with whom Justice Scalia, Justice Ginsburg, and Justice Sotomayor join, dissenting.
Some years ago, the State of California prosecuted a man named John Kocak for rape. At a preliminary hearing, the State presented testimony from an analyst at the Cellmark Diagnostics Laboratory—the same facility used to generate DNA evidence in this case. The analyst had extracted DNA from a bloody sweatshirt found at the crime scene and then compared it to two control samples—one from Kocak and one from the victim. The analyst’s report identified a single match: As she explained on direct examination, the DNA found on the sweatshirt belonged to Kocak. But after undergoing cross-examination, the analyst realized she had made a mortifying error. She took the stand again, but this time to admit that the report listed the victim’s control sample as coming from Kocak, and Kocak’s as coming from the victim. So the DNA on the sweatshirt matched not Kocak, *119but the victim herself. See Tr. in No. SCD110465 (Super. Ct. San Diego Cty., Cal., Nov. 17,1995), pp. 3-4 (“I’m a little hysterical right now, but I think . . . the two names should be switched”), online at http://www.nlada.org/forensics/for_ lib/Documents/1037341561.0/JohnIvanKocak.pdf (as visited June 15, 2012, and available in Clerk of Court’s case file). In trying Kocak, the State would have to look elsewhere for its evidence.
Our Constitution contains a mechanism for catching such errors—the Sixth Amendment’s Confrontation Clause. That Clause, and the Court’s recent cases interpreting it, require that testimony against a criminal defendant be subject to cross-examination. And that command applies with full force to forensic evidence of the kind involved in both the Kocak case and this one. In two decisions issued in the last three years, this Court held that if a prosecutor wants to introduce the results of forensic testing into evidence, he must afford the defendant an opportunity to cross-examine an analyst responsible for the test. Forensic evidence is reliable only when properly produced, and the Confrontation Clause prescribes a particular method for determining whether that has happened. The Kocak incident illustrates how the Clause is designed to work: Once confronted, the analyst discovered and disclosed the error she had made. That error would probably not have come to light if the prosecutor had merely admitted the report into evidence or asked a third party to present its findings. Hence the genius of an 18th-century device as applied to 21st-century evidence: Cross-examination of the analyst is especially likely to reveal whether vials have been switched, samples contaminated, tests incompetently run, or results inaccurately recorded.
Under our Confrontation Clause precedents, this is an open-and-shut case. The State of Illinois prosecuted Sandy Williams for rape based in part on a DNA profile created in Cellmark’s laboratory. Yet the State did not give Williams *120a chance to question the analyst who produced that evidence. Instead, the prosecution introduced the results of Cellmark’s testing through an expert witness who had no idea how they were generated. That approach—no less (perhaps more) than the confrontation-free methods of presenting forensic evidence we have formerly banned—deprived Williams of his Sixth Amendment right to “con£ron[t] . . . the witnesses against him.”
The Court today disagrees, though it cannot settle on a reason why. Justice Alito, joined by three other Justices, advances two theories—that the expert’s summary of the Cellmark report was not offered for its truth, and that the report is not the kind of statement triggering the Confrontation Clause’s protection. In the pages that follow, I call Justice Alito’s opinion “the plurality,” because that is the conventional term for it. But in all except its disposition, his opinion is a dissent: Five Justices specifically reject every aspect of its reasoning and every paragraph of its explication. See ante, at 104 (Thomas, J., concurring in judgment) (“I share the dissent’s view of the plurality’s flawed analysis”). Justice Thomas, for his part, contends that the Cellmark report is nontestimonial on a different rationale. But no other Justice joins his opinion or subscribes to the test he offers.
That creates five votes to approve the admission of the Cellmark report, but not a single good explanation. The plurality’s first rationale endorses a prosecutorial dodge; its second relies on distinguishing indistinguishable forensic reports. Justice Thomas’s concurrence, though positing an altogether different approach, suffers in the end from similar flaws. I would choose another path—to adhere to the simple rule established in our decisions, for the good reasons we have previously given. Because defendants like Williams have a constitutional right to confront the witnesses against them, I respectfully dissent from the Court’s fractured decision.
*121hH
Our modern Confrontation Clause doctrine began with Crawford v. Washington, 541 U. S. 36 (2004). About a quarter century earlier, we had interpreted the Clause to allow the admission of any out-of-court statement falling within a “firmly rooted hearsay exception” or carrying “particularized guarantees of trustworthiness.” Ohio v. Roberts, 448 U. S. 56, 66 (1980). But in Crawford, we concluded that our old approach was misguided. Drawing on historical research about the Clause’s purposes, we held that the prosecution may not admit “testimonial statements of a witness who [does] not appear at trial unless he [is] unavailable to testify, and the defendant . . . had a prior opportunity for cross-examination.” 541 U. S., at 53-54. That holding has two aspects. First, the Confrontation Clause applies only to out-of-court statements that are “testimonial.” Second, where the Clause applies, it guarantees to a defendant just what its name suggests—the opportunity to cross-examine the person who made the statement. See id., at 59.
