Criminal Justice Section  


Criminal Justice Magazine
Summer 2005
Volume 20 Issue 2

What Remains of Reliability:
Hearsay and Freestanding Due Process After Crawford v. Washington

By Andrew Taslitz

Andrew E. Taslitz, a professor at the Howard University School of Law, is the chair of the editorial board of Criminal Justice, the editor of this symposium, an immediate past member of the Criminal Justice Section’s Council, and chair of the Eyewitness Identification Subcommittee of the ABA’s Ad Hoc Committee on Innocence and the Integrity of the Criminal Justice System.

This symposium is a first take on the impact of Crawford v. Washington, 124 S. Ct. 1354 (2004), on hearsay practice in the courts during the approximately one year since that path-breaking Confrontation Clause case was decided. The Confrontation Clause of the Sixth Amendment to the U.S. Constitution entitles a criminal defendant to be “confronted with the witnesses against him.” The Supreme Court has long held that the clause’s scope includes protecting a criminal defendant against the admissibility of certain hearsay at trial. Specifically, before Crawford, the Court applied the test of Ohio v. Roberts, 448 U.S. 56 (1980), under which hearsay may be admitted against a defendant at trial consistently with the Confrontation Clause only if the hearsay statement bears “adequate indicia of reliability.” (Id. at 66.)
“Firmly rooted hearsay,” that is, hearsay admissible under exceptions with a long pedigree or hearsay shown to bear “particularized guarantees of trustworthiness,” meet the Roberts test. (See id. at 66.) For some categories of hearsay, the unavailability of the declarant—the person who made the hearsay statement—also had to be shown, though the scope and continued viability of this requirement was subject to dispute. (See STEVEN I. FRIEDLAND ET AL., EVIDENCE LAW AND PRACTICE 576-77 (1st ed. 2000).) For “core” confrontation rights, however, Crawford substituted an entirely different test: “testimonial” hearsay statements are admissible against a criminal defendant only if the declarant is unavailable at trial and the defendant had a prior opportunity for cross-examination. (See Robert P. Mosteller, Crawford v. Washington: Encouraging and Ensuring the Confrontation of Witnesses, 39 U. RICH. L. REV. 511, 516 (2005) (summarizing holding).)
Although there is much language in the opinion that could be read as suggesting that Crawford forever buried the Roberts test, the Crawford Court did not in fact decide whether Roberts might still live in the world beyond the testimonial core. (Mosteller, supra, at 616-20.) The scope of that core—and thus the meaning of “testimonial”—was not decisively demarcated by Crawford, and many of the essays in this symposium seek to clarify that and other ambiguities of the new way. (See Richard D. Friedman, Adjusting to Crawford: High Court Decision Restores Confrontation Clause Protection, 19 CRIM. JUST. 5, 8-12 (Summer 2004).) Like a vampire, however, Roberts has repeatedly risen from the dead in a significant number of lower court decisions. (See, e.g., Rollins v. Maryland, 2005 WL 181650 (2005); Owens v. Frank, 2005 WL 22702 (2005); State v. Rivera, 2003 WL 23341462 (Conn. 2004); People v. Coker, 2004 WL 62855 (Mich. Ct. App. 2004).) This article’s focus, therefore, is on this two-part question: What remains, and what should remain, of reliability after Crawford?
Note that I have not asked, What remains of Roberts after Crawford? Rather, I asked, What remains of reliability? Roberts is not the only potential source of a reliability analysis. To the contrary, Roberts has proven to be a toothless rule, and it will likely remain so. (See Friedman, supra, at 13.) Indeed, almost all the cases finding that Roberts still lives, albeit in the Confrontation Clause’s periphery, find the hearsay reliable, thus admissible, so little is gained in parsing out lower court opinions so holding with any great care. Instead, I argue here for a reinvigorated due process analysis, drawing on recent lessons from the innocence movement. That movement has not only led to the exoneration of numerous innocent persons wrongly convicted. It has also uncovered the causes of error in ways that suggest too many undiscovered innocents may be languishing in prison. (See generally symposium on Wrongful Convictions, 18 CRIM. JUST. (Spring 2003).)
