Daubert, which we embrace today, delves into the literature on the subject and decides that the polygraph does not have a sufficiently high level of validity and, therefore, determines that its prejudicial impact outweighs its probative value. Second, we did not certify that issue on appeal from the Appellate Court. Rather, we certified two issues: (1) whether the defendant had the right to an evidentiary hearing, on the issue of whether the polygraph evidence was admissible; and (2) whether we should adopt Daubert for the admissibility of scientific evidence. State v. Porter, 236 Conn. 908, 670 A.2d 1308 (1996). In view of the certified issues, neither the state nor the defendant addresses the validity and reliability of a properly administered polygraph examination. Without giving the defendant an opportunity to demonstrate the scientific underpinnings and the accuracy of the polygraph, this court not only deprives the defendant of his constitutional right to present a defense, but also his right to be heard at a meaningful time and in a meaningful manner — both rights which, of course, are at the core of due process under the federal and state constitutions. Armstrong v. Manzo, 380 U.S. 545, 552, 85 S.Ct. 1187, 14 L.Ed.2d 62 (1965); Roundhouse Construction Corp. v. Telesco Masons Supplies Co., 168 Conn. 371, 385, 362 A.2d 778, vacated, 423 U.S. 809, 96 S.Ct. 20, 46 L.Ed.2d 29 (1975), aff'd on remand, 170 Conn. 155, 365 A.2d 393, cert. denied, 429 U.S. 889, 97 S.Ct. 246, 50 L.Ed.2d 172 (1976). Incredibly, not content to bar polygraph evidence from criminal trials, the court today in one sweeping sentence in a footnote also adopts a per se rule excluding polygraph evidence in civil trials.

See footnote 3 of the majority opinion.

See part V of this opinion.

See footnote 36 of the majority opinion.

I

First, it is important that the issues in this case be placed in the context of the evidence presented at the defendant's trial. The defendant was charged with arson for a fire that destroyed his home on July 20, 1992. The defendant and his wife had moved into their new residence in September, 1991, and had extensively renovated the house. During the defendant's trial, his wife testified that an area of the living room, which was identified as one of the sources of the fire, was primed and painted the day before the fire occurred. On the day of the fire, the defendant, his wife and their child smelled smoke and observed a haze in the house. The three of them then went to a neighbor's house across the street. The defendant returned to the family home and emptied a fire extinguisher on the electrical panel in the basement, the location from which the defendant observed smoke originating. The defendant then returned to the neighbor's house and telephoned his brother-in-law, a licensed electrician, for advice. A fire subsequently consumed the house.

The deputy fire marshal for the city of Norwich, who testified as a state's witness, was aware that there were items such as paint thinner, turpentine, aerosol cans, jelly substances and other flammable items used in the painting of the interior of the home during the time that the fire occurred. The deputy fire marshal testified that he found three points of origin of fire, but that he had no opinion as to the source of the ignition of the fire. He also testified that he never considered electrical arcing as a potential source of ignition of the fire because no one communicated to him that there was such electrical arcing in the house. While the firefighters were tending to the fire, however, the electricity in the house was not shut off and other witnesses observed electrical arcing.

An electrical arc is defined as "a sustained brilliantly luminous glow sometimes having the appearance of a curved line of flame that is formed under certain conditions when a break is make in an electric circuit." Webster's Third New International Dictionary. One of the firefighters, a captain of the Yantic fire company, testified for the state at the defendant's trial and he described electrical arcing as "a bluish color, bluish white color that arcs and it sounds . . . like a popping crackling noise, followed by a blue and whitish color. . . . It's kind of like [the] sparkler you see at the 4th of July [that] the kids have." Regarding the general distance of an electrical arc, he testified that "if you're talking 110, 220 voltage in an average house as you have up there . . . it depends on what it's feeding, depends on how much of the line has been burnt, if you had a long electrical cord where the fire had melted the cord down and you put water on it, the whole thing could arc maybe six feet. It could be only two feet. It's very hard to tell." The same firefighter also testified that he did not see any electrical arcing while he was fighting the fire, but that, while at the scene of the fire, it was transmitted to him by portable radio that other firefighters had observed electrical arcing.

The deputy fire marshal identified a floor lamp, contained in a photograph admitted as a state's exhibit, that was near a couch located in one of the three points of origin of the fire. Although the fire marshal testified that electrical arcing can cause fires, he failed to have the lamp or the electrical outlet dismantled and tested in order to determine if either of those were the source of the ignition of the fire. Indeed, the defendant's attorney elicited testimony on cross-examination of the deputy fire marshal that the burn pattern on the lamp was inconsistent with his theory of where the fire started and that he had no explanation for such an inconsistency. In addition, he testified that if a fire is caused by an electrical arc, neither he nor any one else trained in the field of fire investigation would be able to determine the contact between an appliance and the fire.

At the trial, the defendant's homeowner's insurance agent indicated that the home was underinsured by approximately $48,000 at the time of the fire. The insurance agent also testified that an increase in insurance was due to go into effect on August 30, 1992, one month and ten days after the fire. Previous to the fire, the home was evaluated by the insurance company to have been in excellent condition with many renovations having been made. The renovations included new kitchen cabinets, tiled bathrooms, electrical plugs and painted walls, and new carpeting was set to arrive a week after the fire occurred. A certified public accountant testified that the defendant and his wife were current with their bills and had even prepaid their mortgage at the time of the fire. The accountant also testified that the defendant and his wife had invested most of their life savings into the renovation of their home. There were no eyewitnesses to the start of the fire.

The defendant did not testify on his own behalf, but his sworn statement to the police, concerning the denial of his participation in setting the fire and the circumstances surrounding the fire, was introduced into evidence by the state. The defendant submitted to a polygraph examination on January 27, 1993, which was administered by Leighton R. Hammond, a certified polygraphist and president of the Associated Detective Bureau, Inc. The defendant, well before the trial commenced, filed a pretrial motion to admit the results of a polygraph examination that he had taken. In the opinion of Hammond, the defendant was telling the truth when he answered that he neither set the fire nor knew if anyone else had purposely set the fire. Although there was sufficient time to order the defendant to submit to a polygraph examination from an expert retained by the state, as a condition for the admission of the defendant's polygraph test results, the trial court, after argument, denied the defendant's motion for a hearing regarding the admissibility of the polygraph examination. As evidenced by the above facts, the defendant's credibility was crucial to his defense in light of the fact that the state's case was paper-thin. Indeed, the jury agreed that the state failed to prove the fire was started by the defendant for the purpose of collecting insurance proceeds by finding him not guilty on the first count of the information, which alleged a violation of General Statutes § 53a-111 (a) (3). Rather, the jury found him guilty on the second count of starting a fire under circumstances in which a firefighter was subjected to substantial risk of bodily injury in violation of § 53a-111(a)(4). The defendant was sentenced to imprisonment for a period of seventeen years, execution suspended after twelve years.