A few years later, we made clear that Crawford’s, rule reaches forensic reports. In Melendez-Diaz v. Massachusetts, 557 U. S. 305 (2009), the Commonwealth introduced a laboratory’s “‘certificates of analysis’” stating that a substance seized from the defendant was cocaine. Id., at 308. We held that the certificates fell within the Clause’s “ ‘core class of testimonial statements’” because they had a clear “evidentiary purpose”: They were “‘made under circumstances which would lead an objective witness reasonably to believe that [they] would be available for use at a later trial.’ ” Id., at 310-311 (quoting Crawford, 541 U. S., at 51-52). Accordingly, we ruled, the defendant had a right to cross-examine the analysts who had authored them. In reaching that conclusion, we rejected the Commonwealth’s argument that the Confrontation Clause should not apply because the statements resulted from “‘neutral scientific testing,’ ” and so were presumptively reliable. 557 U. S., at *122318. The Clause, we noted, commands that “ ‘reliability be assessed in a particular manner’ ”—through “ ‘testing in the crucible of cross-examination.’” Id., at 317 (quoting Crawford, 541 U. S., at 61). Further, we doubted that the testing summarized in the certificates was “as neutral or as reliable” as the Commonwealth suggested. Citing chapter and verse from various studies, we concluded that “[fjorensic evidence is not uniquely immune from the risk of manipulation” and mistake. 557 U. S., at 318; see id., at 319.
And just two years later (and just one year ago), we reiterated Melendez-Diaz’s analysis when faced with a State’s attempt to evade it. In Bullcoming v. New Mexico, 564 U. S. 647 (2011), a forensic report showed the defendant’s blood-alcohol concentration to exceed the legal limit for drivers. The State tried to introduce that finding through the testimony of a person who worked at the laboratory but had not performed or observed the blood test or certified its results. We held that Melendez-Diaz foreclosed that tactic. The report, we stated, resembled the certificates in Melendez-Diaz in “all material respects,” 564 U. S., at 664: Both were signed documents providing the results of forensic testing designed to “ ‘prov[ej some fact’ in a criminal proceeding,” ibid. (quoting Melendez-Diaz, 557 U. S., at 310). And the State’s resort to a “surrogate” witness, in place of the analyst who produced the report, did not satisfy the Confrontation Clause. Bullcoming, 564 U. S., at 661. Only the presence of “that particular scientist,” we reasoned, would enable Bullcoming’s counsel to ask “questions designed to reveal whether incompetence ... or dishonesty” had tainted the results. Id., at 652, 662. Repeating the refrain of Melendez-Diaz, we held that “[tjhe accused’s right is to be confronted with” the actual analyst, unless he is unavailable and the accused “had an opportunity, pretrial, to cross-examine” him. Bullcoming, 564 U. S., at 652.
This case is of a piece. The report at issue here shows a DNA profile produced by an analyst at Cellmark’s labora*123tory, allegedly from vaginal swabs taken from a young woman, L. J., after she was raped. That report is identical to the one in Bullcoming (and Melendez-Diaz) in “all material respects.” 564 U. S., at 664. Once again, the report was made to establish “‘some fact’ in a criminal proceeding”—here, the identity of L. J.’s attacker. Ibid. (quoting Melendez-Diaz, 557 U. S., at 310); see infra, at 137. And once again, it details the results of forensic testing on evidence gathered by the police. Viewed side-by-side with the Bullcoming report, the Cellmark analysis has a comparable title; similarly describes the relevant samples, test methodology, and results; and likewise includes the signatures of laboratory officials. Compare Cellmark Diagnostics Report of Laboratory Examination (Feb. 15, 2001), Lodging of Petitioner, with App. in Bullcoming v. New Mexico, O. T. 2010, No. 09-10876, pp. 62-65. So under this Court’s prior analysis, the substance of the report could come into evidence only if Williams had a chance to cross-examine the responsible analyst.
But that is not what happened. Instead, the prosecutor used Sandra Lambatos—a state-employed scientist who had not participated in the testing—as the conduit for this piece of evidence. Lambatos came to the stand after two other state analysts testified about forensic tests they had performed. One recounted how she had developed a DNA profile of Sandy Williams from a blood sample drawn after his arrest. And another told how he had confirmed the presence of (unidentified) semen on the vaginal swabs taken from L. J. All this was by the book: Williams had an opportunity to cross-examine both witnesses about the tests they had run. But of course, the State still needed to supply the missing link—it had to show that DNA found in the semen on L. J.’s vaginal swabs matched Williams’s DNA. To fill that gap, the prosecutor could have called the analyst from Cellmark to testify about the DNA profile she had produced from the swabs. But instead, the State called Lambatos as *124an expert witness and had her testify that the semen on those swabs contained Sandy Williams’s DNA:
“Q Was there a computer match generated of the male DNA profile found in semen from the vaginal swabs of [L. J.] to a male DNA profile that had been identified as having originated from Sandy Williams?
“A Yes, there was.
“Q Did you compare the semen . . . from the vaginal swabs of [L. J.] to the male DNA profile . . . from the blood of Sandy Williams?
“A Yes, I did.
[[Image here]]
“Q [I]s the semen identified in the vaginal swabs of [L. J.] consistent with having originated from Sandy Williams?
“A Yes.” App. 56-57.
And so it was Lambatos, rather than any Cellmark employee, who informed the trier of fact that the testing of L. J.’s vaginal swabs had produced a male DNA profile implicating Williams.