At least four broad lessons from the plight of the wrongly convicted have relevance to the problem of hearsay in criminal cases: (1) social science is helpful in identifying and correcting the causes of error; (2) better informed judges and juries are less likely to make mistakes; (3) relatively clear, categorical rules, where feasible, are far preferable to flexible balancing tests; and (4) pretrial procedures are as important as trial ones in promoting accurate verdicts. Due process is meant more than any other constitutional doctrine to bear the load of encouraging correct results, and these four principles can give renewed force to a due process jurisprudence that has sometimes become too limp. (See DONALD DRIPPS, ABOUT GUILT AND INNOCENCE xvii (2003).) But the load is not meant to be borne by the Confrontation Clause, even though some discussion of reliability is unavoidable under that clause. Confrontation jurisprudence is better understood as serving other primary goals, such as preventing governmental abuses in the creation of evidence, promoting the legitimacy of verdicts, and allowing for the catharsis of face-to-face confrontation.
Resolution of the question of what remains of reliability after Crawford is important because so much hearsay will indeed be beyond the core. In the Crawford opinion itself, the Court noted that “[a]n off-hand, overheard remark might be unreliable evidence and thus a good candidate for exclusion under hearsay rules, but it bears little resemblance to the civil-law abuses the Confrontation Clause targeted.” (124 S. Ct. at 1364.) The Court also distinguished “formal statements to government officers” (core) from “a casual remark to an acquaintance” (noncore). (Id.) More specifically, the Court further noted that business records and statements made in furtherance of a conspiracy are not testimonial. (Id. at 1367.) Additionally, without deciding the question, the Court also suggested that even dying declarations that fit within the contours of a testimonial statement may nevertheless, as a sui generis exception for historical reasons of the Framers’ intent, be beyond the core. (See id. at 1367 n.6.) Moreover, lower courts have found a variety of statements within the Confrontation Clause periphery that were admitted under hearsay exceptions for excited utterances, then-existing statement of mind, medical diagnosis and treatment, as well as under the business records and coconspirators exceptions identified by the Court. (See Jeffrey Fisher, Crawford v. Washington: Reframing the Right to Confrontation, at (last updated Feb. 9, 2005; last visited Mar. 15, 2005) (summarizing cases).) Furthermore, some prosecutors have tried to argue, relying on a select portion of the guidance given by the Crawford opinion as to the meaning of the term “testimonial,” that much hearsay by domestic violence victims may not be testimonial, or, if testimonial, that the defendant has forfeited his right to rely on the Confrontation Clause. (See Adam M. Krischer, “Though Justice May Be Blind, It Is Not Stupid”: Applying Common Sense to Crawford v. Washington in Domestic Violence Cases, 1 THE VOICE 1, 1-10 (Nov. 2004).)
These examples alone demonstrate the breadth of hearsay testimony that will be unregulated by the Crawford rule. Although a default position might be that the reliability of such statements should be left to the discretion of local legislatures, courts, and juries, the innocence cases highlight the dangers of mistaken conviction that arise from inadequate trial and pretrial procedures, even when all actors in the system act in good faith, work hard, and take their obligations seriously. Some minimal assurances of trustworthiness beyond the good faith and common sense of the criminal justice system’s participants are required. Forgetting that lesson also means letting too many guilty offenders go free, perhaps to prey on further victims, for each conviction of the innocent means that the guilty escape justice.
The starting point for all this analysis must, of course, be to explore the Crawford Court’s own vision of the role of reliability under the Confrontation Clause and the lessons that we can draw from the Court’s admittedly ambiguous and inconsistent comments on this point. After that analysis, this article will offer sufficient historical detail—perhaps a bit more than is usual in this publication—to enable a practitioner, likely a prosecutor, to argue that the Crawford rule exhausts Confrontation Clause requirements. The last major portion of this article outlines the argument, this time likely to be used by the defense, that due process might nevertheless exclude unreliable hearsay whether or not the declarant is available.

