¶86 (dissenting) — While the majority’s test “avoids the risk of unduly burdening the use of scientific evidence,” majority at 492, it does so only by ignoring and misreading controlling precedent, at the cost of Washington defendants’ Sixth Amendment right to confront the witnesses against them. U.S. Const, amend. VI. The majority’s misguided detour into long-settled questions leads it to conclude that laboratory reports are not testimonial after Melendez-Diaz v. Massachusetts, 557 U.S. 305, 129 S. Ct. 2527, 174 L. Ed. 2d 314 (2009), and that a supervisor can recite the testimony of a subordinate after Bullcoming v. New Mexico,_U.S._, 131 S. Ct. 2705,180 L. Ed. 2d 610 (2011).
*499¶87 These conclusions violate Sione Lui’s federal constitutional right to confront the witnesses against him. Whether the majority believes the confrontation clause should apply to testimony about “scientific” evidence — and the increasingly serious incidents of misconduct at crime laboratories counsel that it should — is immaterial; it is the law under Melendez-Diaz and Bullcoming, and remains the law after Williams v. Illinois,_U.S.__, 132 S. Ct. 2221, 183 L. Ed. 2d 89 (2012) (plurality opinion). We cannot wave away the clear holding of these cases because we would have decided them differently; nor can we adopt a test that categorically violates the confrontation rights of defendants.
¶88 Each of the four reports the prosecution offered against Lui is testimonial under controlling United States Supreme Court precedent. Because the State funneled this testimony into evidence through expert witnesses who lacked any personal knowledge of the facts they recited, Lui was denied his right of confrontation under the Sixth Amendment. This error was not harmless, and he is entitled to a new trial. I respectfully dissent.
ANALYSIS
¶89 The majority claims that our discretion in this case is boundless because case law from the United States Supreme Court “does not provide a controlling rule for cases . . . that involve expert witnesses.” Majority at 470. This is an overstatement at best. In Williams, the Court failed to resolve whether an expert witness may testify to “his independent opinion about underlying testimonial reports that were not themselves admitted into evidence.” Bullcoming, 131 S. Ct. at 2722 (Sotomayor, J., concurring).
¶90 But this is the only question the Court has not answered and the only issue on which we granted review. See Pet. for Review at 1 (“Is the Sixth Amendment Confrontation Clause violated when an expert witness’s testimony *500is based on the work of others who do not testify, and that work was done for the purpose of the criminal prosecution?”). Unfortunately, the majority wanders far afield from this issue without addressing it, instead offering new law in well-traveled areas. Because the testimonial reports introduced against Lui are identical to the reports at issue in Melendez-Diaz and Bullcoming, those precedents control our analysis.
A. Laboratory Reports and Test Results Prepared for Trial Are Testimonial
¶91 The confrontation clause of the Sixth Amendment guarantees that “[i]n all criminal prosecutions, the accused shall enjoy the right... to be confronted with the witnesses against him.” U.S. Const, amend. VI. The majority claims this “plain language” as the source of its test. Majority at 470. The plain language of the confrontation clause, however, does not distinguish between expert and nonexpert witnesses. Moreover, the United States Supreme Court has already determined that “[t]he Constitution’s text does not alone resolve” the scope of the confrontation right. Crawford v. Washington, 541 U.S. 36, 42, 124 S. Ct. 1354, 158 L. Ed. 2d 177 (2004).
¶92 Instead, we must turn to controlling United States Supreme Court precedent. While the majority counts “perspective[s]” and “camps” in these cases, majority at 474 it should count signatures. In Melendez-Diaz and Bullcoming, five justices signed majority opinions of the Court, yielding precisely the authority the majority complains we lack. Compare Melendez-Diaz, 557 U.S. at 306 (“Justice [Antonin] Scalia delivered the opinion of the Court” (emphasis added) (capitalization omitted)),13 and Bullcoming, 131 S. Ct. at 2709 (“Justice [Ruth Bader] Ginsburg deliv*501ered the opinion of the Court, except as to Part IV and footnote 6” (emphasis added) (capitalization omitted)), with Williams, 132 S. Ct. at 2227 (“Justice [Samuel] Alito [Jr.] announced the judgment of the Court and delivered an opinion, in which the Chief Justice, Justice Kennedy, and Justice Breyer join” (emphasis added) (capitalization omitted)).
¶93 Rather than apply the law of these cases to the issues before us, the majority crafts an entirely new test, the centerpiece of which is a distinction between neutral and inculpatory witnesses and between conventional and nonconventional witnesses (i.e., nonexpert and expert witnesses). See majority at 482 n.8. The majority finds support for its distinction in Justice Anthony Kennedy’s dissent in Melendez-Diaz. See id. at 475. The majority embraces Justice Kennedy’s view that laboratory analysts are not “ ‘ “witnesses against” ’ ” a defendant because they make neutral factual findings, and unlike “conventional” inculpatory witnesses they should not be required to testify to their observations. Id. (quoting Melendez-Diaz, 557 U.S. at 333-34 (Kennedy, J., dissenting)); see Bullcoming, 131 S. Ct. at 2726 (Kennedy, J., dissenting). A casual reader of the majority’s account would be forgiven for assuming that Justice Kennedy’s distinction between neutral and inculpatory wit*502nesses is the law. It is not. The United States Supreme Court has rejected the majority’s newfound test four times.
¶94 In Crawford, seven justices held that the confrontation clause applies to all witnesses against the accused regardless of whether the witnesses are neutral or are experts, such as coroners. 541 U.S. at 47 n.2, 66. In Melendez-Diaz, five justices again rejected the notion that the confrontation clause distinguishes between neutral or inculpatory witnesses or that laboratory analysts should be exempted because they are unconventional witnesses. 557 U.S. at 313-14 (majority), 329-30 (Thomas, J., concurring). The Court held that the confrontation clause, when read in conjunction with the adjacent compulsory process clause, “contemplates two classes of witnesses — those against the defendant and those in his favor.” Id. at 313. The Court dismissed the majority’s third class of “neutral” witnesses, holding instead that all witnesses who testify against the defendant must be available for cross-examination, not just those who offer “inculpatory” testimony. Id.-, see also id. at 316 (noting this distinction “would exempt all expert witnesses — a hardly ‘unconventional’ class of witnesses”). The Court further explained that “ [i]t is often, indeed perhaps usually, the case that an adverse witness’s testimony, taken alone, will not suffice to convict.” Id. at 314. The Court emphasized that it has been “longstanding case law” that a police report identifying certain property as stolen triggers the confrontation clause against a defendant on trial for receiving stolen property even though the report is not actually inculpatory. Id. Moreover, the Court refused to exclude neutral scientific testing from confrontation clause protection because it “is little more than an invitation to return to our overruled decision in [Ohio v. ]Roberts,” 448 U.S. 56, 100 S. Ct. 2531, 65 L. Ed. 2d 597 (1980). Melendez-Diaz, 557 U.S. at 317.
¶95 Five justices again rejected the majority’s distinction between neutral and inculpatory testimony in Bull-coming, holding that even results transcribed from a gas *503chromatograph machine without the analyst’s interpretation or independent judgment are subject to confrontation. 131 S. Ct. at 2714-15. Justice Kennedy complained (correctly) that Bullcoming would subject all impartial lab results to confrontation, even those “prepared by experienced technicians in laboratories that follow professional norms and scientific protocols.” Id. at 2726 (Kennedy, J., dissenting). Most recently, in Williams, five justices once again rejected the argument that only testimony accusing a specific, known defendant is subject to cross-examination. 132 S. Ct. at 2273-74 (Kagan, J., dissenting), 2263 (Thomas, J., concurring in the judgment).