Have we not already decided this case? Lambatos’s testimony is functionally identical to the “surrogate testimony” that New Mexico proffered in Bullcoming, which did nothing to cure the problem identified in Melendez-Diaz (which, for its part, straightforwardly applied our decision in Crawford). Like the surrogate witness in Bullcoming, Lambatos “could not convey what [the actual analyst] knew or observed about the events . . . , i. e., the particular test and testing process he employed.” Bullcoming, 564 U. S., at 661. “Nor could such surrogate testimony expose any lapses or lies” on the testing analyst’s part. Id., at 661-662. Like the lawyers in Melendez-Diaz and Bullcoming, Williams’s attorney could not ask questions about that analyst’s “proficiency, the care he took in performing his work, and his veracity.” 564 U. S., at 662, n. 7. He could not probe whether the analyst had *125tested the wrong vial, inverted the labels on the samples, committed some more technical error, or simply made up the results. See App. to Brief for Public Defender Service for the District of Columbia et al. as Amici Curiae 5a, 11a (describing mistakes and fraud at Cellmark’s laboratory). Indeed, Williams’s lawyer was even more hamstrung than Bull-coming’s. At least the surrogate witness in Bullcoming worked at the relevant laboratory and was familiar with its procedures. That is not true of Lambatos: She had no knowledge at all of Cellmark’s operations. Indeed, for all the record discloses, she may never have set foot in Cell-mark’s laboratory.
Under our case law, that is sufficient to resolve this case. “[W]hen the State elected to introduce” the substance of Cellmark’s report into evidence, the analyst who generated that report “became a witness” whom Williams “had the right to confront.” Bullcoming, 564 U. S., at 663. As we stated just last year, “Our precedents] cannot sensibly be read any other way.” Ibid.
II
The plurality’s primary argument to the contrary tries to exploit a limit to the Confrontation Clause recognized in Crawford. “The Clause,” we cautioned there, “does not bar the use of testimonial statements for purposes other than establishing the truth of the matter asserted.” 541 U. S., at 59-60, n. 9 (citing Tennessee v. Street, 471 U. S. 409, 414 (1985)). The Illinois Supreme Court relied on that statement in concluding that Lambatos’s testimony was permissible. On that court’s view, “Lambatos disclosed the underlying facts from Cellmark’s report” not for their truth, but “for the limited purpose of explaining the basis for her [expert] opinion,” so that the factfinder could assess that opinion’s value. 238 Ill. 2d 125, 150, 939 N. E. 2d 268, 282 (2010). The plurality wraps itself in that holding, similarly asserting that Lambatos’s recitation of Cellmark’s findings, *126when viewed through the prism of state evidence law, was not introduced to establish “the truth of any . . . matter concerning [the] Cellmark” report. Ante, at 71; see ante, at 57, 77-78. But five Justices agree, in two opinions reciting the same reasons, that this argument has no merit: Lambatos’s statements about Cellmark’s report went to its truth, and the State could not rely on her status as an expert to circumvent the Confrontation Clause’s requirements. See ante, at 104-110 (opinion of Thomas, J.).
To see why, start with the kind of ease Crawford had in mind. In acknowledging the not-for-the-truth carveout from the Clause, the Court cited Tennessee v. Street as exemplary. See Crawford, 541 U. S., at 59-60, n. 9. There, Street claimed that his stationhouse confession of murder was a sham: A police officer, he charged, had read aloud his alleged accomplice’s confession and forced him to repeat it. To help rebut that defense, the State introduced the other confession into the record, so the jury could see how it differed from Street’s. This Court rejected Street’s Confrontation Clause claim because the State had offered the out-of-court statement not to prove “the truth of [the accomplice’s] assertions” about the murder, but only to disprove Street’s claim of how the police elicited his confession. Street, 471 U. S., at 413. Otherwise said, the truth of the admitted statement was utterly immaterial; the only thing that mattered was that the statement (whether true or false) varied from Street’s.
The situation could not be more different when a witness, expert or otherwise, repeats an out-of-court statement as the basis for a conclusion, because the statement’s utility is then dependent on its truth. If the statement is true, then the conclusion based on it is probably true; if not, not. So to determine the validity of the witness’s conclusion, the fact-finder must assess the truth of the out-of-court statement on which it relies. That is why the principal modern treatise on evidence variously calls the idea that such “basis evi*127dence” comes in not for its truth, but only to help the fact-finder evaluate an expert’s opinion “very weak,” “factually implausible,” “nonsense,” and “sheer fiction.” D. Kaye, D. Bernstein, & J. Mnookin, The New Wigmore: Expert Evidence § 4.10.1, pp. 196-197 (2d ed. 2011); id., § 4.11.6, at 24 (Supp. 2012). “One can sympathize,” notes that treatise, “with a court’s desire to permit the disclosure of basis evidence that is quite probably reliable, such as a routine analysis of a drug, but to pretend that it is not being introduced for the truth of its contents strains credibility.” Id., § 4.10.1, at 198 (2d ed. 2011); see also, e. g., People v. Goldstein, 6 N. Y. 3d 119, 128, 843 N. E. 2d 727, 732-733 (2005) (“The distinction between a statement offered for its truth and a statement offered to shed light on an expert’s opinion is not meaningful”). Unlike in Street, admission of the out-of-court statement in this context has no purpose separate from its truth; the factfinder can do nothing with it except assess its truth and so the credibility of the conclusion it serves to buttress.1
Consider a prosaic example not involving scientific experts. An eyewitness tells a police officer investigating an assault that the perpetrator had an unusual, star-shaped birthmark over his left eye. The officer arrests a person bearing that birthmark (let’s call him Starr) for committing the offense. And at trial, the officer takes the stand and recounts just what the eyewitness told him. Presumably the plurality would agree that such testimony violates the Confrontation Clause unless the eyewitness is unavailable and the defendant had a prior opportunity to cross-examine *128him. Now ask whether anything changes if the officer couches his testimony in the following way: “I concluded that Starr was the assailant because a reliable eyewitness told me that the assailant had a star-shaped birthmark and, look, Starr has one just like that.” Surely that framing would make no constitutional difference, even though the eyewitness’s statement now explains the basis for the officer’s conclusion. It remains the case that the prosecution is attempting to introduce a testimonial statement that has no relevance to the proceedings apart from its truth—and that the defendant cannot cross-examine the person who made it. Allowing the admission of this evidence would end-run the Confrontation Clause, and make a parody of its strictures.
