Crawford v. Washington: The Next Ten Years

Jeffrey L. Fisher*

Imagine a world . . . in which the Supreme Court got it right the first time. That is, imagine that when the Supreme Court first incorporated the Confrontation Clause against the states, the Court did so by way of the testimonial approach.

It’s not that hard to envision. In Douglas v. Alabama—issued in 1965, on the same day the Court ruled that the Confrontation Clause applies to the states—the Court held that a nontestifying witness’s custodial confession could not be introduced against the defendant because, while “not technically testimony,” the confession was “the equivalent in the jury’s mind of testimony” from the nontestifying witness.[1] From that platform, all the Court would have needed to say in Dutton v. Evans[2] and Ohio v. Roberts[3] was that a statement made seemingly in confidence to a cellmate is not “testimonial” in nature, while statements at a preliminary hearing obviously are (although such statements still are admissible when the defendant had an adequate prior opportunity for cross-examination).

Of course, things did not turn out that way. Concerned about the unusual nature of Georgia’s version of the coconspirator hearsay exception at issue in Dutton, the Court tied itself up in knots. Then it announced in Roberts that the Confrontation Clause essentially tracked hearsay law.

For the next twenty-four years, states enjoyed a license to operate, for the most part, free and clear of the Confrontation Clause. No longer did prosecutors have to present their witnesses’ testimony in court, in the presence of the defendant, and subject to cross-examination. In other words, the clause ceased to be a procedural requirement (enforced by an “exclusionary rule”[4] as necessary to preserve the integrity of the right). It became simply a backup to local evidence law.

Still, I introduce this thought exercise in order to address an important issue: many judges (and justices) in the post-Crawford era find it very difficult to believe that so many seemingly entrenched state-law and prosecutorial practices can be illegitimate. Judges raised under the Roberts approach, in other words, find it a great struggle to change their mode of confrontation analysis from an evidentiary approach to one that is meant to protect a purely procedural right.

If the Supreme Court had gotten it right the first time, we would not have this problem. Law-enforcement agencies would not have developed prosecutorial practices—such as so-called “victimless,” or “evidence-based,” prosecutions—that depend on introducing statements obtained during ex parte interviews in lieu of live testimony. Law-enforcement agencies would not have created forensic evidence protocols designed to rely on affidavits instead of live testimony. States likely would have allocated financial resources toward programs aimed at producing lay and expert witnesses in court, instead of assuming it was unnecessary to undertake these burdens. And, dare I say it, law faculties would not have housed the right to confrontation in evidence classes and casebooks. Instead, professors likely would have opted to teach it in criminal procedure alongside the Fifth Amendment right against self-incrimination, the other constitutional right governing “witnesses” in criminal cases.

With these real-world developments and missed opportunities in mind, I will try in the pages that follow to achieve two things. First, I will explain why, despite some continuing resistance in the judiciary and the academy, I firmly believe that Crawford is fundamentally sound. Second, I will map out a handful of things the Supreme Court could do over the next decade to stabilize the Crawford doctrine and to make it more easily digested and applied. My hope and belief is that by the time we celebrate Crawford’s twentieth anniversary, most lawyers and judges will thankfully engage the doctrine without carrying baggage from the Roberts era.By that time, remaining controversies should involve the edges of the doctrine, not the fundamentals.


The testimonial approach starts from the premise that the Confrontation Clause is not a rule of evidence but rather one of criminal procedure. As the Court explained in Crawford:

Admitting statements deemed reliable by a judge is fundamentally at odds with the right of confrontation. To be sure, the Clause’s ultimate goal is to ensure reliability of evidence, but it is a procedural rather than a substantive guarantee. It commands, not that evidence be reliable, but that reliability be assessed in a particular manner: by testing in the crucible of cross-examination. The Clause thus reflects a judgment, not only about the desirability of reliable evidence (a point on which there could be little dissent), but about how reliability can best be determined.

Dispensing with confrontation because testimony is obviously reliable is akin to dispensing with jury trial because a defendant is obviously guilty. This is not what the Sixth Amendment prescribes.[5]

With this insight, the Court, quite properly, decoupled the Confrontation Clause from hearsay law. The Confrontation Clause does not turn on “amorphous notions of ‘reliability,’“ so its exclusionary rule cannot depend on “the vagaries of the rules of evidence.”[6] In other words, the constitutional considerations requiring testimony to be subject to cross-examination in criminal cases “do[] not evaporate when testimony happens to fall within some broad, modern hearsay exception, even if that exception might be justifiable in other circumstances.”[7]

All that remains from these basic principles is to determine which statements fall within this exclusionary rule—that is, which statements are “testimonial.” The Court seems to have settled on the following test: “To rank as ‘testimonial,’ a statement must have a ‘primary purpose’ of ‘establish[ing] or prov[ing] past events potentially relevant to a later criminal prosecution.’“[8] By last count, eight justices endorse this general test.[9] While some justices contend that a different test ought to apply to forensic analysts, there is no serious dispute that this “primary-purpose” test governs all statements made by eyewitnesses and other nonexperts.