Crawford and reliability
Crawford rested its analysis on an examination of the history of the Confrontation Clause. What is most interesting, however, about the Court’s reciting of this history is the purposes for which it did so. The Court sought to divine from this history the hallmarks of the paradigmatic cases that troubled English, colonial, and early state lawmakers and interpreters. But the Court offered a brief, cryptic, hit-and-run analysis of why lawmakers had these concerns, that is, of a guiding rationale to explain the love affair with confrontation. After examining confrontation’s history from Rome to English common-law practices to early state declarations of rights and state decisions, the Court concluded that “the principal evil at which the Confrontation Clause was directed was the civil-law mode of criminal procedure, and particularly its use of ex parte examinations as evidence against the accused,” for it was “these practices that the Crown deployed in notorious treason cases like Raleigh’s; that the Marian statutes invited; that English law’s assertion of a right to confrontation was meant to prohibit; and that the founding-era rhetoric decried.” (124 S. Ct. at 1363.)
That observation, combined with a dictionary-informed reading of the Confrontation Clause’s text (it applies to “witnesses,” who are those who “bear testimony”), led the Court to craft its ultimate testimonial evidence rule. (See id. at 1364.) Without knowing why ex parte affidavits were so feared, however, there is little to guide future courts in deciding the scope of testimonial evidence in the very different circumstances of the modern world. The Court did give one important hint:

We cannot agree with the CHIEF JUSTICE that the fact “[t]hat a statement might be testimonial does nothing to undermine the wisdom of one of these [hearsay] exceptions.” . . . Involvement of government officers in the production of testimony with an eye toward trial presents unique potential for prosecutorial abuse—a fact borne out time and again throughout a history with which the Framers were keenly familiar. This consideration does not evaporate when testimony happens to fall within some broad, modern hearsay exception, even if that exception might be justifiable in other circumstances.
(Id. at 1367 n.7.)

This statement, submerged in a footnote, correctly articulates—or so I will argue shortly—a critical function of the Confrontation Clause that is distinct from reliability enhancement: confrontation permits the uncovering and revealing of intentional or negligent governmental abuses in the creation of testimony, abuses that leave the individual at the mercy of a vastly more powerful state and that may expose the innocent to conviction at the hands of vindictive or incompetent police or prosecutors. Consistent with this argument, the Crawford Court also mocked one of the trial court’s arguments for the admissibility of the hearsay in question—a statement to police made by Crawford’s wife, Sylvia, in which she suggested that Crawford had not harmed his alleged victim in self-defense. The trial court had argued that Sylvia’s statement was reliable in part because she was questioned by a law enforcement officer who was “neutral to her and not someone who would be inclined to advance her interests and shade her version of the truth unfavorably toward the petitioner.” (Id. at 1373.) Said the Court, “The Framers would be astounded to learn that ex parte testimony could be admitted against a criminal defendant because it was elicited by ‘neutral’ government officers.” (Id.)
The Court also noted that even if the officers’ motives were good, that said nothing about how Sylvia perceived the situation, something that only cross-examination could reveal. (See id.) This point might go to the reliability of Sylvia’s testimony, but it also can be read as illustrating how government carelessness or zeal can distort evidence and tip the scales further against an accused even when the police act in good faith. If Sylvia had come to believe, as a result of police investigatory practices, that it was in her own interest to bend the truth to curry favor with the authorities, she may have done just that. Indeed, one of the reasons that the Court gave for discarding a purely reliability-focused approach to the Confrontation Clause was that the distrust of the state embodied in the clause extends even to the judiciary, not merely the constabulary:
We have no doubt that the Courts below were acting in utmost good faith when they found reliability. The Framers, however, would not have been content to indulge this assumption. They knew that judges, like other government officers, could not always be trusted to safeguard the rights of the people; the likes of the dread Lord Jeffreys were not yet too distant a memory. They were loath to leave too much discretion in judicial hands [citing respective Sixth Amendment rights to civil and criminal jury trials]. . . . By replacing categorical constitutional guarantees with open-ended balancing tests, we do violence to their design. Vague standards are manipulable, and, while that might be a small concern in run-of-the-mill assault prosecutions like this one, the Framers had an eye toward politically-charged cases like Raleigh’s—great state trials where the impartiality of even those at the highest levels of the judiciary might not be so clear. It is difficult to imagine Roberts providing any meaningful protection in these circumstances.
(Id. at 1373-74.)