And that example, when dressed in scientific clothing, is no different from this ease. The Cellmark report identified the rapist as having a particular DNA profile (think of it as the quintessential birthmark). The Confrontation Clause prevented the State from introducing that report into evidence except by calling to the stand the person who prepared it. See Melendez-Diaz, 557 U. S., at 310-311; Bullcoming, 564 U. S., at 652. So the State tried another route—introducing the substance of the report as part and parcel of an expert witness’s conclusion. In effect, Lambatos testified (like the police officer above): “I concluded that Williams was the rapist because Cellmark, an accredited and trustworthy laboratory, says that the rapist has a particular DNA profile and, look, Williams has an identical one.” And here too, that form of testimony should change nothing. The use of the Cellmark statement remained bound up with its truth, and the statement came into evidence without any opportunity for Williams to cross-examine the person who made it. So if the plurality were right, the State would have a ready method to bypass the Constitution (as much as in my hypothetical case); a wink and a nod, and the Confrontation Clause would not pose a bar to forensic evidence.
*129The plurality tries to make plausible its not-for-the-truth rationale by rewriting Lambatos’s testimony about the Cell-mark report. According to the plurality, Lambatos merely “assumed” that Cellmark’s DNA profile came from L. J.’s vaginal swabs, accepting for the sake of argument the prosecutor’s premise. Ante, at 72. But that is incorrect. Nothing in Lambatos’s testimony indicates that she was making an assumption or considering a hypothesis. To the contrary, Lambatos affirmed, without qualification, that the Cellmark report showed a “male DNA profile found in semen from the vaginal swabs of [L. J.].” App. 56. Had she done otherwise, this case would be different. There was nothing wrong with Lambatos’s testifying that two DNA profiles— the one shown in the Cellmark report and the one derived from Williams’s blood—matched each other; that was a straightforward application of Lambatos’s expertise. Similarly, Lambatos could have added that if the Cellmark report resulted from scientifically sound testing of L. J.’s vaginal swabs, then it would link Williams to the assault. What Lambatos could not do was what she did: indicate that the Cellmark report was produced in this way by saying that L, J.’s vaginal swabs contained DNA matching Williams’s.2 *130By testifying in that manner, Lambatos became just like the surrogate witness in Bullcoming—a person knowing nothing about “the particular test and testing process,” but vouching for them regardless. 564 U. S., at 661. We have held that the Confrontation Clause requires something more.
The plurality also argues that Lambatos’s characterization of the Cellmark report did not violate the Confrontation Clause because the case “involve[d] a bench trial.” Ante, at 72 (emphasis deleted). I welcome the plurality’s concession that the Clause might forbid presenting Lambatos’s statement to a jury, see ibid.; it indicates that the plurality realizes that her testimony went beyond an “assumption.” But the presence of a judge does not transform the constitutional question. In applying the Confrontation Clause, we have never before considered relevant the decisionmaker’s identity. See, e. g., Davis v. Washington, 547 U. S. 813 (2006). And this case would be a poor place to begin. Lam-batos’s description of the Cellmark report was offered for its truth because that is all such “basis evidence” can be offered for; as described earlier, the only way the factfinder could consider whether that statement supported her opinion (that the DNA on L. J.’s swabs came from Williams) was by assessing the statement’s truth. See supra, at 126-129. That is so, as a simple matter of logic, whether the factfinder is a judge or a jury. And thus, in either case, admission of the statement, without the opportunity to cross-examine, violates the Confrontation Clause. See ante, at 106, n. 1 (opinion of Thomas, J.).
