DUNCAN S. McNAB, J.

Defendant Wilbur Vinson, along with codefendants Michael Orgill and Danny Terry, stands charged with having acted in concert to commit first degree robbery, based on an alleged stick-up at Joe's Grocery Store in Mt. Vernon, New York, on July 4, 1979. Defendant denies any involvement in the alleged crime, contending he has been misidentified. He proffers an alibi defense, namely, that he was home with his mother at the time in question. Additionally, defendant points to a signed written confession made by codefendant Danny Terry on July 18, 1979, wherein Terry exculpated defendant and named two others, John McShaw and Peter Queen, as his cohorts. Against this background, defendant seeks to offer at trial the results of a polygraph examination which he passed with a score of ± 26, far above the minimal passing grade of ± 6 under the numerical scoring system used in the United States Army. Defendant's application is largely based on the decision of my colleague, the Honorable ADOLPH ORLANDO, in People v Daniels ( 102 Misc.2d 540), decided November 27, 1979, which, to this court's knowledge, for the first time in this State, permitted the results of the so-called lie detector test to be admitted at trial.

At defendant's request, a pretrial hearing was held to afford defendant an opportunity to present testimony on the admissibility of the polygraph. Defendant called three expert witnesses, Dr. Gordon Barland, Ph.D., of Salt Lake City, Utah; Mr. Walter Atwood, a self-employed examiner formerly affiliated with the National Security Agency; and Mr. Victor Kaufman, founder and chief examiner of the New York Lie Detection Laboratories who actually administered the test to the defendant. While their testimony will not be rehashed here in detail, suffice it to say that Dr. Barland and Mr. Atwood generally discussed the workings of the polygraph machine, the various physiological responses it is designed to measure, and such refinements in technique as have taken place over the last 30 to 40 years. They also discussed the vast array of variable factors which the examiner should consider in evaluating any test results, such as the intelligence and behavior of the subject, his educational and cultural background, and the nature of the questions propounded to him, among other items.

Insofar as Victor Kaufman is concerned, the parties agreed to limit questioning largely to the actual test administered to this defendant, in view of the fact that Kaufman had previously been examined in the Daniels case. He testified, in substance, that he utilized the so-called "control question" technique in administering the four-pen Stoelting polygraph upon the defendant. This test, first propounded by Backster in 1962 and indorsed both by Kaufman and the United States Army as the most accurate procedure, consists of three types of questions, i.e., (1) so-called neutral (or irrelevant) questions, (2) control questions, those of great concern to the subject, but not pertinent to the matter at issue (i.e., "prior to July 4, 1979 did you ever commit a serious crime?"), and (3) relevant questions, those actually relating to the situation at hand (i.e., "did you stick up Joe's Deli?"). Simply stated, the results of a polygraph are determined by comparing the subject's relative physiological changes in response to a relevant question, vis-a-vis his response to a control question. Mr. Kaufman testified that no comparison is made with the reactions to the neutral questions. As indicated earlier, defendant passed his test with a score of ± 26, far above the minimal passing grade of ± 6. Further comment on this control-question technique will be made at a later point in this opinion.

After hearing all the testimony and the arguments of counsel, this court is constrained to deny defendant's application and will exclude the results of the polygraph examination upon trial. The Daniels case notwithstanding, it is this court's opinion that polygraph results, under the overwhelming weight of the cases both before and since Daniels, have not as yet gained the general scientific recognition necessary to introduce them in court. While it is true that "perfection in test results is not a prerequisite to the admissibility of evidence obtainable by the use of scientific instruments, the rule has been to grant judicial recognition only after the instrument has been sufficiently established to have gained general acceptance in the particular field to which it belongs." (People v Leone, 25 N.Y.2d 511, 517.) The Leone court went on to say (supra, at p 518), "We are all aware of the tremendous weight which such tests would necessarily have in the minds of a jury. Thus, we should be most careful in admitting into evidence the results of such tests unless their reasonable accuracy and general scientific acceptance are clearly recognized."