At oral argument before the trial court regarding the defendant's motion to admit the results of his polygraph examination, the defendant's attorney represented to the trial court that Hammond works for fifty-three police departments and performs all of the polygraph testing for the Windsor police department.

General Statutes § 53a-111 provides in relevant part: "Arson in the first degree: Class A felony. (a) A person is guilty of arson in the first degree when, with intent to destroy or damage a building, as defined in section 53a-100, he starts a fire or causes an explosion, and . . . (3) such fire or explosion was caused for the purpose of collecting insurance proceeds for the resultant loss. . . ."

General Statutes § 53a-111 (a) provides in relevant part: "A person is guilty of arson in the first degree when, with intent to destroy or damage a building, as defined in section 53a-100, he starts a fire or causes an explosion, and . . . (4) at the scene of such fire or explosion a peace officer or firefighter is subjected to a substantial risk of bodily injury."

II

I agree with the majority that for the admissibility of scientific evidence we should abandon the Frye test and adopt the general principles of Daubert v. Merrell Dow Pharmaceuticals, Inc., supra, 509 U.S. 579. Under Daubert, the scientific evidence need not be generally accepted by the scientific community; rather, the scientific evidence must be relevant and reliable. Id., 594-95. In Daubert, the court delineated certain considerations, although not an exhaustive checklist, in order to guide trial courts through their evaluation of the admissibility of scientific evidence: (1) whether the theory or technique can be, or has been, tested; (2) whether the theory or technique has been subjected to peer review and publication, although publication (which is one element of peer review) is not an indispensable condition; (3) the known or potential rate of error and the existence and maintenance of standards controlling the technique's operation; and (4) whether the theory or technique possesses general acceptance within the scientific community. Id., 592-95. The Supreme Court also stated that "[t]he inquiry . . . is, we emphasize, a flexible one. Its overarching subject is the scientific validity — and thus the evidentiary relevance and reliability — of the principles that underlie a proposed submission. The focus, of course, must be solely on principles and methodology, not on the conclusions that they generate." Id., 594-95. If the trial court determines that the scientific evidence is relevant and reliable, then the trial court must determine whether to exclude the evidence because its probative value is substantially outweighed by its prejudicial effect, or whether the scientific evidence will confuse the issues or mislead the jury. Id., 595. In other words, as the majority states it, a trial judge should deem scientific evidence inadmissible "only when the methodology underlying such evidence is sufficiently invalid to render the evidence incapable of helping the fact finder determine a fact in dispute." In exercising this discretion, the trial court, however, must take into account the defendant's right to present a defense under the sixth amendment to the federal constitution and article first, § 8, of the state constitution.

See part V of this opinion.

Our conclusion today that Daubert is the correct standard for determining the admissibility of scientific evidence requires, as I shall subsequently point out, that trial courts determine on a case-by-case basis whether polygraph evidence is admissible. Indeed, the Daubert standard will have the salutary effect of opening the door to a new vista of scientific evidence in order to achieve justice.

See footnote 4 of this opinion with respect to identification evidence. In United States v. Amador-Galvan, 9 F.3d 1414, 1417-18 (9th Cir. 1993), the court held that Daubert overruled a per se rule excluding expert testimony regarding the credibility of eyewitness identification. See also United States v. Cordoba, 104 F.3d 225, 227 (9th Cir. 1997) (applying reasoning in Amador-Galvan to polygraph evidence and holding that per se rule was overruled by Daubert).

III

In my view, there is substantial authority that would at least justify affording the defendant an opportunity to have a hearing regarding the admissibility of the polygraph evidence. In United States v. Piccinonna, 885 F.2d 1529 (11th Cir. 1989), the Court of Appeals for the Eleventh Circuit, sitting en banc, abandoned the per se exclusionary rule regarding polygraph evidence, even before Daubert was decided. "Since the Frye decision, tremendous advances have been made in polygraph instrumentation and technique. Better equipment is being used by more adequately trained polygraph administrators. Further, polygraph tests are used extensively by government agencies. Field investigative agencies such as the FBI, the Secret Service, military intelligence and law enforcement agencies use the polygraph. Thus, even under a strict adherence to the traditional Frye standard, we believe it is no longer accurate to state categorically that polygraph testing lacks general acceptance for use in all circumstances. For this reason, we find it appropriate to reexamine the per se exclusionary rule and institute a [ new] rule more in keeping with the progress made in the polygraph field." (Emphasis added.) Id., 1532.

The court in Piccinonna established certain conditions, as we should, for the admissibility of polygraph evidence. Besides allowing polygraph results to be admitted based on a stipulation, the court in Piccinonna outlined that this evidence is admissible to impeach or corroborate the testimony of a witness if: (1) the party planning to use the polygraph evidence provides adequate notice to the opposing party that the expert testimony will be offered; (2) the opposing party must be given a reasonable opportunity to have its own polygraph examiner administer a test covering substantially the same questions; and (3) any relevant requirements of the Federal Rules of Evidence must be followed — for instance, evidence that a witness passed a polygraph examination cannot be used to corroborate that witness' testimony until after the credibility of that witness has been attacked. United States v. Piccinonna, supra, 885 F.2d 1536. "Even where the above three conditions are met, admission of polygraph evidence for impeachment or corroboration purposes is left entirely to the discretion of the trial judge." Id.

"The instrument used in Frye measured only one physiological response (blood pressure), whereas the modern polygraph measures respiration and galvanic skin resistance in addition to blood pressure. The technique also has been improved through the development of control questions, the pretest interview, and simulation methods." 1 P. Giannelli E. Imwinkelried, Scientific Evidence (2d Ed. 1993) § 8-3, p. 231 n. 75.

"There is no question that in recent years polygraph testing has gained increasingly widespread acceptance as a useful and reliable scientific tool. Because of the advances that have been achieved in the field which have led to the greater use of the polygraph examination, coupled with a lack of evidence that juries are unduly swayed by polygraph evidence, we agree with those courts which have found that a per se rule disallowing polygraph evidence is no longer warranted. Of course, polygraphy is a developing and inexact science, and we continue to believe it inappropriate to allow the admission of polygraph evidence in all situations in which more proven types of expert testimony are allowed. However, as Justice Potter Stewart wrote, ` any rule that impedes the discovery of truth in a court of law impedes as well the doing of justice.' Hawkins v. United States, 358 U.S. 74, 81, 79 S.Ct. 136, 140, 3 L.Ed.2d 125 (1958) [Stewart, J., concurring]. Thus, we believe the best approach in this area is one which balances the need to admit all relevant and reliable evidence against the danger that the admission of the evidence for a given purpose will be unfairly prejudicial." (Emphasis added.) United States v. Piccinonna, supra, 885 F.2d 1535.