In saying that much, I do not doubt that a judge typically will do better than a jury in excluding such inadmissible evidence from his decisionmaking process. Perhaps the judge *131did so here; perhaps, as the plurality thinks, he understood that he could not consider Lambatos’s representation about the Cellmark report, and found that other, “circumstantial evidence” established “the source of the sample that Cell-mark tested” and “the reliability of the Cellmark profile.” Ante, at 75-76. Some indications are to the contrary: In delivering his verdict, the judge never referred to the circumstantial evidence the plurality marshals, but instead focused only on Lambatos’s testimony. See 4 Record JJJ151 (calling Lambatos “the best DNA witness I have ever heard” and referring to Williams as “the guy whose DNA, according to the evidence from the experts, is in the semen recovered from the victim’s vagina”). But I take the plurality’s point that when read “[i]n context” the judge’s statements might be “best understood” as meaning something other than what they appear to say. See ante, at 73-74, n. 6. Still, that point suggests only that the admission of Lambatos’s statement was harmless—that the judge managed to put it out of mind. After all, whether a factfinder is confused by an error is a separate question from whether an error has occurred. So the plurality’s argument does not answer the only question this case presents: whether a constitutional violation happened when Lambatos recited the Cellmark report’s findings.3
*132At bottom, the plurality’s not-for-the-truth rationale is a simple abdication to state-law labels. Although the utility of the Cellmark statement that Lambatos repeated logically depended on its truth, the plurality thinks this case decided by an Illinois rule holding that the facts underlying an expert’s opinion are not admitted for that purpose. See ante, at 69-72; People v. Pasch, 152 Ill. 2d 133, 175-177, 604 N. E. 2d 294, 311 (1992). But we do not typically allow state law to define federal constitutional requirements. And needless to say (or perhaps not), the Confrontation Clause is a constitutional rule like any other. As Justice Thomas observes, even before Crawford, we did not allow the Clause’s scope to be “dictated by state or federal evidentiary rules.” Ante, at 105. Indeed, in Street, we independently reviewed whether an out-of-court statement was introduced for its truth—the very question at issue in this case. See 471 U. S., at 413-416. And in Crawford, we still more firmly disconnected the Confrontation Clause inquiry from state evidence law, by overruling an approach that looked in part to whether an out-of-court statement fell within a “‘firmly rooted hearsay exception.’” 541 U. S., at 60 (quoting Roberts, 448 U. S., at 66). That decision made clear that the Confrontation Clause’s protections are not coterminous with rules of evidence. So the plurality’s state-law-first approach would be an about-face.
Still worse, that approach would allow prosecutors to do through subterfuge and indirection what we previously have held the Confrontation Clause prohibits. Imagine for a moment a poorly trained, incompetent, or dishonest laboratory *133analyst. (The analyst in Bullcoming, placed on unpaid leave for unknown reasons, might qualify.) Under our precedents, the prosecutor cannot avoid exposing that analyst to cross-examination simply by introducing his report. See Melendez-Diaz, 557 U. S., at 311. Nor can the prosecutor escape that fate by offering the results through the testimony of another analyst from the laboratory. See Bullcoming, 564 U. S., at 652. But under the plurality’s approach, the prosecutor could choose the analyst-witness of his dreams (as the judge here said, “the best DNA witness I have ever heard”), offer her as an expert (she knows nothing about the test, but boasts impressive degrees), and have her provide testimony identical to the best the actual tester might have given (“the DNA extracted from the vaginal swabs matched Sandy Williams’s”)—all so long as a state evidence rule says that the purpose of the testimony is to enable the factfinder to assess the expert opinion’s basis. (And this tactic would not be confined to cases involving scientific evidence. As Justice Thomas points out, the prosecutor could similarly substitute experts for all kinds of people making out-of-court statements. See ante, at 109-110.) The plurality thus would countenance the Constitution’s circumvention. If the Confrontation Clause prevents the State from getting its evidence in through the front door, then the State could sneak it in through the back. What a neat trick— but really, what a way to run a criminal justice system. No wonder five Justices reject it.
Ill
The plurality also argues, as a “second, independent basis” for its decision, that the Cellmark report falls outside the Confrontation Clause’s ambit because it is nontestimonial. Ante, at 58. The plurality tries out a number of supporting theories, but all in vain: Each one either conflicts with this Court’s precedents or misconstrues this case’s facts. Justice Thomas rejects the plurality’s views for similar reasons as I do, thus bringing to five the number of Justices who *134repudiate the plurality’s understanding of what statements count as testimonial. See ante, at 103-104, 114-117. Justice Thomas, however, offers a rationale of his own for deciding that the Cellmark report is nontestimonial. I think his essay works no better. When all is said and done, the Cellmark report is a testimonial statement.