Under this test, a unanimous court in Leone, in 1969, excluded the results of a polygraph. Five years later, in 1974, the Court of Appeals in Pereira v Pereira ( 35 N.Y.2d 301) reaffirmed that the reliability of polygraph tests had still not been sufficiently established to admit such results in court. Significantly, the language used by the Court of Appeals in cases decided subsequent to Pereira has not suggested any abandonment of their holding in Leone. In People v Allweiss ( 48 N.Y.2d 40, 49), decided October 16, 1979, the court, in the course of sustaining the admissibility of certain hair samples, had occasion to make the following comment: "He [the defendant] argues that the results of hair analysis are no more reliable than lie detector evidence which has been held to be inadmissible".

While this statement constitutes only a fleeting reference to polygraphs, the language used in no way suggests any relaxation of the Leone rule. Even more recently, the Court of Appeals has again referred to the polygraph, also by way of dicta, in People v Tarsia ( 50 N.Y.2d 1), decided April 1, 1980, well after the Daniels decision. There, in a lengthy opinion sustaining the admissibility of defendant's inculpatory statement notwithstanding his having been subjected to a test known as a voice stress evaluation test, which all authorities agree is unreliable, the court was prompted to draw certain comparisons to the polygraph. In so doing, the court noted (supra, p 7) that while the polygraph had been "the object of [repeated] scientific testing" for a much longer period of time than the voice stress test, "Yet, attempts to demonstrate that the polygraph is possessed of scientific certainty have been held too indecisive to warrant judicial acceptance".

Moreover, a majority of the Third Department in Tarsia went even further, commenting in their opinion of April 5, 1979 that "Evidence obtained through the use of a polygraph test is generally inadmissible in a criminal case because the test's reliability has not yet been sufficiently established to give it an evidentiary standing in the administration of the criminal law". (See People v Tarsia, 67 A.D.2d 210, 212; emphasis added.)

Thus, in view of these cases decided since Leone, this court is constrained to exclude the polygraph results at trial. In so doing, however, the court would not dispute that certain refinements in the field of polygraph examination have been made over the years, as the expert witnesses testified here, and as the Daniels case correctly pointed out. As recently as 1977, the Second Department, in Matter of McGinigle v Town of Greenburgh ( 59 A.D.2d 908, 910), noted that a polygraph machine, by virtue of being able to permanently record the involuntary responses of four separate body functions, possesses sufficient indicia of reliability to make it a "proper investigative tool." (Emphasis added.)

The court would take this opportunity to note approvingly the wide use made of polygraph examinations by various law enforcement agencies and to encourage the Westchester County District Attorney's office to make increased use of this tool in their efforts to screen out those cases for which formal judicial proceedings may prove to be inappropriate.

Nonetheless, even certain of its proponents concede that significant problems inherent in the administration of the polygraph may serve to keep test results from a jury. In his article "`That's the Man': A Sobering Study of Eyewitness Identification and the Polygraph" (49 St. John's L Rev 1), former Supreme Court Judge Frank O'Connor isolates three major factors, namely (1) the reliability of the instrument itself, (2) the competence of the examiner, and (3) the very real possibility that polygraph test results could serve to encroach upon the fact-finding function. In the instant case as in Daniels, problems associated with the competence of the examiner were minimized, since a reputable and experienced examiner, Mr. Victor Kaufman, administered it. Even so, all the expert witnesses upon the hearing agreed that as a general matter, a great burden is placed on the examiner to ask proper questions and to screen out, or allow for, an assortment of variables which may affect each test to some extent, as alluded to previously. Moreover, as the Court of Appeals recognized in the Pereira case ( 35 N.Y.2d 301, supra) and as Mr. Atwood implicitly conceded on this hearing, even an independent analysis of a subject's chart by a second qualified expert cannot cure an incompetently or improperly administered test.