Indeed, in a recent unanimous decision, the Court of Appeals for the Ninth Circuit abandoned the per se rule against the admissibility of polygraph evidence. " Daubert holds that it is the trial judge's task, rather than [an appellate court's task], to conduct the initial weighing of probative value against prejudicial effect. Accordingly, we hold that Daubert also overruled any per se rule [in the Ninth Circuit] that unstipulated polygraph evidence is always inadmissible under [r]ule 403 [of the Federal Rules of Evidence]. Requiring the trial judge to conduct the [r]ule 403 analysis is consistent with the law of other circuits." United States v. Cordoba, 104 F.3d 225, 228 (9th Cir. 1997); see United States v. Posado, 57 F.3d 428, 433-34 (5th Cir. 1995) ("After Daubert, a per se rule is not viable. . . . Remaining controversy about test accuracy is almost unanimously attributed to variations in the integrity of the testing environment and the qualifications of the examiner. . . . We merely remove the obstacle of the per se rule against admissibility, which was based on antiquated concepts about the technical ability of the polygraph and legal precepts that have been expressly overruled by the Supreme Court." [Citations omitted.]); see also United States v. Pettigrew, 77 F.3d 1500 (5th Cir. 1996) (reaffirming elimination of per se rule in Posado); United States v. Pulido, 69 F.3d 192, 205 (7th Cir. 1995) ("[t]his court has persistently refused to adopt [a per se rule against admissibility of polygraphs], choosing rather to leave the decision on admissibility to the sound discretion of the district court" [internal quotation marks omitted]); United States v. Sherlin, 67 F.3d 1208 (6th Cir. 1995) (acknowledging Daubert as standard for admissibility of polygraph evidence but holding that decision to exclude polygraph evidence rests within sound discretion of District Court); McMorris v. Israel, 643 F.2d 458 (7th Cir. 1981), cert. denied, 455 U.S. 967, 102 S.Ct. 1479, 71 L.Ed.2d 684 (1982) (refusal of prosecutor to give reasons for not entering into stipulation regarding admissibility of polygraph test results in state court prosecution was violation of defendant's due process rights); United States v. Galbreth, 908 F. Sup. 877 (D.N.M. 1995) (admitting polygraph evidence, although Tenth Circuit cases prior to Daubert held that polygraph evidence is inadmissible); Ulmer v. State Farm Fire Casualty Co., 897 F. Sup. 299 (D.La. 1995) (admitting polygraph evidence after finding that Daubert must be applied); United States v. Crumby, 895 F. Sup. 1354 (D.Ariz. 1995) (accepting Daubert as proper standard for admissibility of polygraph evidence and allowing admission of polygraph evidence under certain conditions); United States v. Lech, 895 F. Sup. 582 (S.D.N Y 1995) (assuming that Daubert applies to polygraph evidence but holding under facts of case that evidence not admissible); United States v. Scheffer, 44 M.J. 442, 445-46 (C.M.A. 1996), cert. granted, 523 U.S. ___ (65 U.S.L.W. 3761, May 20, 1997) (holding that military rule of evidence excluding polygraph evidence, under the circumstances of that case, violated defendant's sixth amendment right to present a defense); United States v. Mobley, 44 M.J. 453, 454-45 (C.M.A. 1996) (same, companion case to Scheffer).

Rule 403 of the Federal Rules of Evidence provides: "Exclusion of Relevant Evidence on Grounds of Prejudice, Confusion, or Waste of Time
"Although relevant, evidence may be excluded if its probative value is substantially outweighed by the danger of unfair prejudice, confusion of the issues, or misleading the jury, or by considerations of undue delay, waste of time, or needless presentation of cumulative evidence."

In Posado, even the government conceded that the per se rule against admitting polygraph evidence was "no longer viable after Daubert." United States v. Posado, supra, 57 F.3d 432.

See part V of this opinion.

In addition, it appears that the Court of Appeals for the Second Circuit has signaled that it is ready to consider the admissibility of polygraph evidence under Daubert. See United States v. Kwong, 69 F.3d 663, 669 (2d Cir. 1995) ("[t]he record before us simply does not provide the proper opportunity to explore the validity of polygraph evidence"). Likewise, the Court of Appeals for the Fourth Circuit has indicated that it is ready to reconsider the admissibility of polygraph evidence; see United States v. Toth, 91 F.3d 136 (4th Cir. 1996) (decided without published opinion), No. 95-5191, 1996 U.S. App. LEXIS 19017, *12-14 (implying that, in appropriate case, polygraph evidence will be reconsidered in light of Daubert, Fifth Circuit's decision in Posado, and Eleventh Circuit's decision in Piccinonna); and the Eighth Circuit has also implied that it will reconsider the admissibility of polygraph evidence. See United States v. Williams, 95 F.3d 723, 728-30 (8th Cir. 1996) (implying that nonstipulated polygraph evidence may be admissible under Daubert, but that District Court has discretion to exclude the evidence under rule 403).

The Eighth Circuit also allows stipulated polygraph evidence.

The only federal circuit in which the Court of Appeals, or the district courts within the circuit, have not abrogated or rejected a per se rule is the Court of Appeals for the District of Columbia Circuit. United States v. Skeens, 494 F.2d 1050, 1053 (D.C. Cir. 1974). The decision in Skeens, however, dates back to 1974, more than eighteen years before Daubert, and that, of course, was the Circuit Court of Appeals that originally promulgated the Frye test back in 1923. See id., 1053 ("The leading case in this Circuit [regarding the admissibility of polygraph tests] is Frye. . . . This case has been followed uniformly in this and other Circuits and there has never been any successful challenge to it in any federal court." [Citation omitted.]). In addition, the First Circuit, in United States v. Black, 78 F.3d 1, 7 (1st Cir. 1996), stated that "the results of polygraph examinations are generally inadmissible. . . ." (Citation omitted; emphasis added.) This appears not to be a per se rule, but, notwithstanding, Black relied on a Sixth Circuit case for the proposition that polygraph test results are generally inadmissible. See id., citing United States v. Scarborough, 43 F.3d 1021, 1026 (6th Cir. 1994). Even the Sixth Circuit, however, as noted previously, no longer follows that rule. United States v. Sherlin, supra, 67 F.3d 1216-17.