¶96 The majority’s insistence that its confrontation clause test is grounded in the plain language of the Sixth Amendment is particularly ironic because adopting it requires us to ignore the plain language of the United States Supreme Court. While the Court has been divided on many issues involving the confrontation clause, in four cases at least five justices have rejected the majority’s view. After several unsuccessful challenges in which this argument was the centerpiece of the State’s briefing and pressed forcefully at oral argument, there can be no doubt that laboratory analysts are constitutionally indistinguishable from any other witness, and that reports about results observed in a laboratory are no different from reports by any eyewitness about any other subject matter. Most recently, in State v. Jasper, 174 Wn.2d 96, 115-16, 271 P.3d 876 (2012), this court unanimously relied on Melendez-Diaz and Bullcoming as controlling authority for this proposition and rejected the State’s attempt to paint these cases as less authoritative than Crawford. Even the Williams plurality acknowledges that Crawford, Melendez-Diaz, and Bull-coming “are to be deemed binding precedents.” 132 S. Ct. at 2242 n.13 (plurality opinion). The majority may not like these holdings, but it cannot wish them away.
¶97 Instead, we must apply controlling law to the questions before us. The rule is simple: a statement is testimo*504nial, and gives rise to a confrontation right, if its primary-purpose is “to establish or prove past events potentially relevant to later criminal prosecution.” Davis v. Washington, 547 U.S. 813, 822, 126 S. Ct. 2266, 165 L. Ed. 2d 224 (2006). This “primary purpose” test applies equally to a written statement that purports to relay the results of scientific testing. See Melendez-Diaz, 557 U.S. at 310-12 (holding that a certificate reporting the results of drug testing triggers the confrontation clause because it was “ ‘made under circumstances which would lead an objective witness reasonably to believe that the statement would be available for use at a later trial’ ” and the analyst who produced it was well aware of its “evidentiary purpose” (quoting Crawford, 541 U.S. at 52)). The primary purpose test applies whether the witness wears a police jacket or a lab coat, and whether the testimony concerns a crime scene or a test tube.
B. The Laboratory Reports and Test Results at Issue Are Testimonial
¶98 In its most recent confrontation clause decision, a fractured United States Supreme Court affirmed the Supreme Court of Illinois without a holding. Williams, 132 S. Ct. 2221. Justice Samuel Alito Jr. wrote the lead opinion, joined by three of his colleagues. Id. at 2227. This plurality rested on two grounds, the first being that the DNA (deoxyribonucleic acid) profile was not offered for its truth and the second being a narrower version of the “primary purpose” test that covers only inculpatory statements — those made “for the purpose of obtaining evidence to be used against petitioner.”Id. at 2228. This narrower test is virtually identical to the one the majority discovers in the “plain language” of the Sixth Amendment. However, five justices flatly rejected this drastic narrowing of the confrontation clause, and it is no more the law than are Justice Kennedy’s previous dissents. Id. at 2273-74 (Kagan, J, dissenting), 2262-63 (Thomas, J., concurring in judgment) (noting the plu*505rality’s test lacks “any grounding in constitutional text, in history, or in logic”).
¶99 Because five justices failed to sign any one opinion or concur in any one rationale, Williams merely affirmed the lower court’s judgment without a holding. It stands as a single-case deviation from the majority opinions of the Court in Melendez-Diaz and Bullcoming, and it is no broader than its facts. Under those facts, all we know from Williams is that an expert may testify to a DNA profile performed by a nontestifying analyst without triggering the confrontation clause if that profile is an informal report created in order to meet an ongoing emergency and to exclude possible suspects before the defendant was ever targeted as a suspect, and that profile is never admitted, shown, read, or identified to the fact finder as a source of the expert’s opinions in a bench trial. Id. at 2230,2236,2242 (plurality opinion).
¶100 Nonetheless, Melendez-Diaz and Bullcoming remain the law, and they plainly hold that “any document prepared for use in a criminal proceeding” is testimonial. Jasper, 174 Wn.2d at 112. Because each of the reports at issue in this case meets this test, and also meets the Williams plurality’s narrower definition of “primary purpose,” each is testimonial and falls within the confrontation clause.
¶101 At issue are four potentially testimonial forensic reports. First, Dr. Kathy Raven recorded the temperature of Elaina Boussiacos’s body and the outside air at the time and place where her body was found. 10 Report of Proceedings (RP) at 1353-55. Second, Raven prepared an autopsy report that concluded Boussiacos was killed by strangulation. Id. at 1334-40,1405; Ex. 168, at 1. As part of her report, Raven also took photographs of Boussiacos’s injuries and the manner in which she was dressed. 10 RP at 1358-65, 1375-95. Third, Martin Hughes with the Washington State Toxicology Laboratory (WSTL) prepared a toxicology report on blood samples taken from Boussiacos at the request of *506the King County medical examiner. Ex. 168, at 10. Finally, DNA analysts Hunan Nasir of Reliagene Technologies and Ms. Vendetes of Orchid Cellmark prepared separate reports that compared DNA samples taken from Lui, Lui’s son, and Boussiacos’s ex-husband with DNA samples collected from the victim. 12 RP at 1491, 1552, 1567; Ex. 136, at 4.
¶102 Each of these reports is testimonial under the logic of the four Williams dissenters because it was “ ‘made under circumstances which would lead an objective witness reasonably to believe that the statement would be available for use at a later trial.’ ” Crawford, 541 U.S. at 52 (quoting amici brief). Each of these reports is also testimonial under the Williams plurality’s narrower “primary purpose” test and is therefore testimonial under the reasoning of eight justices.
¶103 Unlike the DNA testing in Williams that took place before any suspect had been identified, the reports at issue here were created after investigators identified Lui as a suspect in Boussiacos’s murder on February 8,2001.8 RP at 996. Raven performed the autopsy on February 10, WSTL performed the toxicology screen on February 12, and Orchid Cellmark and Reliagene Technologies conducted the DNA testing even later — all after Lui was a suspect and solely for the purpose of collecting evidence for his prosecution. Ex. 168, at 2, 10. Consequently, here, unlike in Williams, there was a “ ‘prospect of fabrication’ ” because the analysts had an incentive to produce something “other than a scientifically sound and reliable profile.” Williams, 132 S. Ct. at 2244 (plurality opinion) (quoting Michigan v. Bryant, 562 U.S. 344, 131 S. Ct. 1143, 1157, 179 L. Ed. 2d 93 (2011)).
¶104 Also unlike the DNA testing in Williams, which the plurality described as inherently exculpatory because it was designed to exclude huge portions of the population in a search for an unknown rapist, the DNA testing performed by Orchid and Reliagene was inherently inculpatory because it was prepared “for the primary purpose of accusing a targeted individual.” Id. at 2229, 2243 (plurality opinion). *507These lab-tested samples were taken from Lui to determine whether it matched DNA samples recovered from the crime scene and for the sole purpose of proving that he had had recent sexual relations with the victim and had tied her shoelaces after death. 12 RP at 1491.
¶105 The majority claims the DNA reports are neutral and insists that they “do[ ] not . . . identify (let alone inculpate) anyone,” majority at 488, because they are merely “affirmation [s] of fact” that “a given DNA donor has certain generic characteristics.” Id. It is difficult to see how the reports do anything but identify Lui as the source of cells and semen found on the victim’s body. Ex. 136, at 4. Indeed, the majority lauds DNA tests for their reliability in obtaining convictions of identified individuals. Majority at 492-93. Moreover, if the majority is correct that a negative blood test for nicotine is inculpatory, id. at 494, how is a DNA report that positively identifies Lui as the contributor and excludes the only other suspect not at least equally as inculpatory? The majority reasons that the reports are distinguishable because a toxicology test, unlike a DNA profile, is inculpatory without having to compare the results to anything. See id. at 494 n.ll. The majority, however, fails to explain how the toxicology report indicating the victim had no nicotine in her bloodstream is at all inculpatory unless compared to Lui’s statement that she may have been attacked when she went out to smoke a cigarette.
¶106 Additionally, although the majority characterizes the DNA report as “meaningless ‘gobbledygook’ ” that required Gina Pineda’s expertise to decode, id. at 489 (quoting 12 RP at 1538), the graphic Pineda referred to was prepared as a visual aid for trial, and it is both readable and meaningful to a layperson. The graphic Pineda referred to on the stand takes the familiar form of a chart. See Ex. 136, at 4. Each column displays a source of DNA, and each line shows a potential for a genetic match. Id. The more matches between a suspect and a sample, the greater the likelihood *508that the sample contains the suspect’s DNA. While expertise is necessary to calculate precisely how probative any given result is in terms of probability, anyone can see that there are far more matches between the samples and Lui’s DNA than any other potential suspect. Thus, the report is plainly inculpatory and hardly “gobbledygook.” But even if the report were written in hieroglyphics, the Sixth Amendment guarantees a defendant the right to confront all witnesses against him, not just those who offer easy-to-understand testimony or who openly profess the defendant’s guilt.