The actual polygraph machine, and the methodology employed to administer it and evaluate its results, presents a second problem. From the testimony offered at this hearing, and the motion papers and scholarly treatises submitted, it is this court's opinion that since Leone, there has not been any significant improvement in the reliability of the polygraph such as would warrant the admissibility of test results in criminal proceedings. In reality the machines used today, such as the four-pen Stoelting used here, are not all that different from the first Keeler model developed back in 1926. That prototype "channeled", or recorded, three body functions, whereas today's model records four, namely, respiration rate, blood pressure and pulse rate, galvanic skin responses, and muscular activity. Moreover, the so-called "control-question" method of administering the test as discussed earlier, appears to this court to be quite capable of rendering an unreliable result. As Kaufman testified at the hearing, the relative degree of fear exhibited by the subject to these two types of questions is actually what the instrument records. Thus, Kaufman conceded, hypothetically, that in the event the examiner was unaware that the subject had previously committed a murder, for example, as opposed to a petit larceny, then the relative degree of fear supposedly represented by the subject's response could be affected, thereby causing a potentially misleading test result. Hence, while careful question preparation and a thorough pretest interview may serve to minimize this problem, the potential for an unreliable result still exists.

Beyond these technical problems however, there lies perhaps an even more acute problem, i.e., the ever present danger that the introduction of such test results, which unavoidably by their unique nature may be interpreted as answering the ultimate question of guilt or innocence, would, to some degree, necessarily usurp the fact finder's fundamental task of determining whether the witness, defendant or otherwise, is being truthful or not. As noted above, this problem has been recognized in the literature on the subject (see O'Connor, 49 St. John's L Rev 1) as well as in the case law. (See People v Leone, 25 N.Y.2d 511, 518, supra; People v Tarsia, 67 A.D.2d 210, 214, supra [dissenting opn].) It is well settled that it is solely within the province of the jury to determine the credibility of a witness, and the opinion of any expert which in effect vouches for the believability or nonbelievability of a witness improperly usurps that function. (See People v Ciaccio, 47 N.Y.2d 431, and cases cited therein.) This court respectfully disagrees with the suggestion made in People v Daniels ( 102 Misc.2d 540, supra) that careful limiting instructions could serve to overcome this problem. Rather, this court sees no way to prevent a jury from substituting the polygraph result for their reasoned thinking, nor to prevent the proceedings from becoming bogged down in a "minitrial" devoted to a resolution of the potentially conflicting opinions of the polygraph experts and the impact of the myriad of variable factors that may affect the reliability of any test result.

Thus, the court would deny defendant's application to admit the polygraph results both at trial, whether before a jury or the court, and in any and all hearings related thereto, save for the instant hearing, since, as noted above, the "reliability [of polygraph evidence] has not yet been sufficiently established to give it evidentiary standing in the administration of the criminal law". (See People v Tarsia, supra, at p 212; emphasis added.) See People v Guerin ( 47 A.D.2d 788) wherein the Third Department precluded the use of any polygraph test upon a hearing pursuant to CPL 440.10 to vacate the judgment of conviction. But see People v Vernon ( 89 Misc.2d 472 ) where, on the particular facts presented, defendant was permitted to introduce favorable polygraph results upon a Clayton hearing to dismiss in the interest of justice. In view of the breadth of the language used in Tarsia, this court is constrained to follow Guerin (supra), and will not permit any testimony concerning the polygraph in the course of any nonjury trial.

Finally, while not at all determinative of the question before the court, and by way of postscript, the picture defense counsel paints of a potential miscarriage of justice should the test results be excluded is somewhat misleading. First, this is not a case resting on the fleeting eyewitness identification of one witness. Rather, the People have two employees of Joe's Grocery Store who were able to make a legally admissible in-court identification of the defendant based on their midafternoon observations of him at the time in question. Secondly, while codefendant Danny Terry did make a written confession on July 18, 1979 exculpating the defendant, the court's file and testimony during the trial indicates that he later gave a verbal statement that same day which did inculpate the defendant, which he again reiterated during the colloquy during his plea. Thus, given the full background, this case would seem to fall short of the classic "wrong man" scenario urged on the court by defendant.