The Court favorably cited the work of two well-known law professors, Akhil Reed Amar and Richard D. Friedman, in its opinion. (See id. at 1370.) Amar stressed in his work the close link among confrontation, jury trial, and public trial rights as a way of ensuring that the whole truth be exposed to the public by avoiding “private and secret examinations.” (See AKHIL REED AMAR, THE CONSTITUTION AND CRIMINAL PROCEDURE 125-31 (1997).) Although Amar emphasized truth-seeking more than did Friedman, Friedman was the principal author of an amicus brief whose reasoning the Court carefully tracked. (Compare Motion for Leave to File an Amicus Curiae Brief of Law Professors Sherman J. Clark et al. in Support of Petitioner Crawford v. Washington, No. 02-9410 (authored primarily by Friedman), with Akhil Reed Amar, Confrontation Clause: First Principles: A Reply to Professor Friedman, 86 GEO. L.J. 1045 (1998) (articulating a variant from Friedman’s position); AMAR, CRIMINAL PROCEDURE, supra, at 125-27 (stressing truth).) In that brief, Friedman likewise stressed the openness of confrontation procedure as discouraging and revealing testimony that is the “product of torture or of milder forms of intimidation” while rendering the making of a “false accusation much more difficult than it would be otherwise.” (Amicus Brief, supra, at 14.) Friedman’s brief also emphasized the clause’s symbolic purposes, quoting the Court’s earlier opinion in Coy v. Iowa, 487 U.S. 1012 (1988), reiterated in Maryland v. Craig, 497 U.S. 836 (1990), that “there is something deep in human nature that regards face-to-face confrontation between accused and accuser as ‘essential to a fair trial in a criminal prosecution’ ” (Amicus Brief, supra, at 16 (quoting Craig, 497 U.S. at 846))—statements neither inconsistent with nor disavowed by the Crawford opinion.
That opinion may implicitly have embraced the clause’s symbolic function as well, noting that “[d]ispensing 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.” (124 S. Ct. at 1371.) The Court’s point here seems to be that there are good political and symbolic reasons to insist on jury trials even when evidence of an accused’s guilt is overwhelming, namely, that we should trust the people—as embodied in the jury—to make these judgments more than we do the state, and that doing so promotes the perceived fairness and legitimacy of the justice system. The same may be said of the Confrontation Clause.
But did the Court retain any role for reliability under the clause? The Court attacked reliability analysis as an “amorphous notion” whose vagaries the Framers would not have left to the Sixth Amendment’s protections. (Id.) Nevertheless, the Court described the clause’s “ultimate goal” as ensuring the 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.” (Id. at 1370; see also Amicus Brief, supra, at 14-20 (noting confrontation promotes reliability but that is not and cannot be seen as the Confrontation Clause’s guiding purpose).) The Court thus belittled the idea of letting hearsay evidence “untested by the adversary process” reach the jury upon a “mere judicial determination of reliability.” (124 S. Ct. at 1370.) The constitutionally prescribed method of assessing reliability is by live cross-examination before a jury rather than by judicial command.
Moreover, concluded the Court, the Roberts reliability test had proven unworkable in practice before the lower courts, being so unpredictable that it failed to provide any “meaningful protection” against core clause violations. (Id. at 1371.) Thus the Court noted that the many multifactor balancing tests used by lower courts enabled some courts to find that the very factors favoring reliability are precisely those that other courts have declared undermine reliability. (See id. at 1371-72.) Furthermore, some courts found reliability based on the factors that make the statements testimonial, for example, by finding that statements made to police were more clearly against penal interest, therefore more reliable, when the giving of inculpatory statements in a testimonial setting is “not an antidote to the confrontation problem, but rather the trigger that makes the Clause’s demands most urgent.” (Id. at 1372.)