Thus, six federal Circuit Courts of Appeals (Fifth, Sixth, Seventh, Ninth, Eleventh and the Court of Appeals for the Armed Forces) and federal district courts in two other circuits (Second and Tenth) no longer apply a per se rule excluding polygraph evidence; three federal Circuit Courts of Appeals (Second, Fourth and Eighth) have either indicated a willingness to review the admissibility in a proper case or implicitly indicated that it was admissible. The Court of Appeals for the Third Circuit has not addressed the issue of polygraph admissibility. As the Court of Appeals for the Fourth Circuit just recently pointed out, "circuits that have not yet permitted evidence of polygraph results for any purpose are now in the decided minority." (Internal quotation marks omitted.) United States v. Toth, supra, 1996 U.S. App. LEXIS 19017, *13.

I have two observations with respect to the case law that the majority relies upon. First, with respect to the state jurisdictions, the vase number of the cases were decided before Daubert. Second, the majority downplays how the federal courts have dealt with Daubert in the context of polygraph evidence. Although the majority today embraces Daubert, it chooses to ignore the wave of compelling federal precedent, following in the wake of Daubert, holding that the admissibility of polygraph evidence must be decided by trial courts on a case-by-case basis.

In addition, state courts have ruled that polygraph evidence is admissible in varying situations. See, e.g., State v. Catanese, 368 So.2d 975, 981-83 (La. 1979) (polygraph evidence admissible in posttrial, and possibly pretrial, proceedings); Conner v. State, 632 So.2d 1239, 1257-59 (Miss. 1993) (evidence of willingness to take polygraph examination admissible to rehabilitate impeached witness' credibility); State v. Sanders, 117 N.M. 452, 872 P.2d 870, 877 (1994) (polygraph evidence admissible under state rules of evidence previously enacted by state Supreme Court); State v. Wright, 322 S.C. 253, 255, 471 S.E.2d 700 (1996) ("[t]he admission of evidence in a criminal prosecution is within the discretion of the trial judge and his ruling will not be disturbed on appeal unless abuse of discretion is shown"). Finally, many other jurisdictions allow polygraph evidence when its admissibility is stipulated to in advance by the defendant and the state.

At one time, Massachusetts admitted polygraph evidence without stipulation. Commonwealth v. A Juvenile, 365 Mass. 421, 425-26, 313 N.E.2d 120 (1974). In 1989, however, the Supreme Judicial Court of Massachusetts abandoned the position it had adopted in that case and ruled polygraph evidence inadmissible in criminal trials. Commonwealth v. Mendes, 406 Mass. 201, 212, 547 N.E.2d 35 (1989). In the opinion in Mendes, Chief Justice Liacos stated that "[t]he court today rushes headlong into the wholesale rejection of a carefully crafted set of rules in this Commonwealth governing the admissibility of polygraph evidence." Id., 212. He further stated that "[w]ith proper oversight and the appropriate use of judicial discretion by trial judges, the polygraph can serve as an effective tool in the truth-seeking process." Id., 216 (Liacos, C.J., dissenting).

"Of the state jurisdictions, New Mexico has gone the furthest in admitting polygraph evidence. In State v. Dorsey, [ 88 N.M. 184, 539 P.2d 204 (1975)] the New Mexico Supreme Court held that polygraph results were admissible if (1) the operator is qualified; (2) the testing procedures were reliable; and (3) the test of the particular subject was valid. [Id., 184-85.]" P. Giannelli, "Forensic Science: Polygraph Evidence: Part II," 30 Crim. L. Bull. 366, 373 (1994). "Several subsequent New Mexico appellate decisions have addressed procedural details flowing from the general admission authorized in Dorsey. . . . These holdings were incorporated in New Mexico Rule of Evidence 11-707, a comprehensive polygraph examination rule adopted by the Supreme Court of New Mexico in 1983." (Citations omitted.) J. McCall, "Misconceptions and Reevaluation — Polygraph Admissibility after Rock and Daubert," 1996 U. Ill. L. Rev. 363, 386 (1996).
"The rule changed the prior law in New Mexico by limiting admission of polygraphs to purposes of impeaching or corroborating the testimony of a witness, and by requiring thirty-day written notice of a party's intention to offer polygraph evidence at trial. Additionally, the following features of the rule set out clear cut solutions to four major concerns about the admission of polygraph results: a set of minimum qualifications a polygrapher must meet to give testimony about a polygraph test result in New Mexico court, a recording requirement for all aspects of the polygraph examination, a prohibition on compelling any witness to take a polygraph examination, and a set of protocol requirements for a polygraph examination that must be followed if the results of the test are to be admitted in New Mexico courts." Id., 387-88.

See footnote 69 of the majority opinion.

IV

The majority acknowledges — by way of an assumption — that polygraph evidence "may have enough demonstrated validity to pass the Daubert threshold for admissibility." The majority then illogically decides, as a matter of law, that the prejudicial impact outweighs its probative value. Notwithstanding our adoption of Daubert in this case, the majority then completely ignores the underpinnings of Daubert. In Daubert, the United States Supreme Court held: "Faced with a proffer of expert scientific testimony . . . the trial judge must determine at the outset . . . whether the expert is proposing to testify to (1) scientific knowledge that (2) will assist the trier of fact to understand or determine a fact in issue." (Citation omitted; emphasis added.) Daubert v. Merrell Dow Pharmaceuticals, Inc., supra, 509 U.S. 592; see also United States v. Cordoba, supra, 104 F.3d 228 (pursuant to Daubert, it is trial judge's task, rather than appellate level court's task, to weigh probative value against prejudicial effect).

Although the determination of the admissibility of scientific evidence in the first instance is a question for the trial court, I must comment on the majority's finding that, as a matter of law, the prejudice of polygraph evidence outweighs its probative value. Applying what the majority formulates as our general standard for the admissibility of evidence — whether the prejudicial impact outweighs its probative value — the majority concludes that polygraph evidence is Notre prejudicial for essentially three reasons: (more) it lacks accuracy and, therefore, is of limited probative value; (2) it is time-consuming; and (3) it invades the jury's fact-finding mission. I find none of these reasons sufficient to support a rule that directs our trial courts to omit polygraph evidence under all circumstances, in civil as well as criminal trials, regardless of whether the omission is stipulated to in advance by both the state and the defendant in a criminal case or by the plaintiff and the defendant in a civil case.

See part II C of the majority opinion.