¶107 Even the State concedes that the reports offered against Lui implicate the confrontation clause under the Williams plurality’s primary purpose test. State’s Suppl. Br. Addressing Williams v. Illinois and Article I, Section 22 of the Wash. Constitution at 5 n.5 (admitting that the plurality’s narrow definition “might not apply’ to the reports in this case because Lui had already been identified as a suspect). The State is correct. Whether the DNA report meets the majority’s pliable definition of “inculpatory,” it and each of the other reports at issue satisfy the definition of “testimonial” applied by eight of the justices who decided Williams and by a majority of the Court in Melendez-Diaz and Bullcoming.
¶108 Raven’s recorded temperature measurements are also testimonial under Melendez-Diaz and Bullcoming, and under the Williams plurality’s narrower “primary purpose” test. Raven recorded these readings from Boussiacos’s body and the outside air where her body was found on February 8, shortly before the State identified Lui as a suspect. See 10 RP at 1353-56. However, unlike the DNA testing in Williams, these temperature readings were not recorded to meet an ongoing emergency or exclude potential suspects— they were designed to determine the time of the victim’s death. 14 RP at 1809.
¶109 Although the majority claims recording this temperature reading was not an inculpatory act, majority at *509493, the only conceivable purpose for recording it, and for determining the time of Boussiacos’s death, was to support a criminal prosecution. An objective witness in Raven’s circumstances would anticipate the use of her measurements in a later trial. See State v. Mason, 160 Wn.2d 910, 921-23, 162 P.3d 396 (2007) (noting that “the test is objective”), abrogated in part on other grounds by Giles v. California, 554 U.S. 353, 128 S. Ct. 2678, 171 L. Ed. 2d 488 (2008); see also Williams, 132 S. Ct. at 2243 (plurality opinion) (noting that “the primary purpose that a reasonable person would have ascribed to the statement” controls the inquiry). Because the temperature measurements were recorded not by a third-party private laboratory but by a state investigator at a murder scene for the primary purpose “to establish or prove past events potentially relevant to later criminal prosecution,” Davis, 547 U.S. at 822, these results would satisfy even the narrower test for testimonial statements advanced by the Williams plurality.
fllO Raven’s autopsy report and the WSTL toxicology report were no less inculpatory or testimonial. Even the majority concedes that they met its test. Majority at 494-95. They were prepared to prove the cause and manner of Boussiacos’s death, and were relied on at trial to prove that the killing was intentional and that she had no nicotine in her system — not to exclude a class of suspects. 10 RP at 1375-99; 14 RP at 1830-31, 1850-51. None of these tests were designed to meet an “ongoing emergency” or to catch an unidentified criminal who “was still at large.” Williams, 132 S. Ct. at 2243 (plurality opinion).
fill Like the testimonial DNA reports created by Nasir and Vendetes that were introduced to the jury through Pineda’s testimony, Raven’s autopsy report was testimonial and introduced into evidence through Dr. Richard Harruff’s testimony. Raven’s autopsy report qualifies as “testimonial” under any test. It is testimonial under the United States Supreme Court’s “primary purpose” test, the narrower Williams plurality test, and even the majority’s new “wit*510ness against” test. See majority at 494-95. Like any other autopsy report, it is not “machine-generated” and it contains numerous attestations of fact, and Raven’s account of the victim’s body when found and the cause and manner of death can be described only as inculpatory.
|112 While medical examiners may perform autopsies for reasons both related and unrelated to trial, the confrontation clause does not require that trial testimony be the sole purpose for an out-of-court statement — only that it be the primary purpose. Bryant, 131 S. Ct. at 1155; Davis, 547 U.S. at 822. This test is objective, not subjective; it asks whether a reasonable person standing in the examiner’s shoes would understand the report’s “evidentiary purpose” and “ ‘believe that the statement would be available for use at a later trial.’ ” Melendez-Diaz, 557 U.S. at 311 (quoting Crawford, 541 U.S. at 52). Raven’s autopsy report meets this test.
¶113 On the front page of the report, Raven signed her name to her opinion that “[t]he cause of death ... is due to asphyxia due to neck compression. The manner of death is classified as homicide.” Ex. 168, at 1. Regardless of whether Raven complied with a statute when she wrote those words, no reasonable person would have any doubt that the report would be offered against a defendant at a murder trial. Indeed, the very reason medical examiners have a statutory duty to prepare written autopsy reports is to further and support the criminal prosecution of persons accused of assault, rape, murder, and other serious crimes. No one subject to the statutory command to provide such a report “upon the request of the prosecuting attorney” could misunderstand its purpose or significance.14 RCW 68.50.106.
*511¶114 If there was ever any doubt that autopsy reports satisfy the primary purpose test for testimonial evidence, Melendez-Diaz erased it. See 557 U.S. at 318 n.5 (naming autopsies among the class of forensic analyses subject to the confrontation clause), 322 (“whatever the status of coroner’s reports at common law in England, they were not accorded any special status in American practice” (citing Crawford, 541 U.S. at 47 n.2)), 335 (Kennedy, J., dissenting) (acknowledging that autopsy reports would satisfy the majority’s test).
C. The Results of Genetic Testing Are Testimonial
¶115 The majority contends that the test results reported by analysts at Orchid and Reliagene are not subject to the confrontation clause because genetic tests yield “ ‘raw, machine-produced data’ ” that is not testimonial. Majority at 479 (quoting Bullcoming, 131 S. Ct. at 2714). The majority is wrong, both about the law and about these tests. Although Pineda testified that she examined raw machine data to prepare her report, 12 RP at 1507, the State did not offer this data into evidence and the jury never saw it. The jury did see excerpts from the reports prepared by analysts Nasir and Vendetes that purported to summarize the results of this testing.15 Ex. 136, at 4; see 12 RP at 1552,1540. *512It is this testimonial report that implicates the confrontation clause, and Lui had a right to confront Nasir and Vendetes about the results they recorded, no matter how these were generated.
¶116 As for the DNA reports, they were plainly the work of human hands and a human mind, as was the rest of the presentation of which they were a part. Ex. 136, at 1-4. It does not matter that they summarized the results of repeated DNA testing, much of which was performed by machine; virtually all forensic lab work is performed by or in conjunction with machines, just as police work involves radar detectors, breath test devices, and license plate readers. Justice Sonia Sotomayor’s caution about the scope of the Court’s holdings in Bullcoming carves out only raw, machine-produced data, not all testimony that purports to relate the results of machine testing, about which the witness may just as easily lie or be mistaken as when relaying any other fact. See 131 S. Ct. at 2722 (Sotomayor, J., concurring).
¶117 The majority concedes that the WSTL toxicology report was inculpatory and therefore implicated Lui’s confrontation rights, majority at 494, though chemical analysis of blood is hardly less machine aided than DNA testing. Reports attesting to machine-generated results were squarely at issue in both Melendez-Diaz and Bullcoming.
¶118 In Melendez-Diaz, a majority of the Court rejected the distinction between testimony about tests that are “ ‘neutral’ ” and “ ‘scientific,’ ” and less-reliable “testimony recounting historical events, which is ‘prone to distortion or manipulation.’ ” 557 U.S. at 317. Instead, the Court insisted that analysts must testify about recorded laboratory results even if they possess “the scientific acumen of Mme. Curie and the veracity of Mother Teresa.” Id. at 319 n.6. And in Bullcoming, the Court categorically rejected the argument that an analyst was not subject to confrontation because he *513was a “ ‘mere scrivener’ ” who reported factual output from a machine, exactly as DNA analysts Nasir and Vendetes did here when they transcribed the results of testing into a human-readable report. 12 RP at 1552, 1567; see Bull-coming, 131 S. Ct. at 2714-15 (rejecting the distinction between neutral and inculpatory facts, and noting it would reach far beyond laboratory testing to all testimony about “factual conditions or events,” such as “ ‘the light was green’ ”). Moreover, Nasir and Vendetes were not mere scriveners; “[Nasir] was the one to interpret the results and write the reports,” 12 RP at 1552, and Vendetes decided whether to exclude certain peaks in the DNA from the Cellmark report, id. at 1569, 1572.