Additionally, noted the Court, judicial determinations of reliability may be based upon unwarranted assumptions that only cross-examination might uncover. Sylvia Crawford herself was told that whether she would be released depended upon how the investigation continued, and she responded to leading questions posed by the police, thus raising dangers of distorting the truth. Yet the trial court assumed that Sylvia was reliable because she was “an eyewitness with direct knowledge of the events,” even though she had also told the police that her eyes were closed during many of the relevant events, that she really did not watch part of them, and that she was “in shock,” matters crying out for further exploration. (Id. at 1372-73.) The Court noted the unpredictability of the Roberts test again, this time as evidenced by the conflicting conclusions on reliability reached by the trial and appellate courts in the Crawford case itself. Roberts was simply too malleable a standard for the Court’s taste. (See id. at 1371-72.)
Yet, after all this gnashing of teeth about amorphous reliability tests, the Court also added in dictum that, for nontestimonial hearsay, it is “consistent with the Framers’ design to afford the States flexibility in their development of hearsay law—as does Roberts” and as would an approach that “exempted such statements from Confrontation Clause scrutiny altogether,” leaving a deep ambiguity concerning whether there is a Confrontation Clause periphery to which Roberts’ reliability analysis applies or whether reliability judgments made by the judiciary are wholly beyond the Confrontation Clause, though not necessarily (as the Court did not even tangentially address the question) entirely outside constitutional regulation, such as via due process. (See id. at 1374.)
The Court’s reasoning concerning the role of reliability analysis can be made most coherent by reading Crawford thus: the Confrontation Clause is designed primarily to prevent intentional or accidental creation of flawed evidence by governmental investigatory abuses, abuses that can be both discouraged and exposed by public cross-examination. These qualifications are important, for under this reading there is no right to cross-examination of evidence neither created by the government nor reasonably expected by a hearsay declarant to be used by the government, an expectation that may itself taint the evidence. This latter point reflects the reality that once the government enters the picture, evidence-distortion may follow in its wake, no matter how honorable are police and prosecutors’ intentions. Reliability does not therefore matter in itself. Nor indeed does the manner of testing reliability by cross-examination matter in itself, for if it did, an entitlement to such testing would exist even for nontestimonial hearsay. What matters, therefore, are the risks of oppression of the individual and of unreliable evidence stemming from the government’s actual or reasonably perceived involvement in the creation of evidence. Thoroughly unreliable evidence entirely causally unconnected with governmental action is thus not the Confrontation Clause’s concern. Roberts is indeed dead. (Cf. Dale Nance, Rethinking Confrontation after Crawford, 2 INT’L COMMENT. ON EVIDENCE 1, 13 (2004) (“Crawford’s seeming rejection of the Roberts reliability framework should signal the rejection of a jury distrust theory of Confrontation and its replacement with a theory grounded in distrust of the government especially the executive and legislative branches.”) (emphasis added).)
If I am right, does that and should that mean that a criminal defendant has no constitutional recourse whatsoever when faced with powerful, damning hearsay that also raises a grave risk of error? My answer is no. He does have recourse under the fundamental fairness guarantee of the Due Process Clause. Before exploring why that is so, it is first important to explain why the Confrontation Clause’s job is best understood as limited to preventing and exposing governmental evidentiary abuses, leaving the real work of reliability promotion as a worthwhile goal in itself to due process.