A
According to the plurality, we should declare the Cellmark report nontestimonial because “the use at trial of a DNA report prepared by a modern, accredited laboratory ‘bears little if any resemblance to the historical practices that the Confrontation Clause aimed to eliminate.’” Ante, at 86 (quoting Michigan v. Bryant, 562 U. S. 344, 379 (2011) (Thomas, J., concurring in judgment)). But we just last year treated as testimonial a forensic report prepared by a “modern, accredited laboratory”; indeed, we declared that the report at issue “fell within the core class of testimonial statements” implicating the Confrontation Clause. Bullcoming, 564 U. S., at 665 (internal quotation marks omitted); see Brief for New Mexico Department of Health, Scientific Laboratory Division, as Amicus Curiae in Bullcoming, O. T. 2010, No. 09-10786, pp. 1-2 (discussing accreditation). And although the plurality is close, it is not quite ready (or able) to dispense with that decision. See ante, at 82, n. 13 (“Experience might yet show that the holdings in [Bullcoming and other post-Crawford] cases should be reconsidered”). So the plurality must explain: What could support a distinction between the laboratory analysis there and the DNA test in this case?4
*135As its first stab, the plurality states that the Cellmark report was “not prepared for the primary purpose of accusing a targeted individual.” Ante, at 84. Where that test comes from is anyone’s guess. Justice Thomas rightly shows that it derives neither from the text nor from the history of the Confrontation Clause. See ante, at 114-117 (opinion concurring in judgment). And it has no basis in our precedents. We have previously asked whether a statement was made for the primary purpose of establishing “past events potentially relevant to later criminal prosecution”— in other words, for the purpose of providing evidence. Davis, 547 U. S., at 822; see also Bullcoming, 564 U. S., at 664; Bryant, 562 U. S., at 361; Melendez-Diaz, 557 U. S., at 310-311; Crawford, 541 U. S., at 51-52. None of our cases has ever suggested that, in addition, the statement must be meant to accuse a previously identified individual; indeed, in Melendez-Diaz, we rejected a related argument that laboratory “analysts are not subject to confrontation because they are not ‘accusatory’ witnesses.” 557 U. S., at 313.
Nor does the plurality give any good reason for adopting an “accusation” test. The plurality apparently agrees with Justice Breyer that prior to a suspect’s identification, it will be “unlikely that a particular researcher has a defendant-related motive to behave dishonestly.” Ante, at 97 (Breyer, J., concurring); see ante, at 84-85 (plurality opinion). But surely the typical problem with laboratory analyses—and the typical focus of cross-examination—has to do with careless or incompetent work, rather than with per*136sonal vendettas. And as to that predominant concern, it makes not a whit of difference whether, at the time of the laboratory test, the police already have a suspect.5
The plurality next attempts to invoke our precedents holding statements nontestimonial when made “to respond to an ‘ongoing emergency,’” rather than to create evidence for trial, Bryant, 562 U. S., at 358; here, the plurality insists, the Cellmark report’s purpose was “to catch a dangerous rapist who was still at large.” Ante, at 84. But that is to stretch both our “ongoing emergency” test and the facts of this case beyond all recognition. We have previously invoked that test to allow statements by a woman who was being assaulted and a man who had just been shot. In doing so, we stressed the “informal [and] harried” nature of the statements, Bryant, 562 U. S., at 377—that they were made as, or “minutes” after, id., at 374, the events they described “actually happened],” Davis, 547 U. S., at 827 (emphasis deleted), by “frantic” victims of criminal attacks, ibid., to officers trying to figure out “what had . . . occurred” and what threats remained, Bryant, 562 U. S., at 376 (internal quotation marks omitted). On their face, the decisions have nothing to say about laboratory analysts conducting routine tests far away from a crime scene. And this case presents a peculiarly inapt set of facts for extending those precedents. Lambatos testified at trial that “all reports in this case were prepared for this criminal investigation . . . [a]nd for the purpose of the eventual litigation,” App. 82—in other words, for the purpose of producing evidence, not enabling emergency re*137sponders. And that testimony fits the relevant timeline. The police did not send the swabs to Cellmark until November 2008—nine months after L. J.’s rape—and did not receive the results for another four months. See id., at 30-34, 51-52, 54. That is hardly the typical emergency response.
Finally, the plurality offers a host of reasons for why reports like this one are reliable: “[T]here [i]s no prospect of fabrication,” ante, at 85 (internal quotation marks omitted); multiple technicians may “work on each DNA profile,” ibid.) and “defects in a DNA profile may often be detected from the profile itself,” ibid. See also ante, at 95-99 (opinion of Breyer, J.). But once again: Been there, done that. In Melendez-Diaz, this Court rejected identical arguments, noting extensive documentation of “[sjerious deficiencies ... in the forensic evidence used in criminal trials.” 557 U. S., at 319; see supra, at 122; see also Bullcoming, 564 U. S., at 654, n. 1 (citing similar errors in laboratory analysis); Brief for Public Defender Service for the District of Columbia et al. as Amici Curiae 13 (discussing “[sjystemic problems,” such as sample contamination, sample switching, mislabeling, and fraud, at “ ‘flagship’ DNA labs”). Scientific testing is “technical,” to be sure, ante, at 86 (opinion of Breyer, J.); but it is only as reliable as the people who perform it. That is why a defendant may wish to ask the analyst a variety of questions: How much experience do you have? Have you ever made mistakes in the past? Did you test the right sample? Use the right procedures? Contaminate the sample in any way? Indeed, as scientific evidence plays a larger and larger role in criminal prosecutions, those inquiries will often be the most important in the case.6
*138And Melendez-Diaz made yet a more fundamental point in response to claims of the über alies reliability of scientific evidence: It is not up to us to decide, ex ante, what evidence is trustworthy and what is not. See 557 U. S., at 317-318; see also Bullcoming, 564 U. S., at 660. That is because the Confrontation Clause prescribes its own “procedure for determining the reliability of testimony in criminal trials.” Crawford, 541 U. S., at 67. That procedure is cross-examination. And “[dispensing with [it] because testimony is obviously reliable is akin to dispensing with jury trial because a defendant is obviously guilty.” Id., at 62.