¶119 It is true that photographs, blood samples containing DNA, and indeed all forms of physical evidence are not testimonial, but this is because they do not testify to anything. Physical evidence does not assert any fact beyond its own existence. State v. Appleby, 289 Kan. 1017, 221 P.3d 525, 551 (2009) (“DNA itself is physical evidence and is nontestimonial”); Herrera v. State, 367 S.W.3d 762, 773 (Tex. Ct. App. 2012) (“[a]n autopsy photograph ... is not a testimonial statement”). But physical evidence is made a part of testimony, and a confrontation right arises, when a human witness attests to some fact about the physical evidence that is not apparent from the fact of the evidence itself.
¶120 The majority’s concern that respecting defendants’ Sixth Amendment confrontation rights in this context will swamp the system with unnecessary witnesses is hyperbolic. No one disputes that the State can offer evidence without producing a witness to testify to each step in the chain of custody. See Melendez-Diaz, 557 U.S. at 311 n.l (“we do not hold . . . that anyone whose testimony may be relevant in establishing the chain of custody, authenticity of the sample, or accuracy of the testing device, must appear in person as part of the prosecution’s case”). The majority’s argument that not every analyst who operates a machine or *514performs a step in the analysis becomes a “witness” is a straw man. The confrontation clause obviously does not require every person who touches physical evidence to be cross-examined. No one claims that it does, and the majority needlessly belabors this point.
¶121 By definition, the confrontation guaranty applies only to witnesses who testify against the defendant. Crawford, 541 U.S. at 52. Witnessing a laboratory test is a necessary condition of being a “witness” under Bullcoming, but it is not a sufficient condition that gives rise to a confrontation right under Crawford and Melendez-Diaz. An analyst who touches a sample does not testify against a defendant any more than a police officer does when he handles physical evidence at a crime scene. A confrontation right arises only when that witness testifies against a criminal defendant and when the State offers that testimony against the defendant.
¶122 The confrontation clause obviously does not give Lui the right to cross-examine the analysts who took shoelace cuttings and processed the samples, but not because they work in a laboratory or because their analysis was based on the output of machines. Lui has no right to cross-examine these analysts for the simple reason that their testimony was not offered against him. Unlike the other analysts who worked on these samples, Nasir and Vendetes each prepared a report testifying to their observations of the results. 12 RP at 1552, 1567; Ex. 136, at 4. Moreover, Nasir and Vendetes did more than simply record the data — they created it. See Melendez-Diaz, 557 U.S. at 322-23. They interpreted the data, deciding what to report and what to exclude. 12 RP at 1552, 1569, 1572. These reports are testimonial, and because the State offered it against Lui at trial, he had the right to cross-examine the analysts who prepared them.
¶123 Not only does the confrontation clause draw no distinction between testimony about laboratory results and testimony about other events, but the majority’s character*515ization of this testimony as inherently accurate is itself flawed. Scientific evidence from test results is not as objective and factual as the majority believes. Before a DNA test result is reported on a graph or chart, a human analyst picked up each sample, made notations on a report (or did not), used care to ensure the sample’s integrity (or did not), followed the laboratory’s protocol to ensure accuracy (or did not), reported all DNA peaks (or did not), and recorded his or her observations accurately (or invented them out of whole cloth).
¶124 To the extent that forensic testing depends on a human analyst, it contains the same potential for human error and calls just as strongly for cross-examination as testimony by a “conventional” witness about “conventional” topics. See Williams, 132 S. Ct. at 2264-65 (Kagan, J., dissenting) (arguing that “[c]ross-examination of the analyst is especially likely to reveal whether vials have been switched, samples contaminated, tests incompetently run, or results inaccurately recorded”).
¶125 The serious and growing problem of “drylabbing,” in which forensic analysts report results of tests that were never performed, also belies the majority’s confidence that the results of supposedly objective, factual machine data cannot easily be manipulated or manufactured. In Massachusetts, a state crime lab analyst with fraudulent scientific credentials attested to fictional drug test results in roughly 34,000 cases without drawing any concern from her supervisors. See J. Hr’g of the H. Comms. on Pub. Health, Pub. Safety & Post-Audit Oversight, 187th Gen. Court (Mass. 2012) (statement of Dr. JudyAnn Bigby, Sec’y, Exec. Office of Health & Human Servs., Nov. 28, 2012, available at http:// www.mass.gov/eohhs /docs /bigby-testimony-crime-lab-112812 . doc; Brittany Brady, Chemist in Massachusetts Drug Sample Case Lied about Degree, CNN Justice (Sept. 26, 2012, 9:59 AM), http://www.cnn.com/2012/09/25/justice/massachusetts -chemist. Despite the majority’s attempt to pin these incidents on isolated and “hypothetical rogue analyst [s],” at least one *516other analyst at a Massachusetts crime lab was also tampering with evidence. Majority at 492; Zach Howard, Second Chemist Charged in Massachusetts Drug Lab Scandal, Reuters (Apr. 1 2013, 5:23 PM), http://www.reuters.com/art icle/2013/04/01/us-usa-massachusetts-crimelab-idUSBRE93 00KJ20130401.
¶126 Laboratory misconduct is not limited to Massachusetts. Forensic examiner Jonathan Salvador’s “lack of attention to detail” and “lack of understanding of chemistry” were rewarded with numerous promotions at the Public Safety crime lab in Houston, where his falsification of results has put close to 5,000 drug cases into question. James Pinkerton & Brian Rogers, Crime Lab Analyst Kept on Job Despite Shoddy Work, Houston Chron. (Apr. 6, 2013), http://www.houstonchronicle.com/news/houston-texas/hous ton/article/Crime-lab-analyst-kept-on-job-despite-shoddy -work-4413046.php (subscription required). Other examples abound. See Melendez-Diaz, 557 U.S. at 318-19 (describing incidents); Thomas J. Lueck, After Falsified Test Results, Kelly Orders Forensic Shakeup, N.Y. Times (Apr. 20, 2007), available at http://www.nytimes.com/2007/04/20/nyr egion/20chief.html?_r=l&0 (reporting that a police crime lab analyst falsely reported results of drug tests).
¶127 The majority’s breezy dismissal of the potential for laboratory fraud is particularly inappropriate given Washington’s recent history of such problems. In 2008, the head of the state crime and toxicology labs resigned after King County judges found “ethical lapses and a climate of compromise” at the facility, including a senior manager who fabricated toxicology results. State Toxicology Lab Chief Resigns over DUI Errors, KOMO News (Feb. 14, 2008, 3:38 PM), available at http://www.komonews.com/news/local /15643687.html (last updated Sept. 27, 2010). And most recently, the manager of the state crime lab in Cheney resigned after it was discovered that he had fabricated arson investigation reports in “at least five cases.” Gene Johnson, State Patrol Says Crime Lab Manager Falsified *517Work, KOMO News (Apr. 16, 2013, 8:46 AM), available at http://www.komonews.com/news/local / State-crime-lab-man ager-resigns-amid-investigation-203214851.html.
¶128 The majority’s argument that exempting laboratory reports from the confrontation clause will actually reduce drylabbing is nonsensical. Majority at 492-93. Nothing in Melendez-Diaz or any other case prevents forensic technicians from operating in teams or jointly analyzing samples, and nothing in the many recent incidents of drylabbing supports the majority’s hollow claim that teams of workers are a natural check on the problem. Criminal defendants are entitled to more than the majority’s empty assurances of reliability; they are entitled to cross-examine the witnesses who testify against them, including those witnesses whose testimony is submitted in written form. Crawford, 541 U.S. at 61.