Controlling government misconduct
It can no longer fairly be disputed that the primary, although not necessarily sole, purpose of the Confrontation Clause is preventing governmental misconduct in the creation of evidence. (See, e.g., Margaret Berger, The Deconstitutionalization of the Confrontation Clause: A Proposal for a Prosecutorial Restraint Model, 76 MINN. L. REV. 557 (1992) (first articulating the history supporting this conclusion); Joshua C. Dickinson, The Confrontation Clause and the Hearsay Rule: The Current State of a Failed Marriage in Need of a Quick Divorce, 33 CREIGHTON L. REV. 763, 803-07 (2000) (arguing that there are important, but subtle, differences between Berger’s prosecutional restraint model of history and Friedman’s alternative); Richard D. Friedman & Bridget McCormack, Dial-In Testimony, 150 U. PA. L. REV. 1171, 1246-52 (2002) (noting similarities of Friedman’s historical approach to Berger’s but rejecting her insistence on governmental involvement as permitting avoidance of the mandates of the Confrontation Clause).) A few commentators deny that controlling governmental involvement in evidence creation is the direct purpose of the clause, preferring to describe it as creating an affirmative right in the accused to ensure the sound workings of the adversarial system. (See, e.g., Randolph K. Jonakait, Restoring the Confrontation Clause to the Sixth Amendment, 35 U.C.L.A. L. REV. 557 (1988) (making just this point); Penny J. White, Rescuing the Confrontation Clause, 54 S.C. L. REV. 537, 623 (2003) (confrontation is a right essential to the functioning of the presumption of innocence and the rights to counsel and to a jury).) Yet these authors generally recognize that one of the primary functions of our jury-centered version of the adversary system is to serve as a check on abuses of governmental power. (See RANDOLPH JONAKAIT, THE AMERICAN JURY SYSTEM 18-40 (2003).) Although they recognize truth-seeking as another goal of the adversary system, they do not see advancing “the accuracy of the truth-determining process” as the Confrontation Clause’s goal. (See JONAKAIT, AMERICAN JURY SYSTEM, supra, at 41-64 (adversary system and accuracy in truth-finding); Randolph N. Jonakait, A Response to Professor Berger: The Right to Confrontation: Not a Mere Restraint on Government, 76 MINN. L. REV. 615, 615 (1992) (rejecting truth-determination as the clause’s overriding goal); cf. John G. Douglass, Beyond Admissibility: Real Confrontation, Virtual Cross-Examination, and the Right to Confront Hearsay, 67 GEO. WASH. L. REV. 191, 232, 272 n.21 (1999) (due process, not confrontation, is the proper doctrinal home for reliability concerns).)