So the plurality’s second basis for denying Williams’s right of confrontation also fails. The plurality can find no reason consistent with our precedents for treating the Cellmark report as nontestimonial. That is because the report is, in every conceivable respect, a statement meant to serve as evidence in a potential criminal trial. And that simple fact should be sufficient to resolve the question.
B
Justice Thomas’s unique method of defining testimonial statements fares no better. On his view, the Confrontation Clause “regulates only the use of statements bearing Indicia of solemnity.’ ” Ante, at 111 (quoting Davis, 547 U. S., at 836-*139837). And Cellmark’s report, he concludes, does not qualify because it is “neither a sworn nor a certified declaration of fact.” Ante, at 111. But Justice Thomas’s approach grants constitutional significance to minutia, in a way that can only undermine the Confrontation Clause’s protections.
To see the point, start with precedent, because the Court rejected this same kind of argument, as applied to this same kind of document, at around this same time just last year. In Bullcoming, the State asserted that the forensic report at issue was nontestimonial because—unlike the report in Melendez-Diaz—it was not sworn before a notary public. We responded that applying the Confrontation Clause only to a sworn forensic report “would make the right to confrontation easily erasable”—next time, the laboratory could file the selfsame report without the oath. 564 U. S., at 664. We then held, as noted earlier, that “[i]n all material respects,” the forensic report in Bullcoming matched the one in Melendez-Diaz. 564 U. S., at 664; see supra, at 122. First, a law enforcement officer provided evidence to a state laboratory assisting in police investigations. See 564 U. S., at 664. Second, the analyst tested the evidence and “prepared a certificate concerning the result[s].” Id., at 665. Third, the certificate was “formalized in a signed document . . . headed a ‘report.’” Ibid. (some internal quotation marks omitted). That was enough.
Now compare that checklist of “material” features to the report in this case. The only differences are that Cellmark is a private laboratory under contract with the State (which no one thinks relevant), and that the report is not labeled a “certificate.” That amounts to (maybe) a nickel’s worth of difference: The similarities in form, function, and purpose dwarf the distinctions. See supra, at 122-123. Each report is an official and signed record of laboratory test results, meant to establish a certain set of facts in legal proceedings. Neither looks any more “formal” than the other; neither is any more formal than the other. See ibid. The variances *140are no more (probably less) than would be found if you compared different law schools’ transcripts or different companies’ cashflow statements or different States’ birth certificates. The difference in labeling—a “certificate” in one case, a “report of laboratory examination” in the other—is not of constitutional dimension.
Indeed, Justice Thomas’s approach, if accepted, would turn the Confrontation Clause into a constitutional geegaw— nice for show, but of little value. The prosecution could avoid its demands by using the right kind of forms with the right kind of language. (It would not take long to devise the magic words and rules—principally, never call anything a “certificate.”)7 And still worse: The new conventions, precisely by making out-of-court statements less “solem[n],” ante, at 104, would also make them less reliable—and so turn the Confrontation Clause upside down. See Crawford, 541 U. S., at 52-53, n. 3 (“We find it implausible that a provision which concededly condemned trial by sworn ex parte affidavit thought trial by unsworn ex parte affidavit perfectly OK”). It is not surprising that no other Member of the Court has adopted this position. To do so, as Justice Thomas rightly says of the plurality’s decision, would be to “diminisfh] the Confrontation Clause’s protection” in “the very cases in which the accused should ‘enjoy the right . . . to be confronted with the witnesses against him.’” Ante, at 118.
IV
Before today’s decision, a prosecutor wishing to admit the results of forensic testing had to produce the technician responsible for the analysis. That was the result of not one, but two decisions this Court issued in the last three years. *141But that clear rule is clear no longer. The five Justices who control the outcome of today’s case agree on very little. Among them, though, they can boast of two accomplishments. First, they have approved the introduction of testimony at Williams’s trial that the Confrontation Clause, rightly understood, clearly prohibits. Second, they have left significant confusion in their wake. What comes out of four Justices’ desire to limit Melendez-Diaz and Bullcoming in whatever way possible, combined with one Justice’s one-justice view of those holdings, is—to be frank—who knows what. Those decisions apparently no longer mean all that they say. Yet no one can tell in what way or to what extent they are altered because no proposed limitation commands the support of a majority.
The better course in this case would have been simply to follow Melendez-Diaz and Bullcoming. Precedent-based decisionmaking provides guidance to lower court judges and predictability to litigating parties. Today’s plurality and concurring opinions, and the uncertainty they sow, bring into relief that judicial method’s virtues. I would decide this case consistently with, and for the reasons stated by, Melendez-Diaz and Bullcoming. And until a majority of this Court reverses or confines those decisions, I would understand them as continuing to govern, in every particular, the admission of forensic evidence.
I respectfully dissent.
In responding to this reasoning, the plurality confirms it. According to the plurality, basis evidence supports the “credibility of the expert’s opinion” by showing that he has relied on, and drawn logical inferences from, sound “factual premises.” Ante, at 78. Quite right. And that process involves assessing such premises’ truth: If they are, as the plurality puts it, “unsupported by other evidence in the record” or otherwise baseless, they will not “allay [a factfinder’s] fears” about an “expert’s reasoning.” Ibid. I could not have said it any better.