¶129 The majority wrings its hands over the possibility that laboratory analysts might miss trials “due to sickness, travel, inclement weather, or being called to testify in another trial,” majority at 492, but, of course, so might police officers or any other eyewitnesses or expert witnesses. The majority’s concern for scheduling proves too much and would subordinate all confrontation rights of Washington citizens to the convenience of a trial calendar. Nor does the majority explain why the mortality of laboratory analysts sets them apart from other witnesses. Id. Crawford plainly anticipates that witnesses may be unavailable, holding that the testimonial statement of a witness who does not appear and is “unavailable to testify” is inadmissible unless the defendant had “a prior opportunity for cross-examination.” Crawford, 541 U.S. at 53-54.
¶130 This does not leave the State helpless or make a medical examiner’s lifespan a “ ‘statute of limitations for murder.’ ” Melendez-Diaz, 557 U.S. at 335 (Kennedy, J., dissenting) (quoting Carolyn Zabrycki, Comment, Toward a Definition of “Testimonial”: How Autopsy Reports Do Not Embody the Qualities of a Testimonial Statement, 96 Cal. L. *518Rev. 1093,1094,1155 (2008)). Photographs and other physical evidence — including fingerprints, tool marks, and samples of blood and semen — are not testimonial, and any analyst can examine them if the original analyst dies or is otherwise unavailable. Indeed as described below, although Raven was not available to testify, Harruff testified to the cause of Boussiacos’s death based almost entirely on nontestimonial photographs in Raven’s autopsy report. Photographing routine field measurements also ensures this evidence survives the investigator — -just as photographing an officer’s observations at the crime scene ensures this testimony will survive him.
¶131 Nor does the Sixth Amendment require the State to keep physical evidence it would otherwise throw away. Washington State crime labs already store, control, or reference samples for trial and for later retesting as a matter of course because of the realities of appeal and postconviction challenge, and the utility of this evidence for solving other crimes, including “cold cases.” See generally RCW 43.43.753 (establishing a state bank for DNA samples); Wash. State Patrol Forensic Laboratory Servs. Bureau, Forensic Services Guide (Jan. 2012) (describing long-term storage methods for DNA and other physical evidence). Forensic testing is neither as unique or uniquely reliable as the majority claims nor as vulnerable as it fears.
D. A Supervisor Is Not a “Witness” to a Test He or She Did Not Personally Conduct or Observe
¶132 The majority and the State assert that even if the various reports offered against Lui were testimonial, there was no error because he had an opportunity to cross-examine Harruff and Pineda, both of whom were competent to testify about the testing protocols in place at their respective workplaces. Majority at 490 (characterizing Pineda as “an experienced supervisor” who “was well informed *519about the procedures used”); see State’s Suppl. Br. at 21; State’s Suppl. Br. Addressing Williams v. Illinois and Article I, Section 22 of the Wash. Constitution at 9 (characterizing these experts as “hands-on supervisors who were intimately familiar with the relevant laboratory’s procedures”). This line of reasoning is foreclosed after Bullcoming, as there is simply no room to argue that a supervisor who did not personally observe a test is a “witness” to the results of this test.
¶133 The majority insists that the problem in Bullcoming was that the defendant “was denied effective cross-examination.” Majority at 491 (emphasis added). Because Lui was able to “effectively” cross-examine Pineda and Harruff, the majority asserts his confrontation right was satisfied. Not so. The problem in Bullcoming was not with the effectiveness of the cross-examination; the problem was that the analyst who testified about the results of a blood test was not a “witness” under the plain language of the confrontation clause. 131 S. Ct. at 2710 (holding that “[t]he accused’s right is to be confronted with the analyst who made the certification”). The majority’s endorsement of the “effectiveness” of Lui’s cross-examination also runs directly counter to Crawford?s rejection of extrinsic measures of reliability as the benchmark for the confrontation right. See 541 U.S. at 67, 69.
¶134 While the confrontation clause is concerned with the reliability of testimony, it provides “a procedural rather than a substantive guarantee.” Id. at 61. The only way to satisfy a defendant’s confrontation right is to allow him to cross-examine the witness who offers testimony — not that witness’ supervisor. In Bullcoming, the Court gave no weight to the fact that the testifying analyst was a “knowledgeable representative of the laboratory” who could “explain the lab’s processes and the details of the report.” 131 S. Ct. at 2723 (Kennedy, J., dissenting). Instead, the Court reiterated that “ ‘it will not permit the testimonial statement of one witness to enter into evidence through the *520in-court testimony of a second.’ ” Id. at 2715 (quoting Melendez-Diaz, 557 U.S. at 334 (Kennedy, J., dissenting)).
|135 Just as in Bullcoming, the experts who testified against Lui were supervisors who could testify about general testing procedures but lacked even a limited “personal . . . connection to the scientific test at issue.” Id. at 2722 (Sotomayor, J., concurring). Their earnestness and the fact that they were “well informed about the procedures used,” majority at 490, are constitutionally irrelevant. It is clear from the record that Harruff did not perform the autopsy and was not even in the building when it was performed, 10 RP 1339, and Pineda testified that she did not observe or participate in the testing at Reliagene and that she was not necessarily in Texas during the testing at Cellmark, 12 RP at 1489, 1494-95, 1567. The analysts who actually performed the autopsy, recorded the victim’s body temperature, measured blood for nicotine, and reported the results of DNA testing never took the stand, and Lui never had the opportunity to cross-examine them about their educational backgrounds, employment history, methods, or conclusions — the majority’s assurance about “effectiveness” notwithstanding.
¶136 A police supervisor would never be allowed to testify to eyewitness observations about a crime scene that were recorded by his subordinate, under the guise of providing expert testimony, no matter how closely he reviewed the notes, but that is precisely what the majority would allow here. See Davis, 547 U.S. at 826 (holding that the confrontation clause may not be “evaded by having a note-taking policeman recite the . . . testimony of the declarant” (emphasis omitted)). The only difference in Lui’s case is the presence of a science degree and a lab coat, and after Melendez-Diaz and Bullcoming, these trappings have no constitutional significance. Lui was denied his right to confront the witnesses who had firsthand knowledge of the testimonial evidence the State offered against him.
*521 E. Expert Testimony about Reports Prepared by a Nontestifying Witness Violates the Confrontation Clause
¶137 The State also argues that Lui’s confrontation right was not violated because Harruff and Pineda did not offer any reports into evidence but only relied on them as the basis for their expert opinions. We stayed our decision in hopes that Williams would resolve the question of expert testimony about reports prepared by nontestifying witnesses. Unfortunately, it did not; indeed, Williams left the area muddier than before. Nor does the majority resolve this question. We should confront it and recognize that expert testimony violates a defendant’s confrontation right when the expert acts as a “conduit” by reading the substance of a testimonial report into evidence.
¶138 In Lui’s case, both the trial court and the Court of Appeals held that his confrontation rights were not violated by expert testimony because the underlying reports were offered as a basis for expert opinion and not for their truth. State v. Lui, 153 Wn. App. 304, 322-25, 221 P.3d 948 (2009); 10 RP at 1368. The Williams plurality advanced this theory as an alternative to its narrowed primary purpose test, opining that “[o]ut-of-court statements that are related by the expert solely for the purpose of explaining the assumptions on which that opinion rests are not offered for their truth and thus fall outside the scope of the Confrontation Clause.” Williams, 132 S. Ct. at 2228. As noted, five justices have repeatedly rejected this theory.
¶139 Justice Thomas protested that there is “no meaningful distinction between disclosing an out-of-court statement so that the factfinder may evaluate the expert’s opinion and disclosing that statement for its truth.” Id. at 2257 (Thomas, J., concurring in judgment). The four dissenters agreed that the DNA report was offered for its truth, reasoning that “to determine the validity of the witness’s conclusion, the factfinder must assess the truth of *522the out-of-court statement on which it relies.” Id. at 2268-69 (Kagan, J., dissenting). The Washington Court of Appeals agreed, concluding that whether a statement falls under a hearsay exception “does not save it from confrontation clause analysis” if it is also offered for its truth. State v. Fraser, 170 Wn. App. 13, 23, 282 P.3d 152 (2012).