First, as to its accuracy, the courts and the literature have indicated that a properly administered polygraph examination is quite accurate. In fact, the majority concedes that there are impressive statistics as to its accuracy, based upon a very recent study. The majority opinion indicates that David Raskin, a noted authority, reviewed laboratory studies (simulated studies) that indicate that the polygraph has an 89 percent sensitivity rate (correctly labels a deceptive subject as being deceptive 89 percent of the time) and 91 percent specificity rate (correctly labels a truthful subject as being truthful 91 percent of the time), whereas the field studies (based on data from polygraph examiners in actual criminal cases) indicate that the polygraph has an 87 percent sensitivity rate and 59 percent specificity rate. D. Raskin, "The Scientific Status of Research on Polygraph Techniques," in West Companion to Scientific Evidence 2 (Faigman et al. eds., forthcoming 1996), cited in C. Honts B. Quick, "The Polygraph in 1995: Progress in Science and the Law," 71 N.D. L. Rev. 987, 995, 1018-19 (1995); see also D. Raskin, "Hoffmann, Hypnosis, and the Polygraph," 3 Utah B.J. 7, 8 (1990) ("[E]xtensive scientific data have clearly demonstrated that polygraph examinations that are properly conducted in appropriate situations have an accuracy rate that exceeds 90 percent. . . . Major advances in instrumentation, improvements in examination procedures, and our development of computerized methods for interpreting the outcome of polygraph tests . . . have produced accuracy rates of approximately 95 percent in federal criminal investigations. . . ." [Citations omitted.]); A. Moenssens, F. Inbau J. Starrs, Scientific Evidence in Criminal Cases (3d Ed. 1986) § 14.09, p. 712 ("when the technique is properly applied by a trained, competent examiner, it is very accurate in its indications, with a known error percentage of less than one percent"); D. Raskin, "The Polygraph in 1986: Scientific, Professional and Legal Issues Surrounding Application and Acceptance of Polygraph Evidence," 1986 Utah L. Rev. 29, 72 (1986) ("[t]he existing literature suggests an accuracy of 90 [percent] or higher when examinations are conducted to assess the credibility of suspects in criminal investigations and other specific incidents"). In addition to Raskin's most recent study, the majority also concedes that "other studies indicate higher levels of accuracy," citing generally to P. Giannelli, "Forensic Science: Polygraph Evidence: Part I," 30 Crim. L. Bull. 262, 271-73 (1994), and 1 C. McCormick, Evidence (4th Ed. 1992) § 206, pp. 909-10.

See part II B 2 a of the majority opinion.

The majority, in part II B 2 b of its opinion, discusses "predictive values" based on the polygraph's sensitivity and specificity. Based upon Raskin's most recent field-derived figures of 87 percent sensitivity and 59 percent specificity, the majority calculates that "we should only be 68 percent confident that a subject really is lying if the subject fails a polygraph exam, and only 82 percent confident that the subject is being truthful if the subject passes." Using only Raskin's laboratory-derived figures, the majority calculates even higher "predictive values."
In the present case, using the majority's figures, we can be 82 percent confident that the defendant was being truthful when he passed his polygraph examination, an examination where he was asked whether he set fire to his house or if he knew if anyone else had purposely set the fire. See part I of this opinion. With this high degree of certainty, not to allow this evidence is simply unconscionable.
With respect to the truth-seeking ability of the polygraph, I find the "predictive value" percentages calculated by the majority compelling enough to show that a per se rule is unwarranted, especially in light of the significant debate about the accuracy of the polygraph. The trial courts in the first instance must conduct case-specific inquiries regarding offered polygraph evidence, in order to screen out incompetent polygraph practioners and improperly conducted polygraph examinations. See footnotes 15, 23 and 28 of this opinion. As the Fifth Circuit Court of Appeals aptly stated, "[r]emaining controversy about test accuracy is almost unanimously attributed to variations in the integrity of the testing environment and the qualifications of the examiner." United States v. Posado, supra, 57 F.3d 434.

In concluding that the traditional reasons for the per se rule are unpersuasive, one commentator has stated: "The most frequently mentioned [criticism] is that the technique is `unreliable' due to inherent failings, a shortage of qualified operators, and the prospect that `coaching' and practicing would become commonplace if the evidence were generally admissible. Yet, by themselves, such doubts are not sufficient to warrant a rigid exclusionary rule. A great deal of lay testimony routinely admitted is at least as unreliable and inaccurate, and other forms of scientific evidence involve risks of instrumental or judgmental error." 1 C. McCormick, supra, § 206, pp. 914-15. Indeed, with respect to psychiatric testimony used to predict a defendant's future potential for dangerous behavior, the United States Supreme Court has stated, despite the unanimous expert opinion that such conclusions were accurate only one out of three times; Barefoot v. Estelle, 463 U.S. 880, 920-22, 103 S.Ct. 3383, 77 L.Ed.2d 1090 (1983) (Blackmun, J., dissenting); that "[a]ll of these professional doubts about the usefulness of psychiatric predictions can be called to the attention of the jury. Petitioner's entire argument, as well as that of Justice Blackmun's dissent, is founded on the premise that a jury will not be able to separate the wheat from the chaff. We do not share in this low evaluation of the adversary process." Id., 899-901 n. 7.

Many of the arguments against the admissibility of polygraph evidence center around undesirable and incompetent examiner practices. "Polygraph examiners in the United States, as a whole, are poorly trained. Their techiques lack standardization, and polygraph tests are subject to manipulation by unethical examiners. Given this state of affairs, the probative value of the work-product of most of the polygraph professionals should rightly be questioned. . . . [M]any of the problems associated with examiner practices could be exposed and/or controlled by an evidentiary requirement that all polygraph tests offered for admission be tape recorded in their entirety. However, to date only New Mexico has taken this step to assure the quality of polygraph tests offered for evidence. Problems of attorneys shopping for a particular test outcome could be controlled by an evidentiary requirement that if one test is offered, all polygraph tests conducted on a particular subject must be disclosed. Again, only New Mexico has taken this simple and seemingly logical step." C. Honts M. Perry, "Polygraph Admissibility: Changes and Challenges," 16 Law Hum. Behav. 357, 375-76 (1992).
"The wealth of examples provided by actual cases shows that a great deal of questionable evidence is produced by incompetent, poorly trained and unethical polygraph examiners. . . . To reduce these problems, the court could appoint polygraph experts after consultation with all parties. That would minimize the pressures placed on experts hired by one party and maximize the likelihood of screening out so-called experts who are incompetent or lack integrity." D. Raskin, supra, 1986 Utah L. Rev. 71.

