(dissenting). Petitioner had the right to confront the persons who performed the tests leading to his termination from the Police Department. I, therefore, dissent.
The right to confront witnesses is imbedded in American law. "In almost every setting where important decisions turn on questions of fact, due process requires an opportunity to confront and cross-examine adverse witnesses” (Goldberg v Kelly, 397 US 254, 269). In another context this Court has stated that a decision dispensing with the right of confrontation must be based on a finding of good cause (People ex rel. McGee v Walters, 62 NY2d 317, 322 [parole revocation hearing]). No good cause for the absence of the witnesses was established here.
This Court should not apply a lesser standard of confrontation to a person who has meritoriously served the police force *581for a period of 17 years than it does to a person who has been found guilty of a crime and who faces a parole revocation hearing. The sole evidence against the petitioner was a report which found benzoylecgonine, a cocaine derivative, in his urine. Petitioner had been chosen randomly to undergo the test. The evidence is that he used one of two vials, and when he was unable to use the other, a portion of his urine was poured from the first into the second vial.
In People ex rel. McGee v Walters (supra), this Court stated that when the right of confrontation was dispensed with, the administrative tribunal had to consider (1) its favored status, (2) the nature of the evidence at issue, (3) the potential utility of cross-examination, and (4) the State’s burden in being required to produce the witness (62 NY2d, at 319-320). Each of these standards, applied in this case, is favorable to the petitioner. First, the right of confrontation is favored here as it is in all similar contexts where decisions turn on questions of fact. Second, as to the nature of the evidence, the test is the only evidence against petitioner, is decisive on his career and livelihood and is contrary to an outstanding service record. Third, the potential utility of cross-examination lies not only in any notes or records that the tester may have made as the tests were performed but also in the experience of those individuals with determinations that were unreliable. While some of the evidence attempted to demonstrate how carefully the procedures for both the EMIT test and the GC/MS tests were performed, this evidence was no substitute for having the testimony of the persons who actually performed the tests. This is particularly true where the persons performing the tests were still employed by the laboratory and where the petitioner repeatedly objected to testimony concerning the tests in their absence. Since the burden was on the respondent to establish the petitioner’s use of and possession of drugs, it is no answer to say that the petitioner could have called the absent witnesses.
As to the fourth standard, the respondent’s assertions of how burdensome production of the actual testing witnesses would be is not supported by the record. No information whatsoever is provided us to how many proceedings similar to this one are held per year at which the testers might be called to testify.
While the majority cites testimony that the chances of error were close to zero, the testimony of the one expert who testified that there were at least 500 erroneous samples per *582year out of the 50,000 tests performed cannot be ignored.1 This testimony must be weighed in the face of petitioner’s sworn denial that he had ever used drugs. In addition, while the majority also points to some chronic absenteeism on the part of the petitioner between 1987 and 1989, the record does not reveal what is meant by "Chronic Absent, category A” and "Chronic Absent, category B.”2 Moreover, the absences were explained by petitioner as related to a line of duty injury to his knee, and there was no showing that these absences were related to drug use.
Finally, the criminal cases which the majority cites in support of its conclusion that confrontation here was unnecessary are clearly distinguishable and inapplicable here. In each case the substance tested had been obtained from the person of the defendant or his residence. In Minner v Kerby (30 F3d 1311), one issue was whether the notes of a chemist trainee which found cocaine in the substance tested was properly admitted into evidence. The defendant had been found guilty of possession of cocaine with intent to distribute. First, the evidence was that defendant had thrown the substance under a car as the police approached. Second, the chemist trainee was no longer employed even though it was not shown that he was unavailable. Third, the absent trainee’s results had been checked and verified by his superior who did testify. Fourth, separate testing of the same substance had been done by another expert who did testify. Under these circumstances the court ruled that defendant’s right of confrontation had not been violated.
In United States v Smith (964 F2d 1221), a report of a chemist who did not testify was not the sole evidence linking the defendant to crack cocaine. The defendant had been convicted of drug trafficking activity. The evidence was that he had sold cocaine to undercover officers on three occasions. A search of his home pursuant to a search warrant produced *583more cocaine. The chemist who performed the tests on the crack cocaine found during the search of his home did not testify because of back problems. The trial court permitted the tester’s supervisor to give his opinion of the substance seized based on Federal Rules of Evidence, rule 703 ("Bases of Opinion Testimony by Experts”). No comparable rule of evidence exists in New York State.
In Reardon v Manson (806 F2d 39 [State case is reported at State v Reardon, 172 Conn 593, 376 A2d 65]), the United States Court of Appeals found no confrontation problem when a supervisor gave testimony that certain substances contained marihuana and cocaine. The defendant had been convicted of possession of marihuana and the sale of cocaine. The Second Circuit permitted testimony from the supervisor who personally gave the substances to the toxicologists and to whom test results were brought. Some of the tests were "mechanically objective in nature” (806 F2d, at 41). It is clear that the supervisor both personally supervised and confirmed the tests undertaken.
Finally, while the majority cites Matter of McBarnette v Sobol (83 NY2d 333), for the proposition that there is a limited right to cross-examine adverse witnesses in administrative proceedings, the primary issue there was whether Public Health Law § 230 (11) (a) required written misconduct complaints against physicians to remain confidential where the complainants were known to the physician involved in disciplinary proceedings alleging sexual misconduct and had testified against him. The case certainly does not stand for the proposition that either cross-examination or confrontation of witnesses is limited.
I would reverse and send the case back for a new hearing at which the absent witnesses would be required to testify.
Judges Simons, Bellacosa, Levine and Ciparick concur with Chief Judge Kaye; Judge Smith dissents and votes to reverse in a separate opinion in which Judge Titone concurs.
Judgment affirmed, with costs.
. When petitioner’s attorney asked respondent’s expert whether the error range could be as high as 10% (or 5,000) the expert replied, “I think taken out of context you could say that, yes”. He further indicated that such estimates of error involved "[o]lder contemporary technologies”. He stated further that the newer GC/MS instruments were "more reliable” and that the persons using these instruments in New York State were closely regulated. There was no further elaboration of the differences in the older or newer GC/MS instruments.
. Petitioner was chronic absent, category A, four times in 1987 and one time in 1988. He was chronic absent, category B, on five occasions in 1988 and two occasions in 1989.