I don’t think this formulation is perfect, but it will suffice. The critical thing it does is focus a court’s attention during litigation on the Confrontation Clause’s core concern: whether the speaker was doing the functional equivalent of offering testimony. Or, put another way, it focuses attention on whether the jury would perceive the statement as a substitute for live testimony. If so, the statement is testimonial, regardless of its perceived reliability.

To be sure, the Court noted in Bryant that, in determining whether a statement was procured “with a primary purpose of creating an out-of-court substitute for trial testimony[,] . . . standard rules of hearsay, designed to identify some statements as reliable, will be relevant.”[10] But Justice Sotomayor, the author of Bryant, quickly clarified this passage. She rightly explained that, when a statement satisfies a hearsay exception, it is likely to be nontestimonial because many hearsay exceptions “rest on the belief that certain statements are, by their nature, made for a purpose other than use in a prosecution.”[11] That is, some of the same things that hearsay law takes to indicate reliability can also indicate that a statement is nontestimonial. But there is no causal connection between the two—merely an overlap. The fact that a statement is reliable does not make it nontestimonial, for “[t]he rules of evidence, not the Confrontation Clause, are designed primarily to police reliability.”[12]

One other overarching principle is important to bear in mind. Unlike exclusionary rules that flow from other provisions of the Bill of Rights, the restriction the Confrontation Clause imposes on introducing testimonial statements is not designed to identify or deter police misconduct. There is nothing unconstitutional or even improper about a police officer’s asking specific questions to determine what happened or executing a witness affidavit. To the contrary, we want the police to question witnesses and to conduct probing inquiries. As the Court has put it:

Police investigations . . . are . . . in no way impugned by our characterization of their fruits as testimonial. Investigations of past crimes prevent future harms and lead to necessary arrests. While prosecutors may hope that inculpatory “nontestimonial” evidence is gathered, this is essentially beyond police control. . . . The Confrontation Clause in no way governs police conduct, because it is the use of, not the investigatory collection of, ex parte testimonial statements which offends that provision.[13]

Accordingly, courts that ask—in the face of an argument that a statement is testimonial—how the police should have acted differently are asking the wrong question. Neither the police nor any other agent of law enforcement has a legitimate interest in trying to obtain statements from witnesses that are immune from adversarial testing. Instead, investigators are simply supposed to try to find out what happened. Then the prosecution is supposed to prove its case by putting live witnesses on the stand. When the prosecution is unwilling or unable (through no fault of the defendant) to do so, the Confrontation Clause’s exclusionary rule comes into play.


Let me now outline a few things the Court should do over the next decade to clarify and solidify Crawford’s exclusionary rule.

First, the Court should decide one or two cases involving statements made to investigators other than police officers or their immediate agents (such as 911 operators). The Court has held that statements to police or their agents in the aftermath of potentially criminal events are testimonial—at least once any ongoing emergency arising from the events has been quelled.[14] But the Court has not yet addressed whether statements made under similar circumstances to other kinds of investigators or victim-support organizations are testimonial.

Most states, for example, employ personnel such as child protection services workers and other social workers to investigate suspected instances of child abuse. Most states or local law-enforcement agencies also direct (or at least encourage) “sexual-assault nurse examiners” or other personnel with medical training to conduct interviews of suspected crime victims, when possible, for forensic and investigatory purposes.[15] Finally, many localities have private victims’ services organizations that interview and help identify victims of abuse.

Lower courts, by and large, have held that statements made in these kinds of interviews are testimonial. But there are outliers in each of the three subcategories, particularly in situations in which the police are not present or otherwise directly involved.[16]

The Supreme Court would do well to make clear that these kinds of statements are testimonial. All are given under investigatory circumstances. All are passed on directly to law enforcement insofar as they are useful to potential criminal prosecutions. And all, in a jury’s eyes, are a nearly perfect substitute for in-court testimony. The statements recount under structured questioning “how potentially criminal past events began and progressed.”[17]

Issuing a decision holding that such statements are testimonial would erase any doubt about the Court’s general commitment to Crawford. It would also curtail the last remaining practice that law enforcement agencies developed to take advantage of the Roberts regime – a practice that threatens to allow a system in which victims and other witnesses can give their testimony without having to appear in court.