Second, the claims that it will be time-consuming for our courts can be disposed of in short shrift. The need to conserve precious court time pales in significance when a person's liberty or life is at stake. Surely, we cannot be that callous to a person who may be innocent. Any problems with polygraph evidence can be overcome through rules developed by this court regarding its admissibility, the availability of a polygraph examination conducted by the state, cross-examination and jury instructions.

The United States Court of Appeals for the Armed Forces answered the "floodgate" argument that their decision regarding polygraph evidence would generate an unreasonable burden on the military courts as follows: "Apart from the speculative nature of such an argument, we think that it is just as likely that polygraph evidence will prevent needless litigation by avoiding some meritless prosecutions as well as smoking out bogus [defenses]. . . . [O]ur measure should be the scales of justice, not the cash register." United States v. Scheffer, supra, 44 M.J. 448.

The United States Supreme Court, in Daubert, rejected an argument that the abandonment of the "general acceptance" standard for the admissibility of scientific evidence will "result in a "free-for-all" in which befuddled juries are confounded by absurd and irrational pseudoscientific assertions. [ This argument is] overly pessimistic about the capabilities of the jury and of the adversary system generally. Vigorous cross-examination, presentation of contrary evidence, and careful instruction on the burden of proof are the traditional and appropriate means of attacking shaky but admissible evidence." (Emphasis added.) Daubert v. Merrell Dow Pharmaceuticals, Inc., supra, 509 U.S. 595-96.
Arguments that polygraph evidence tends to mislead or confuse the jury are generally not supported in the literature. "Studies tend to show that juries are more inclined not to give extraordinary weight to polygraph evidence. Field and laboratory studies have consistently supported the notion that juries are capable of weighing and evaluating all evidence and rendering verdicts that may be inconsistent with polygraph results. . . .
"Several theories exist as to why polygraph evidence doesn't mislead or confuse a jury: jurors distrust scientific evidence, cross-examination brings out the weaknesses in polygraph evidence giving jurors the ability to give polygraph tests . . . the weight they are due, or there may be other significant evidence to support either a conviction or acquittal (or finding for [the] plaintiff or [the] defendant in a civil action) that simply outweighs the impact of the polygraph evidence. . . . Despite the logic and research to support these theories, none seem sufficient to persuade many courts from what appears to be a disparaging, and somewhat condescending, perspective of the abilities of American juries." C. Honts M. Perry, "Polygraph Admissibility: Changes and Challenges," 16 Law Hum. Behav. 357, 366-67 (1992); see also D. Raskin, supra, 1986 Utah L. Rev. 66 ("The available [empirical] evidence indicates that testimony by polygraph experts has no detrimental effects on the trial process, nor does it have an undue influence on the trier of fact. Since 1975 New Mexico has admitted polygraph evidence without a stipulation between the parties. Ten years of experience [1975 to 1985] has failed to reveal any inherent problems with that type of evidence."); R. Peters, "A Survey of Polygraph Evidence in Criminal Trials," 68 A.B.A. J. 162, 165 (1982) (on basis of empirical study conducted in Wisconsin, one commentator has indicated that "[t]he actual trial results clearly support the belief that juries are capable of weighing and evaluating all evidence and rendering verdicts that may be inconsistent with the polygraph evidence").

The greatest tragedy in our judicial system, with all its fallibility and misidentifications, would be to take away the life or liberty of one innocent person for the sake of preventing guilty persons from escaping punishment. See 4 W. Blackstone, Commentaries on the Laws of England (1769) c. 27, p. 352 ("it is better that ten guilty persons escape, than that one innocent suffer").

Finally, the majority curiously rejects polygraph evidence on the ground that it invades the fact-finding province of the jury. There are three answers to that argument. First, any reliable device that can aid the jury in its truth seeking mission should be available under rules and standards set forth by this court. As stated by the United States Court of Appeals for the Armed Forces: "We believe that the truth-seeking function is best served by keeping the door open to scientific advances." United States v. Scheffer, supra, 44 M.J. 446. Second, as we approach the twenty-first century, "science, for better or for worse, has become more a part of our daily lives. Scientific evidence, in turn, has become more a part of the ordinary trial so that jurors may be more likely to use polygraph evidence with discretion." McMorris v. Israel, supra, 643 F.2d 462. Third, we allow the prosecution to introduce expert testimony, which is far less reliable than the polygraph, to bolster the credibility of the state's case in other situations. See, e.g., State v. Ali, 233 Conn. 403, 432, 660 A.2d 337 (1995) (expert allowed to testify as to typical behavior patterns of victims of sexual assault); State v. Freeney, 228 Conn. 582, 592, 637 A.2d 1088 (1994) (same); State v. Borrelli, supra, 227 Conn. 173-74 (expert allowed to testify as to typical behavior patterns of victims of battered women's syndrome); State v. Spigarolo, 210 Conn. 359, 378-79, 556 A.2d 112, cert. denied, 493 U.S. 933, 110 S.Ct. 322, 107 L.Ed.2d 312 (1989) (expert allowed to testify that it is not unusual for child sexual abuse victim to give inconsistent or incomplete statements).

Indeed, this court recently allowed expert testimony to bolster the testimony of a child sexual abuse victim where there was a prolonged delay in the disclosure of the abuse to the authorities. See State v. Christiano, 228 Conn. 456, 459-63, 637 A.2d 382, cert. denied, 513 U.S. 821, 115 S.Ct. 83, 130 L.Ed.2d 36 (1994). In fact, in the court's opinion in that case, authored by Justice Borden, we observed that "[t]he admissibility of opinion testimony of expert witnesses is a matter within the discretion of the trial court. . . . Generally, expert testimony is admissible if (more) the witness has a special skill or knowledge directly applicable to a matter in issue, (2) that skill or knowledge is not common to the average person, and (3) the testimony would be helpful to the court or jury in considering the issues." (Citation omitted; internal quotation marks omitted.) Id., 461.

Although some of the previously noted cases can be distinguished because the expert testimony was admitted for the purpose of establishing the behavior patterns of the victims, I see little difference in the distinction. Nevertheless, that distinction has little substance because in those situations the evidence is submitted to bolster the credibility of the state's case, the state's witness, or both.

V

Finally, by not allowing the defendant an opportunity to have a hearing before the trial court in order to demonstrate the reliability of the polygraph evidence, he is deprived of his right to present a defense under the sixth amendment to the federal constitution and article first, § 8, of the state constitution. The United States Supreme Court has made it clear that the right of an accused to present testimony that is relevant and material may not be denied arbitrarily. Washington v. Texas, 388 U.S. 14, 23, 87 S.Ct. 1920, 18 L.Ed.2d 1019 (1967).