¶140 In Lui’s case, the State did not offer the lab results to impeach his credibility or rebut his testimony. Instead, just as in Williams, the only conceivable purpose for expert testimony about DNA testing, temperature readings, autopsy findings, and toxicology results was to introduce these results for their truth. See David H. Kaye, David E. Bernstein, & Jennifer L. Mnookin, The New Wigmore: A Treatise on Evidence: Expert Evidence § 4.10.1, at 196 (2d ed. 2011) (noting that “[t]o use the inadmissible information in evaluating the expert’s testimony, the jury must make a preliminary judgment about whether this information is true”).
¶141 The fiction that these results could “assist the jury in deciding what weight to give” the expert’s opinion is baffling. Lui, 153 Wn. App. at 323 n.19. Before the jury could give testimony about these test results any weight, it had to first determine whether the underlying reports were true. And, it is unclear what else it could conclude from reports stating that genetic material found at the crime scene matched genetic material taken from the suspect, that no drugs were found in the victim’s system, and the temperature of the victim’s body when she was found. Ex. 136, at 4; Ex. 168, at 10.
¶142 Most courts to consider this question since Melendez-Diaz and Bullcoming have rejected the categorical approach advanced by the Williams plurality and by the State. Instead of concluding as a matter of law that expert testimony cannot violate a defendant’s confrontation right because it does not come in for its truth, these courts ask whether the expert testified from personal knowledge and opinion or was merely a “conduit” for the underlying testi*523monial report. See, e.g., State v. Gonzales, 2012-NMCA-034, ¶ 8, 274 P.3d 151,153 (cautioning that without safeguards, Fed. R. Evid. 703 would allow prosecutors to “use surrogate witnesses disguised as experts in order to introduce evidence that otherwise would trigger the Sixth Amendment’s Confrontation Clause”); United States v. Pablo, 696 F.3d 1280, 1287-89 (10th Cir. 2012) (warning that expert testimony can be “little more than a backdoor conduit for an otherwise inadmissible statement”); State v. Kennedy, 229 W. Va. 756, 773, 735 S.E.2d 905 (2012) (holding “to the extent that [a witness] is a ‘mere conduit’ for the opinions of the authoring pathologist, such testimony violates the Confrontation Clause”). This “conduit” test respects that evidence rules like Fed. R. Evid. 703 are not legal conclusions and have no claim to precedence over the Sixth Amendment.
¶143 The conduit test’s distinction between permissible and impermissible expert opinion testimony closely follows the rejection of other forms of surrogate testimony in Davis, Melendez-Diaz, and Bullcoming. It also addresses the concern that a bright-line rule permitting experts to read testimonial reports into evidence would put the confrontation clause once again at the mercy of the evidence code. See, e.g., Williams, 132 S. Ct. at 2256 (Thomas, J., concurring in judgment) (“I do not think that rules of evidence should so easily trump a defendant’s confrontation right”); Kaye, Bernstein & Mnookin, supra, § 4.10.1, at 196-97 (urging courts not to “permit an end-run around a constitutional prohibition” by accepting “[t]he factually implausible, formal claim that experts’ testimony is being introduced only to help in the evaluation of the expert’s conclusions, but not for its truth”). As this court has previously held, ER 703 “ ‘ “was not designed to enable a witness to summarize and reiterate all manner of inadmissible evidence.” ’ ” State v. DeVries, 149 Wn.2d 842, 848 n.2, 72 P.3d 748 (2003) (quoting State v. Martinez, 78 Wn. App. 870, 880, 899 P.2d 1302 (1995) (quoting 3 David W. Louisell & Christopher B. Mueller, Federal Evidence § 389, at 663 (1979))).
*524¶144 We should hold that ER 703 is not a per se exemption from the confrontation clause for expert testimony and formally adopt the conduit test. Expert testimony based on inadmissible testimony satisfies the defendant’s confrontation right if the expert testifies to his own opinion and conclusions, but not if he merely funnels the testimony of a nontestifying witness into evidence.
¶145 Applying the conduit test to the circumstances at hand, we should conclude that Harruff’s testimony about temperature measurements taken at the scene by Raven violated Lui’s confrontation rights. Harruff did not take these measurements or observe the test, and his recitation of Raven’s figures did not reflect his personal knowledge or opinion. 10 RP at 1354, 1369. While the majority correctly holds that Harruff’s testimony about the results of the WSTL toxicology screen violated Lui’s confrontation right, majority at 494, we should explain why this is so. Harruff had no personal knowledge about the actual testing performed by the laboratory, or even who performed the tests, and did not offer his professional opinion about the results. 10 RP at 1397-98. Harruff’s role in introducing these results was that of “a well-credentialed conduit for testimonial hearsay.” United States v. Ramos-González, 664 F.3d 1, 5 (1st Cir. 2011).
¶146 While Pineda’s testimony about the results of DNA testing also included her expert opinion about what the results reported by Nasir and Vendetes meant, she testified to the results of six tests that she did not perform or observe and about which she had no personal knowledge. 12 RP at 1484, 1489, 1494-95. Each of these tests attested to the presence or absence of genetic information in a particular sample. Ex. 136, at 4. Although the report was not admitted, it was shown to the jury, and Pineda effectively read the results into evidence by her testimony, describing results that were inextricably linked to her conclusions. 12 RP at 1517 (“we obtained a 10 locus profile”; “[w]e noted that [Lui] cannot be excluded as one of the major *525DNA donors”), 1519 (“because the peaks that [Lui] possesses [ ] were also detected in the shoelaces ... he cannot be excluded as a major donor”).
¶147 The confrontation clause requires testimony by at least one analyst who personally performed or observed the test at issue, and the expert witness evidence rule offers no exception. The confrontation clause cannot be evaded by substituting “a note-taking policeman” for the declarant. Davis, 547 U.S. at 826. Transforming the policeman into a DNA expert and the governing rule from ER 701 to ER 703 does not alter the constitutional equation. Although Pineda added her expertise to the analysis, she also read the results to the jury and denied Lui his right to confront the witnesses, analysts Nasir and Vendetes, who actually observed these tests and testified to the results in their reports.
¶148 Harruff’s testimony about Raven’s autopsy report and autopsy photographs presents a different matter. As discussed above, photographs and other physical evidence are not testimonial and an expert may refer to them without violating a defendant’s confrontation right. See State v. Roberts, 142 Wn.2d 471, 522, 14 P.3d 713 (2000) (permitting testimony from a blood splatter expert who reviewed only documentary evidence of photographs and videotape and examined the victim’s clothing and chair).
¶149 Harruff’s testimony about the nature of Boussiacos’s injuries and the cause of her death was based almost entirely on photographs that the State entered into evidence. 10 RP at 1375-97. Although Harruff testified to bruising of the neck muscles and bruising under the scalp that were not apparent in the photographs, this testimony constituted a tiny fraction of his overall testimony. Id. at 1392. Moreover, his testimony that it takes four minutes to die from strangulation came entirely from his expertise and personal knowledge; it did not parrot Raven’s report or even refer to it. Id. at 1385, 1397.
*526¶150 We should hold that Harruff’s testimony about Boussiacos’s injuries, the cause of her death, and the condition of her body based on autopsy photographs was an appropriate exercise of his expert opinion and did not violate the confrontation clause. See Commonwealth v. Avila, 454 Mass. 744, 912 N.E.2d 1014, 1029 n.19 (2009) (holding that a medical examiner’s opinion about the cause of a wound is permissible if based on a photograph properly admitted into evidence, but not if it recites findings in the autopsy report).
F The Confrontation Clause Is Not an Unbearable Burden
¶151 Ultimately, the majority argues that we must deny defendants the right to confront laboratory analysts because the burden it places on the State is too heavy, apparently indignant that defendants granted this right by Melendez-Diaz have actually exercised it.16 Majority at 491-92. While the confrontation clause places a burden on the courts and prosecutors of Washington State, this is hardly a persuasive argument for dispensing with one of the bedrock guaranties of our criminal justice system. The Sixth Amendment also guarantees to criminal defendants the right to a speedy and public trial, to have facts (even “neutral” and “scientific” facts) found by a jury, and to be appointed a competent lawyer at no cost. See U.S. Const. amend. VI. Each of these guaranties has cost our State incalculable money, time, and lost convictions, and the costs continue to mount. If the majority is willing to exempt laboratory analysts from cross-examination to save a little, why not strike confrontation entirely, or do away with jury trials and court-appointed attorneys, and save much more?