The plurality suggests that Lambatos’s testimony is merely a modern, streamlined way of answering hypothetical questions and therefore raises no constitutional issue, see ante, at 57, 67-70; similarly, the plurality contends that the difference between what Lambatos said and what I would allow involves only “slightly revising]” her testimony and so can be of no consequence, see ante, at 72, n. 3. But the statement “if X is true, then Y follows” differs materially—and constitutionally—from the statement “Y is true because X is true (according to Z).” The former statement is merely a logical proposition, whose validity the defendant can contest by questioning the speaker. And then, assuming the prosecutor tries to prove the statement’s premise through some other witness, the defendant can rebut that effort through cross-examination. By contrast, the latter statement as well contains a factual allegation (that X is true), which the defendant can only effectively challenge by confronting the person who made it (Z). That is why recognizing the difference between these two forms of testimony is not to insist on an archaism or a formality, but to *130ensure, in line with the Constitution, that defendants have the ability to confront their accusers. And if prosecutors can easily conform their conduct to that constitutional directive, as the plurality suggests, so much the better: I would not have thought it a ground of complaint that the Confrontation Clause, properly understood, manages to protect defendants without overly burdening the State.
The plurality asserts (without citation) that I am “reaeh[ing] the truly remarkable conclusion that the wording of Lambatos’ testimony confused the trial judge,” ante, at 73, and then spends three pages explaining why that conclusion is wrong, see ante, at 73-75. But the plurality is responding to an argument of its own imagining, because I reach no such conclusion. As I just stated, the trial judge might well have ignored Lambatos’s statement about the Cellmark report and relied on other evidence to conclude that “the Cellmark profile was derived from the sample taken from the victim,” ante, at 73. All I am saying is that the admission of that statement violated the Confrontation Clause even if the judge ultimately put it aside, because it came into evidence for nothing other than its truth. See supra, at 126-129.
Similarly, the plurality claims (still without citation) that I think the other evidence about the Cellmark report insufficient, see ante, at 75. But once again, the plurality must be reading someone else’s opinion. I express no view on sufficiency of the evidence because it is irrelevant to *132the Confrontation Clause issue we took this case to decide. It is the plurality that wrongly links the two, spending another five pages trumpeting the strength of the Cellmark report, see ante, at 76-78, 85-86. But the plurality cannot properly decide whether a Confrontation Clause violation occurred at Williams’s trial by determining that Williams was guilty. The American criminal justice system works the opposite way: determining guilt by holding trials in accord with constitutional requirements.
Justice Breyer does not attempt to distinguish our precedents, opting simply to adhere to “the dissenting view set forth in Melendez-Diaz and Bullcoming.” See ante, at 93 (concurring opinion). He principally worries that under those cases, a State will have to call to the witness stand “[s]ix to twelve or more technicians” who have worked on a report. See ante, at 90; see also ante, at 88-89, 101-103. But none of our cases— including this one—has presented the question of how many analysts must *135testify about a given report. (That may suggest that in most cases a lead analyst is readily identifiable.) The problem in the cases—again, including this one—is that no analyst came forward to testify. In the event that some future case presents the multiple-technician issue, the Court can focus on “the broader ‘limits’ question” that troubles Justice Breyer, ante, at 93. But the mere existence of that question is no reason to wrongly decide the case before us—which, it bears repeating, involved the testimony of not twelve or six or three or one, but zero Cellmark analysts.
Neither can the plurality gain any purchase from the idea that a DNA profile is not “inherently inculpatory” because it “tends to exculpate all but one of the more than 7 billion people in the world today.” Ante, at 58; see ante, at 85. All evidence shares this feature: the more inculpatory it is of a single person, the more exculpatory it is of the rest of the world. The one is but the flipside of the other. But no one has ever before suggested that this logical corollary provides a reason to ignore the Constitution’s efforts to ensure the reliability of evidence.
Both the plurality and Justice Breyer warn that if we require analysts to testify, we will encourage prosecutors to forgo DNA evidence in favor of less reliable eyewitness testimony and so “increase the risk of convicting the innocent.” Ante, at 98 (Breyer, J., concurring); see ante, at 58 (plurality opinion). Neither opinion provides any evidence, even by way of anecdote, for that view, and I doubt any exists. DNA evidence *138is usually the prosecutor’s most powerful weapon, and a prosecutor is unlikely to relinquish it just because he must bring the right analyst to the stand. Consider what Lambatos told the factfinder here: The DNA in L. J.’s vaginal swabs matched Williams’s DNA and would match only “1 in 8.7 quadrillion black, 1 in 390 quadrillion white, or 1 in 109 quadrillion Hispanic unrelated individuals.” App. 56-57. No eyewitness testimony could replace that evidence. I note as well that the Innocence Network— a group particularly knowledgeable about the kinds of evidence that produce erroneous convictions—disagrees with the plurality’s and Justice Breyer’s view. It argues here that “[c]onfrontation of the analyst... is essential to permit proper adversarial testing” and so to decrease the risk of convicting the innocent. Brief for the Innocence Network as Amicus Curiae 3, 7.
Justice Thomas asserts there is no need to worry, because “the Confrontation Clause reaches bad-faith attempts to evade the formalized process.” Ante, at 113; see ante, at 111, n. 5. I hope he is right. But Justice Thomas provides scant guidance on how to conduct this novel inquiry into motive.