The illogic of the majority's opinion is highlighted in its response to this concurring and dissenting opinion, where the majority indicates that a per se rule does not violate a defendant' constitutional right to present a defense. The majority states that "as a matter of logic, there can be no constitutional right to an evidentiary hearing regarding evidence that, as a matter of law, is inadmissible." See footnote 80 of the majority opinion. In other words, based on the majority's circular reasoning, the defendant's right to present a defense can be subverted merely by the court deciding, as a matter of state law, that exculpatory evidence is not admissible. Not only is that reasoning illogical, it is also devious.

In Rock v. Arkansas, 483 U.S. 44, 107 S.Ct. 2704, 97 L.Ed.2d 37 (1987), the United States Supreme Court reversed the conviction of a defendant by holding that her sixth amendment right under the federal constitution was violated when she was deprived of the right to call herself as a witness to testify on her own behalf, regardless of the fact that she needed to be hypnotically refreshed and the state's law did not allow this form of testimony. " A State's legitimate interest in barring unreliable evidence does not extend to per se exclusions that may be reliable in an individual case. . . . The State would be well within its powers if it established guidelines to aid trial courts in the evaluation of posthypnosis testimony and it may be able to show that testimony in a particular case is so unreliable that exclusion is justified." (Emphasis added.) Id., 61. "Despite the unreliability that hypnosis concededly may introduce . . . the procedure has been credited as instrumental in obtaining investigative leads or identifications that were later confirmed by independent evidence. . . . The inaccuracies the process introduces can be reduced, although perhaps not eliminated, by the use of procedural safeguards." (Citations omitted.) Id., 60.

Some of the procedural safeguards for the admission of hypnotically refreshed testimony are: (1) a qualified person performed the hypnosis; (2) an audiotape or videotape was made before, during and after the hypnosis was conducted; (3) cross-examination to reveal inconsistencies is allowed; and (4) the jury is educated about the risks of hypnotically refreshed testimony through the use of opposing expert testimony and cautionary instructions. Rock v. Arkansas, supra, 483 U.S. 60-61. These same safeguards, among others, can be use in the admission of polygraph testimony.

On the basis of Rock, one commentator has pointed out that a per se exclusionary rule regarding polygraph evidence violates a defendant's constitutional right to present a defense. "A recent [United States] Supreme Court interpretation of the constitutional right of a defendant in a criminal prosecution to call witnesses to testify in his or her behalf requires a reconsideration of the denial position when a criminal defendant attempts to introduce exculpatory polygraph evidence. Beginning with its earlier opinions in Chambers v. Mississippi [ 410 U.S. 284, 93 S.Ct. 1038, 35 L.Ed.2d 297 (1973)] and Green v. Georgia, [442 U.S. 96, 99 S.Ct. 2150, 60 L.Ed.2d 738 (1979)] the Supreme Court started the development of a constitutional right to introduce exculpatory evidence in criminal trials. The Court's 1987 opinion in [ Rock] established the basic components of that right. . . . The Court stated in dicta that reasonable restrictions on the right to call a witness to give posthypnotic recall, as opposed to a per se prohibition, would be constitutional." J. McCall, "Misconceptions and Reevaluation — Polygraph Admissibility after Rock and Daubert," 1996 U. Ill. L. Rev. 363, 392-93 (1996).

"The holding in Rock v. Arkansas [supra, 483 U.S. 44] should be significant in the reconsideration of polygraph evidence because it could dispel the long-standing judicial hostility toward such evidence. This negative attitude is akin to that historically taken toward hypnotically refreshed testimony, and Rock teaches that the Sixth Amendment requires courts to give more positive treatment toward a previously hypnotized witness. If the putative Sixth Amendment light of a defendant to present reliable polygraph evidence is acknowledged, courts must consider polygraph evidence with a different attitude." J. McCall, supra, 1996 U. Ill. L. Rev. 406. "[T]he important points from Rock on this issue are that the denial position probably violates the Sixth Amendment and that sophisticated judicial and legislative efforts should now be made to limit admission of polygraph evidence so that only the most reliable evidence of this type is admitted." Id., 406 n. 309.

"The right to present defense evidence was also cited in McMorris v. Israel, [supra, 643 F.2d 458] in which the defendant offered to stipulate to the admission of a polygraph examination. Although stipulated polygraph results were admissible under state law at that time, the prosecutor, without offering any reasons, refused to stipulate. . . . The court [in granting habeas corpus relief to the defendant] . . . rested its decision on [narrow] grounds; that is, the prosecution's refusal to stipulate without offering a valid ground for the refusal deprived the defendant of due process. . . ." P. Giannelli, "Forensic Science: Polygraph Evidence: Part II," 30 Crim. L. Bull. 366, 375 (1994).

Just recently, in United States v. Scheffer, supra, 44 M.J. 445-46, the Court of Appeals for the Armed Forces held, notwithstanding the military rule of evidence with respect to the inadmissibility of polygraph evidence, that such evidence could not be excluded in a situation where the defendant testified, placed his credibility in issue, and was accused by the prosecution of being a liar. "A per se exclusion of polygraph evidence offered by an accused to rebut an attack on his credibility, without giving him an opportunity to lay a foundation under [the military rule of evidence regarding scientific testimony] and Daubert, violates his Sixth Amendment right to present a defense." Id., 445. In Scheffer, the court relied on Rock for the proposition that per se exclusions of evidence that may be reliable in an individual case cannot stand; Rock v. Arkansas, supra, 483 U.S. 61; and stated that "[w]hile Rock concerned exclusion of a defendant's testimony and this case concerns exclusion of evidence supporting the truthfulness of a defendant's testimony, we perceive no significant constitutional difference between the two. In either case, the Sixth Amendment right to present a defense is implicated." United States v. Scheffer, supra, 446; see also United States v. Mobley, supra, 44 M.J. 454-45. Likewise, in the present case, the defendant's right to present a defense is also implicated.

With respect to the defendant's right to lay the foundation for the admission of the polygraph evidence, the Scheffer court stated: "[W]e, like the Fifth Circuit, cannot determine `whether polygraph technique can be said to have made sufficient technological advance in the seventy years since Frye to constitute the type of "scientific, technical, or other specialized knowledge" envisioned by Rule 702 [of the Federal Rules of Evidence] and Daubert.' United States v. Posado, [supra, 57 F.3d 433]. We will never know, unless we give [the defendant] an opportunity to lay the foundation." United States v. Scheffer, supra, 44 M.J. 446.