¶152 Notwithstanding the United States Supreme Court’s admonition that the confrontation clause, like the *527right to trial by jury and the privilege against self-incrimination, “is binding, and we may not disregard it at our convenience,” Melendez-Diaz, 557 U.S. at 325, this burden is not as heavy as the majority claims. Indeed, we already live with all of the consequences the majority imagines, and have for the four years since Melendez-Diaz settled the questions with which the majority belatedly grapples. The sky has not fallen in this time, here or in any other state. There is at least one practical reason for this: the United States Supreme Court has now twice endorsed the constitutionality of notice-and-demand statutes exactly like Washington’s, which condition the defendant’s confrontation right on the timely filing of an objection to the State’s offer of evidence. See id. at 326-27; Bullcoming, 131 S. Ct. at 2718; CrR 6.13(b). While CrR 6.13(b) does not excuse the State from its obligations under the confrontation clause, it places the burden of requesting analyst witnesses squarely where it belongs: on the criminal defendant. See Melendez-Diaz, 557 U.S. at 327 (“The defendant always has the burden of raising his Confrontation Clause objection; notice-and-demand statutes simply govern the time within which he must do so.”); State v. Schroeder, 164 Wn. App. 164,167-68, 262 P.3d 1237 (2011) (finding that the defendant waived his right to confrontation on a piece of evidence by failing to file a timely objection).
¶153 Finally, although the State has a strong interest in the efficient provision of criminal justice, I cannot agree that confrontation runs entirely counter to this interest. The sheer number of reversals and new trials that attend each new drylabbing scandal illustrate the cost of insulating laboratory analysts from cross-examination. Requiring analysts to testify to their own qualifications and represent their own work gives the State a strong incentive to weed out incompetent technicians and gives laboratories a strong incentive to develop programs and procedures that thrive in “the crucible of cross-examination.” Crawford, 541 U.S. at 61. While it may be bitter, confrontation is good medicine for *528the criminal justice system and the integrity of the convictions it obtains. Washington prosecutors and courts are more than able to shoulder the burden the Sixth Amendment demands.
G. The Error Is Not Harmless
¶154 Confrontation clause errors are subject to constitutional harmless-error analysis. Jasper, 174 Wn.2d at 117 (citing Delaware v. VanArsdall, 475 U.S. 673,684,106 S. Ct. 1431, 89 L. Ed. 2d 674 (1986)). This test is significantly more stringent than that for violations of court rules and other nonconstitutional errors. See State v. Robinson, 153 Wn.2d 689, 697, 107 P.3d 90 (2005) (citing State v. Templeton, 148 Wn.2d 193, 220, 59 P.3d 632 (2002)). Under the constitutional error standard, prejudice is presumed and the State must show “beyond a reasonable doubt that the error complained of did not contribute to the verdict obtained.” Chapman v. California, 386 U.S. 18, 24, 87 S. Ct. 824, 17 L. Ed. 2d 705 (1967); State v. Stephens, 93 Wn.2d 186, 190-91, 607 P.2d 304 (1980).
¶155 The Court of Appeals accepted the State’s argument that any confrontation violation at Lui’s trial was harmless, finding that if the State’s experts tunneled testimonial evidence to the jury, “[t]here is no reasonable probability this evidence contributed prejudicially to the verdict.” Lui, 153 Wn. App. at 320 n.15. Because the State did not produce overwhelming untainted evidence of Lui’s guilt, I disagree.
¶156 The “ ‘overwhelming untainted evidence’ ” test considers the untainted evidence admitted at trial to determine “if it is so overwhelming that it necessarily leads to a finding of guilt.” State v. Smith, 148 Wn.2d 122,139, 59 P.3d 74 (2002) (citing State v. Guloy, 104 Wn.2d 412, 426, 705 P.2d 1182 (1985)). This test “ensures that a conviction will be reversed where there is any reasonable possibility that the use of inadmissible evidence was necessary to reach a guilty verdict.” Id.
*529¶157 Although the State offered sufficient evidence of Lui’s guilt to permit a reasonable jury to convict him of murder, the untainted evidence was hardly overwhelming. This untainted evidence included Harruff’s testimony that Boussiacos was intentionally strangled and testimony impeaching Lui’s credibility. Lui told investigators that the night Boussiacos disappeared he slept on the couch and heard nothing, and that when he awoke between 7:00 and 8:00 a.m., Boussiacos’s car was gone. Ex. 43, at 34. However, phone records show that Lui called his sister at roughly 1:00 a.m. and left a message; when confronted with these records, he gave an unlikely story that he had rolled over on his phone while sleeping. Id. at 31-32; 7 RP at 809-10, 827. The jury also heard that the couple’s downstairs neighbors were awakened shortly after 3:00 a.m. by the sound of someone walking around upstairs. 5 RP at 566, 583-84. During questioning, Lui downplayed Boussiacos’s discovery of his infidelity, insisting that they had “put that aside” despite his statements to friends and a 911 operator that the two had called off the wedding shortly before her disappearance. 14 RP at 1819-20; Ex. 43, at 25-28.
¶158 Lui’s credibility was most damaged by his lie about Boussiacos’s ring. Police did not find her engagement ring on her body or in a search of her purse, where friends testified she kept it. 14 RP at 1703. Lui denied having the ring and speculated that Boussiacos’s mother had it or, alternatively, that Boussiacos had been wearing it when she left for California. 10 RP at 1431; Ex. 169, at 50-51, 80. At trial, the State established that Lui gave an identical ring to his current wife, who wore it until police took it into evidence. 12 RP at 1609-22, 1628-29; 14 RP at 1702-08. Even Lui’s lawyer acknowledged that he had lied about the ring, calling it a “monumental mistake.” 14 RP at 1866.
¶159 While a jury certainly could have found that this cumulative untainted evidence was probative of Lui’s guilt, Washington courts have never found confrontation clause errors harmless beyond a reasonable doubt under such *530fragmentary and circumstantial evidence. See State v. Gonzales Flores, 164 Wn.2d 1, 18-20, 186 P.3d 1038 (2008) (finding harmless error where the State presented untainted testimony, audio and video recordings, physical evidence, and defendant admissions); State v. Davis, 154 Wn.2d 291, 304-05, 111 P.3d 844 (2005) (State presented untainted testimony by victim and officers and photographs of victim injuries); State v. Damon, 144 Wn.2d 686, 693, 25 P.3d 418 (2001) (defendant did not contest that he had committed all the criminal acts charged); State v. Powell, 126 Wn.2d 244, 268, 893 P.2d 615 (1995) (untainted eyewitness testimony of repeated episodes of past abuse, including prior attempts at strangulation); State v. Whelchel, 115 Wn.2d 708, 728-29, 801 P.2d 948 (1990) (untainted eyewitness testimony and defendant confession); Guloy, 104 Wn.2d at 420-21 (undisputed evidence that defendant engaged in criminal conspiracy to murder both victims); State v. Hopkins, 134 Wn. App. 780, 792, 142 P.3d 1104 (2006) (untainted victim testimony, eyewitness testimony, and defendant confession); State v. Saunders, 120 Wn. App. 800, 812-13, 86 P.3d 232 (2004) (untainted defendant admissions); State v. Thomas, 91 Wn. App. 195, 203, 955 P.2d 420 (1998) (untainted eyewitness testimony and physical evidence); State v. Folkerts, 43 Wn. App. 67, 73-75, 715 P.2d 157 (1986) (untainted eyewitness testimony, physical evidence, and defendant admissions); State v. Roberts, 31 Wn. App. 375, 380, 642 P.2d 762 (1982) (untainted eyewitness testimony). The State’s untainted evidence against Lui was not in the same class.