The other area of current confusion, of course, involves forensic reports. Even though the Court has held that reports identifying a substance as an illegal drug or determining blood-alcohol content are testimonial, lower courts are divided over whether the same is true with respect to autopsy reports concluding that a homicide occurred.[18] This should not be a difficult issue, and the Court should say so by explaining that autopsy reports are testimonial. Just like drug and alcohol reports, autopsies indicating that homicide has caused death are designed to aid investigations into past criminal events. Autopsy reports may not target or accuse a particular individual of committing the crime, but neither do numerous other sorts of plainly testimonial statements.

That leaves the final thing the Court should address in the near term—the rules governing the introduction of testimonial forensic reports through expert witnesses. In Williams v. Illinois, five justices concluded that, when one expert testifies on the basis of another expert’s report, that report (or any statement from it that the testifying expert transmits orally) is introduced for the truth of the matter asserted—and thus is subject to the ordinary rules governing the admissibility of testimonial statements.[19] This reasoning is manifestly correct. If the underlying report is true, it bolsters the expert’s opinion; if it is not, the expert’s opinion is less likely to be accurate. The Court should expressly endorse this principle in a majority opinion so as to remove any doubt that it constitutes the law.

Furthermore, the Court should make plain that the Confrontation Clause prohibits a testifying expert witness from transmitting the substance of a nontestifying witness’s testimonial statements, even if the expert also offers a so-called independent opinion concerning the situation. Again, most courts have already recognized as much. But not all have.[20] Just as with the “not-for-truth” argument that the state pressed in Williams, this independent-opinion argument would allow a bald circumvention of Crawford. It should therefore be walled off.


It is worth remembering that the project of developing rules of confrontation to apply against the states is roughly forty years behind similar efforts for other major criminal procedure rights. And nearly fifty years after the Supreme Court first coined various criminal procedure doctrines to implement other such rights—tests like the “reasonable expectation of privacy” and “reasonable person would feel free to leave” formulations, which are used to determine when “searches” and “seizures” take place—courts still struggle at the margins to apply those tests. This is because doctrines designed to translate ancient (or at least colonial-era) protections to modern circumstances cannot eliminate all hard cases. Rather, the most we can hope for is that such doctrines will allow courts to get the easy cases right and force courts to ask the right questions when considering the hard cases.

Crawford is well on its way to accomplishing that objective. (No one, for example, argues anymore about whether nontestifying witnesses’ custodial confessions or grand-jury testimonies are admissible.) But we’re not quite there yet. Hopefully in another ten years we will be.

*        Professor of Law, Stanford Law School.

[1].     380 U.S. 415, 419 (1965).

[2].     400 U.S. 74 (1970).

[3].     448 U.S. 56 (1980).

[4].     Davis v. Washington, 547 U.S. 813, 826 (2006).

[5].     541 U.S. 36, 61–62 (2004). See also Melendez-Diazv. Massachusetts,567 U.S. 305, 319 n.6 (2009) (emphasizing that the Confrontation Clause would bar the introduction of forensic lab reports as a substitute for live testimony even “if all analysts always possessed the scientific acumen of Mme. Curie and the veracity of Mother Theresa”).

[6].     541 U.S. at 61; see also United States v. Cromer,389 F.3d 662, 679 (6th Cir. 2004) (“If there is one theme that emerges from Crawford, it is that the Confrontation Clause confers a powerful and fundamental right that is no longer subsumed by the evidentiary rules governing the admission of hearsay statements.”).

[7].     Crawford,541 U.S. at 56 n.7. In Melendez-Diaz, the Court elaborated on this principle while responding to the Commonwealth’s argument that forensic laboratory reports were admissible as business or official records: “Business and public records are generally admissible [in criminal cases] absent confrontation not because they qualify under an exception to the hearsay rules, but because—having been created for the administration of an entity’s affairs and not for the purpose of establishing or proving some fact at trial—they are not testimonial. Whether or not they qualify as business or official records, the analysts’ statements here—prepared specifically for use at petitioner’s trial—were testimony against petitioner, and the analysts were subject to confrontation under the Sixth Amendment.” 567 U.S. at 324.

[8].     Bullcoming v. New Mexico, 131 S. Ct. 2705, 2714 n.6 (2011) (plurality opinion) (quoting Davis v. Washington, 547 U.S. 813, 822 (2006)); accord Michigan v. Bryant, 131 S. Ct. 1143, 1154 (2011).

[9].     See Davis v. Washington, 547 U.S. 813, 822 (2006); see also Bryant, 131 S. Ct. at 1154 (new justices sign on to the opinion applying the Davis primary-purpose formulation).