Although the defendant did not testify in this case, his credibility was at issue as a result of the introduction by the state of his exculpatory out-of-court statement. Furthermore, since the decision to exclude the polygraph evidence was made in a ruling on a pretrial motion, there was no need for the defendant to offer his own testimony as a predicate to raising the question of whether he was entitled to a hearing on the issue of admissibility.

VI

In conclusion, I would remand this case to the trial court in order to furnish the defendant with an opportunity to prove the validity of his polygraph examination and to demonstrate that it was correctly administered in this case. Both the state and the defendant should have the opportunity to be heard on the issue of accuracy and the conditions for the admissibility of the polygraph evidence. Only then do I believe that a court would be in the position to rule on its admissibility. "Where credibility is as critical as in the instant case, the circumstances are such as to make the polygraph evidence materially exculpatory within the meaning of the [federal] [c]onstitution." McMorris v. Israel, supra, 643 F.2d 462.

Accordingly, I dissent with respect to the issue of the admissibility of polygraph evidence.


I write this separate dissenting opinion based upon what occurred after this court released its opinion in this case on May 20, 1997. See State v. Porter, 241 Conn. 57, 698 A.2d 739 (1997). In part I of the court's opinion, in which I concurred, we unanimously agreed to abandon the Frye test for the admissibility of scientific evidence in order to adopt the standards under the Federal Rules of Evidence as interpreted in Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579, 113 S.Ct. 2786, 125 L.Ed.2d 469 (1993). State v. Porter, supra, 75-76. Daubert did not formulate a court-made rule, but merely interpreted the federal rules. As its starting point, the court in Daubert noted that rule 402 of the Federal Rules of Evidence provides the "baseline" for interpreting the federal rules. Daubert v. Merrell Dow Pharmaceuticals, Inc., supra, 587. Rule 402 provides that "[a]ll relevant evidence is admissible, except as otherwise provided by the Constitution of the United States, by Act of Congress, by these rules, or by other rules prescribed by the Supreme Court pursuant to statutory authority. Evidence which is not relevant is not admissible." With respect to expert testimony, Daubert relied on rule 702 of the Federal Rules of Evidence, which provides: "If scientific, technical, or other specialized knowledge will assist the trier of fact to understand the evidence or to determine a fact in issue, a witness qualified as an expert by knowledge, skill, experience, training, or education, may testify thereto in the form of an opinion or otherwise." With these rules in mind, the court in Daubert outlined certain "considerations" in order to guide trial courts when considering, as part of their "gatekeeper" role, whether to admit proffered scientific evidence. Interrelated with rule 702, as Daubert points out, is rule 403, which "permits the exclusion of relevant evidence `if its probative value is substantially outweighed by the danger of unfair prejudice, confusion of the issues, or misleading the jury. . . .'" (Emphasis added.) Daubert v. Merrell Dow Pharmaceuticals, Inc., supra, 595. Accordingly, in part I of our original opinion, we not only adopted the "considerations" that the court in Daubert delineated under the auspices of rule 702, but we also made clear that the admissibility of scientific evidence was subject to the limitations of rule 403 by stating that "scientific evidence, like all evidence, is properly excluded if its prejudicial impact substantially outweighs its probative value, even if it is otherwise admissible. See, e.g., Daubert v. Merrell Dow Pharmaceuticals, Inc., supra, [595]. . . ." (Citations omitted; emphasis added.) State v. Porter, Connecticut Law Journal, Vol. 53, No. 47, p. 92 (May 20, 1997).

Frye v. United States, 293 F. 1013 (D.C. Cir. 1923).

On June 24, 1997, the majority of this court amended the original version of the Porter opinion by deleting the word "substantially" from the previously quoted material when it substituted the original page 90 with a replacement page. See State v. Porter, supra, 241 Conn. 90. The rule that the court now adopts in Porter is that such scientific evidence can be excluded from evidence, after passing Daubert's threshold considerations for admissibility, if "its prejudicial impact outweighs its probative value"; id.; a rule neither party advocated.

The replacement pages appear in the Connecticut Law Journal, Vol. 53, No. 52, pp. iii-iv (June 24, 1997). The original version of the page, without the change, appears in the Connecticut Law Journal, Vol. 53, No. 47, p. 92 (May 20, 1997).

I neither question that this court has the authority to hone the rules of evidence interpreted by Daubert, in order to fit the needs of our Connecticut jurisprudence, nor do I necessarily believe that the word "substantially" should not have been deleted from the formulation in Porter of the prejudicial impact versus probative value test. I believe, however, that the decision to make such a substantial change in the opinion should have been made only in the context of the advocacy of a case, so that this court could act with the input of the parties, as well as of others who have a potential interest.

Provided, however, that the court does not trample on constitutional rights.

Furthermore, I note that although our prior case law has not been consistent, this court has stated in the past that relevant evidence is to be excluded only if it is "unduly prejudicial." See State v. Jeffrey, 220 Conn. 698, 707, 601 A.2d 993 (1991) (holding that admitted evidence "had probative value and that its admission into evidence was not unduly prejudicial to the defendant and was, consequently, a proper exercise of the trial court's discretion" [emphasis added]); see also State v. Crafts, 226 Conn. 237, 255, 627 A.2d 877 (1993) ("The defendant's final objection to the admission of the statements is that the testimony should have been excluded because it was more prejudicial than probative. The test for determining whether evidence is unduly prejudicial is not whether it is damaging to the defendant but whether it will improperly arouse the emotions of the jury." [Emphasis added.]); State v. Santiago, 224 Conn. 325, 339-40, 618 A.2d 32 (1992) (enumerating situations in which "unduly prejudicial" effect of relevant evidence would counsel its exclusion); State v. Rinaldi, 220 Conn. 345, 356, 599 A.2d 1 (1991) (same).

Because we should leave the decision of whether to delete the word "substantially" from our rule, with respect to the admissibility of scientific evidence, to such time when that issue is properly raised within the context of a case or controversy, I disagree with the change in the original opinion of this court.


I concur in Justice Berdon's dissent concerning a substantial change in this court's May 20, 1997 opinion. State v. Porter, 241 Conn. 57, 164, 698 A.2d 739 (1997). The change should not have been made outside the context of the advocacy of the case.

I concurred in the original opinion because I believe that the Daubert rule, under which scientific evidence is excluded only when its prejudicial impact substantially outweighs its probative value, should be our rule. I continue in that belief.

Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579, 113 S.Ct. 2786, 125 L.Ed.2d 469 (1993).