¶160 Washington courts have consistently refused to find harmless error in cases like Lui’s, where the untainted evidence was incomplete and suggested, but did not overwhelmingly establish, the State’s theory. See State v. Grenning, 169 Wn.2d 47, 59-60,234 P.3d 169 (2010); State v. Maupin, 128 Wn.2d 918, 928-30, 913 P.2d 808 (1996); State v. Easter, 130 Wn.2d 228, 242-43, 922 P.2d 1285 (1996); State v. St. Pierre, 111 Wn.2d 105, 119-20, 759 P.2d 383 *531(1988); Stephens, 93 Wn.2d at 190-91; State v. Romero, 113 Wn. App. 779, 794-95, 54 P.3d 1255 (2002); State v. McDaniel, 83 Wn. App. 179, 187-88, 920 P.2d 1218 (1996); State v. Vargas, 25 Wn. App. 809, 815-16, 610 P.2d 1 (1980).
¶161 Moreover, the tainted testimony in this case was substantial and crucial to the State’s theory. The State argues that the DNA results obtained by Orchid and Reliagene were inconclusive and that omitting Pineda’s testimony would not have affected the outcome. It is true that the presence of Lui’s DNA on the victim’s shoelaces was not conclusive because the two shared a home. The “Y-STR” testing used in this case also could not exclude the possibility that Lui’s son had touched her shoelaces. 12 RP at 1517-19. However, the DNA sample from the victim’s shoelaces was the only piece of evidence linking Lui to the State’s theory that he had dressed her after death and carried her body to the trunk of her car. See 14 RP at 1832-35. Jodi Sass, a forensic scientist in the DNA unit of the Washington State Patrol Crime Laboratory, testified that she had obtained a trace male genetic component from the oddly tied shoelaces found on Boussiacos’s tennis shoes, but she was not able to generate a DNA profile from the sample. 9 RP at 1228-33. Without Pineda’s testimony about the results of these tests, the State would have presented photographs of oddly tied shoelaces but no evidence that Lui tied them after her death.
¶162 Because Lui and the victim had a consensual sexual relationship, the State contends that Pineda’s testimony about trace amounts of Lui’s semen in the vaginal swab was also not decisive and was not relevant to a material element of the crime. However, Pineda’s testimony undercut Lui’s claim that he and the victim had not had intercourse for two weeks before her disappearance. 14 RP at 1828-29; Ex. 43, at 21-23. The State highlighted this inconsistency to the jury, not just to undercut Lui’s credibility but as a possible motive for the killing. 14 RP at 1828-30.
*532¶163 The State argues that Harruff’s tainted testimony about the time of Boussiacos’s death, based on Raven’s temperature readings, was equally supportive of Lui’s defense that she was killed by a third person after leaving for the airport some time after the night of February 2. 10 RP at 1355-56. While the State was required to prove only that Boussiacos died sometime between February 2, when she disappeared, and February 8, when her body was found, Raven’s temperature readings supported the State’s theory that Lui murdered her the night she disappeared. See 14 RP at 1809.
¶164 Finally, the State contends that Harruff’s testimony about the lack of nicotine in Boussiacos’s blood was irrelevant since it did nothing more than undercut idle conjecture by Lui. 10 RP at 1430. Whether the victim was a smoker was not a material element of the crime charged, but the testimony hurt Lui’s credibility and went to the State’s larger narrative that Lui had lied to investigators from the beginning.
¶165 The tainted evidence, particularly the DNA report, was crucial to the State’s case. Pineda’s testimony about the results of this testing linked Lui to the victim’s shoelaces and undercut his story that the two had not had sexual relations in weeks. Harruff’s testimony about the results of the WSTL toxicology screen made Lui’s initial story look less like speculation and more like an intentional lie, and his testimony about the time of death lent support to the State’s theory. The most the State can show is that it introduced sufficient untainted evidence to permit a reasonable jury to find Lui guilty. This showing does not satisfy the constitutional harmless error test, and we should conclude that the errors were not harmless beyond a reasonable doubt.
CONCLUSION
¶166 Consistent with United States Supreme Court precedent, I would recognize that the reports at issue in this *533case are testimonial and Lui was denied his Sixth Amendment right to confront the witnesses against him. The majority’s newfound interpretation of the Sixth Amendment is based on the very rationale a majority of the United States Supreme Court has rejected in cases from Crawford to Williams.
¶167 Because the four reports at issue are testimonial, Lui was entitled to cross-examine the witnesses who prepared them and not settle with questioning well-credentialed conduits about their testimonial evidence.
¶168 The majority’s fear that respecting the constitutional rights of Lui and similarly situated defendants will unduly burden the criminal justice system is exaggerated. Moreover, it provides no justification for charting an entirely new course in conflict with established precedent. I would reverse the Court of Appeals and remand this matter for a new trial. Accordingly, I respectfully dissent.
Owens and Fairhurst, JJ., concur with Stephens, J. Chambers, J. Pro Tem., concurs in result only.Reconsideration denied March 13, 2014.
While Justice Clarence Thomas also wrote separately in Melendez-Diaz, this does not erase the precedential value of the majority opinion he signed. See United States v. King, 194 F.R.D. 569, 576 n.7 (E.D. Va. 2000) (explaining that “ ‘[a] decision with only a simple concurrence should not be considered a plurality decision’ ” (quoting Ken Kimura, Note, A Legitimacy Model for the Interpretation *501of Plurality Decisions, 77 Cornell L. Rev. 1593,1595 n.13 (1992))). Justice Thomas knows how to concur in a judgment without signing the opinion; he did it in Davis v. Washington, which was decided two years before Melendez-Diaz. 547 U.S. 813, 834, 126 S. Ct. 2266, 165 L. Ed. 2d 224 (2006) (Thomas, J., concurring) (writing separately because “I concur only in the judgment in Davis v. Washington”). Instead, his simple concurrence explained that the report also bore sufficient “ ‘indicia of formality ” to come within the confrontation clause. Melendez-Diaz, 557 U.S. at 329-30 (Thomas, J., concurring) (quoting Giles v. California, 554 U.S. 353, 378, 128 S. Ct. 2678, 171 L. Ed. 2d 488 (2008)). Moreover, the Court’s definition in Melendez-Diaz was not “qualified by a requirement of formality,” as the majority believes, majority at 475, but rather explained that there are three categories of testimonial statements: “ ‘ex parte in-court testimony or its functional equivalent,’ ” “ ‘extrajudicial statements . . . contained in formalized testimonial materials,’ ” and “ ‘statements that were made under circumstances which would lead an objective witness reasonably to believe that the statement would be available for use at a later trial.’ ” Melendez-Diaz, 557 U.S. at 310 (alteration in original) (quoting Crawford, 541 U.S. at 51-52).
This same logic applies to our analysis of the toxicology report, which the majority correctly concludes is testimonial, albeit for the wrong reasons. In Washington, the state forensic lab is required by statute to provide “all necessary toxicology procedures requested by all coroners [and] medical examiners.” RCW 68.50.107; RCW 43.43.670(l)(c). The toxicology report in this case was requested by the King County medical examiner and was no less mandated by statute than the autopsy report. Ex. 168, at 10.
The majority concludes that “[n]othing in the record states that the jury saw the reports prepared by [other lab analysts,]” majority at 490, based on Pineda’s testimony that “I did look at the electronic data from the results in the samples in this case. I did draw my own interpretation and my own conclusions from it.” 12 RP at 1507. The record, however, is not as clear as the majority believes. Pineda’s aforementioned testimony was in response to questions regarding how she arrived at her expert conclusions. Id. She did not specify whether the chart that the jury saw was based on raw data from the electropherogram or from the reports of Nasir and Vendetes. However, when explaining the data in the chart, Pineda explained that the data was based on whether it was reported. Pineda specifically stated, “If you look at the other column, starting with the right most column, the vaginal wash, you can see that at some row you have a number and other have an NR, which stands for no results or not reportable for any reason. That means that we didn’t detect peaks at all at that locus, or the peaks did not meet the minimum threshold for us to use it in our interpretation and to put it in a report.” Id. at 1540. Based on Pineda’s testimony, it seems the jury did see data from the DNA reports *512prepared by Nasir and Vendetes, though it was tunneled through Pineda’s testimony and chart.
The majority’s claim that the United States Supreme Court has not decided “the confrontation clause status of forensic reports” cannot be squared with its complaint that Melendez-Diaz is to blame for an increase in demand for laboratory analyst testimony. Compare majority at 479, with majority at 491-92.