[10].     Bryant, 131 S. Ct. at 1155; see also Bullcoming, 131 S. Ct. at 2720 (Sotomayor, J., concurring in part).

[11].     131 S. Ct. at 1157 n.9.

[12].     Bullcoming, 131 S. Ct. at 2720 n.1 (Sotomayor, J., concurring in part).

[13].     Davis, 547 U.S. at 832 n.6; see also id. at 830 (noting that “the primary, if not indeed the sole, purpose of the interrogation was to investigate a possible crime—which is, of course, precisely what the officers should have done”).

[14].     Bryant, 131 S. Ct. at 1147; Davis, 547 U.S. at 814.

[15].     As the Nevada Supreme Court put it, in reference to sexual-assault nurse examiners (“SANEs”), such “nurses . . . are trained to conduct sexual assault examinations. A particular duty of a SANE nurse is to gather evidence for possible criminal prosecution in cases of alleged sexual assault. SANE nurses do not provide medical treatment. They only examine the individual to get vital signs and a history from the victim.” Medinav. Nevada,143 P.3d 471, 473 (Nev. 2006). See generally United States v. Gardinier, 65 M.J. 60 (C.A.A.F. 2007).

[16].     Courts have held that statements to state social workers or child protection services workers during risk-assessment interviews are testimonial, regardless of whether the law-enforcement personnel are involved in the interview or the interview is conducted at the behest of law enforcement. See, e.g., State v. Hopkins,154 P.3d 250 (Wash. Ct. App. 2007). But see State v. Arnold, 933 N.E.2d 775 (Ohio 2010). If police are even indirectly involved with an interrogation conducted by a state social worker or child protection services worker, courts will likely find the statement testimonial. See, e.g., In re S.P.,215 P.3d 847 (Or. 2009). But a significant minority of courts treats such statements as nontestimonial when the police are not yet directly involved. See, e.g., Seely v. State, 282 S.W.3d 778, 789–90 (Ark. 2008).

When private personnel interview child victims in coordination with law enforcement, courts have held that the resulting statements are testimonial. See, e.g., State v. Blue,717 N.W.2d 558, 564 (N.D. 2006). When the police are not directly involved, however, courts are divided over whether statements to such private personnel are testimonial. Compare D.G. v. State, 76 So. 3d 852 (Ala. Crim. App. 2011), with Bishop v. State,982 So. 2d 371 (Miss. 2008).

Every court but one has held that statements to sexual-assault nurse examiners and similar interviewers are testimonial even if the police are not yet involved. See, e.g., Medina v. State,143 P.3d 471 (Nev. 2006). But see State v. Scacchetti,711 N.W.2d 508 (Minn. 2006).

[17].     Davis, 547 U.S. at 830.

[18].     Compare United States v. Ignasiak, 667 F.3d 1217 (11th Cir. 2012) (testimonial), United States v. Moore, 651 F.3d 30 (D.C. Cir. 2011), Malaska v. State, 88 A.3d 805 (Md. Ct. Spec. App. 2014), Commonwealth v. Carr, 986 N.E.2d 380 (Mass. 2013), State v. Navarette, 294 P.3d 435 (N.M. 2013),Miller v. State, 313 P.3d 934 (Okla. Crim. App. 2013), Lee v. State, 418 S.W.3d 892 (Tex. Ct. App. 2013), State v. Lui, 315 P.3d 493 (Wash. 2014), and State v. Kennedy, 735 S.E.2d 905 (W. Va. 2012), with State v. Medina, 306 P.3d 48 (Ariz. 2013) (nontestimonial), People v. Dungo, 286 P.3d 442 (Cal. 2012) (objective statements in report not testimonial), People v. Leach, 980 N.E.2d 570 (Ill. 2012), and State v. Maxwell, 9 N.E.3d 930 (Ohio 2014).

[19].     132 S. Ct. 2221, 2257 (2012) (Thomas, J., concurring); id. at 2268 (Kagan, J., dissenting).

[20].     Compare United States v. Soto, 720 F.3d 51 (1st Cir. 2013), Ignasiak, 667 F.3d 1217, State v. McLeod, 66 A.3d 1221 (N.H. 2013), and Kennedy, 735 S.E.2d 905, with State v. Brewington, 743 S.E.2d 626 (N.C. 2013) (ignoring in-court witness’s disclosure of testimonial statements because the witness offered “independent opinion”), and Lui, 315 P.3d 493 (testifying expert may transmit any noninculpatory statement from nontestifying analyst).

Download as PDF.