Reading the Confrontation Cause holistically as part of the Bill of Rights
The commentators rely on two primary rationales relevant here. The first is that the Confrontation Clause must be viewed as part of a greater whole, the Bill of Rights. (See Berger, supra, at 557-62.) The structure of the Bill of Rights and of Supreme Court precedent thereunder reveals a vital concern with restraining governmental misconduct. Thus the Fourth Amendment prevents the police from searching, arresting, and seizing property, unless the police have a good enough reason—usually probable cause—to believe that a particular defendant has committed a crime and, generally, unless the police also have sought prior judicial approval for their action. Prior judicial approval is also meant to limit police discretion and place decision-making power in a more neutral arbiter. (See INGA L. PARSONS, FOURTH AMENDMENT: PRACTICE AND PROCEDURE 61-62 (2005).)
Similarly, under the Fifth Amendment privilege against self-incrimination, the police are barred from using physical or psychological coercion to obtain confessions. (See generally JOHN B. TAYLOR, RIGHT TO COUNSEL AND PRIVILEGE AGAINST SELF-INCRIMINATION (2004).) Indeed, so great is the Court’s fear of police abuses in obtaining confessions that the Court, in Miranda v. Arizona, 384 U.S. 436 (1966), created a Fifth Amendment right to the presence of counsel during questioning and required a series of warnings to ensure the defendant’s awareness of this and related rights. Additionally, under the Sixth Amendment right to the assistance of counsel, an indicted defendant’s uncounseled statements may not be induced by the government, thus demonstrating a concern with the government’s ability to shape evidence in secret, through inquisitions. (See Massiah v. United States, 377 U.S. 201 (1964).) Likewise, the Court has held that there is a Sixth Amendment right to counsel at a postindictment lineup because of the dangers that an improperly conducted lineup arranged by the government will result in a suggestive identification. (See United States v. Wade, 388 U.S. 218 (1967).)
The Sixth Amendment’s Confrontation Clause helps to control government misconduct in a particular way, through disclosure, specifically, by opening that conduct up to scrutiny in a public trial. This scrutiny, inherent in the opportunity for “effective” cross-examination, may, of course, serve to reveal weaknesses and inaccuracies in the evidence to the jury. (See generally Douglass, supra.) But disclosure serves a different purpose than promoting accurate decision making in the case before the court. Public disclosure serves as well to give government a disincentive to engage in future wrongful conduct. Berger has put this well, noting that such an interpretation “complement[s] the other rights that the [Sixth] Amendment grants—trial by jury, a public trial, specification of the charges, and right to counsel—by providing yet another mechanism for making crucial workings of the government visible and keeping the overwhelming prosecutorial powers of the government in check.” (See Berger, supra, at 562.)
Deterring wrongful government conduct is, of course, recognized as a major purpose behind the Court’s jurisprudence under much of the Bill of Rights. Most notably, the primary function of the exclusionary rule is to deter Fourth, Fifth, and Sixth Amendment violations by the government. (See ANDREW E. TASLITZ & MARGARET L. PARIS, CONSTITUTIONAL CRIMINAL PROCEDURE 9 (2d ed. 2004).) The Confrontation Clause serves as a deterrent as well, not solely by excluding evidence, but also by encouraging admission of certain forms of evidence and exposing them to attack at trial. (See also Berger, supra, at 561 (exposure of governmental investigative processes creates a disincentive for secret prosecutorial shaping of witness’s answers while preventing the state from hiding “behind the cloak of the hearsay rule” and allowing the “scrutinizing [of] a process in which jurors should play a role as political participants.”).)
Confrontation Clause history and restraining government. These scholars also argue that the history of the Confrontation Clause supports the holistic reading of its text as part of the Bill of Rights in reaching the same conclusion: The clause is best understood as being more about the collective political and systemic function of citizen jurors deterring government abuses in evidence creation than about truth and accuracy in an individual case. Although avoiding abusive evidence creation will promote accuracy, that is a benefit rather than the function or purpose of the clause. The English trials famously raising demands for witness confrontation and for rejection of civil law-like trials by affidavit arose before lawyers, and thus before cross-examination and its purported role in promoting truth, played a central role in criminal cases. (See Berger, supra, at 572.) The oath was considered the primary guarantor of reliability, yet still defendants demanded confrontation. (See id. at 572-73.) Gilbert’s The Law of Evidence, written at the turn of the eighteenth century, after the most infamous of the confrontation-demand trials but at a time when the hearsay rule and the need for cross-examination were in their infancy and were accorded little significance (see id. at 574), recognized that confrontation could promote reliability but critically linked that result to the importance of the method by which it came about. Thus, in disparaging proof by depositions, he said:

[T]he Credit of Depositions ceteris paribus [all else being equal] falls much below the Credibility of a present Examination viva voce, for the Examiners and Commissioners in such Cases do often dress up secret Examinations, and set up a quite different Air upon them from [what] they would seem, if the same Testimony had been plainly delivered under the strict and open Examination of the Judges at the Assizes.
(GEOFFREY GILBERT, THE LAW OF EVIDENCE 45 (Garland Publ. 1979) (1754).)

The historical trend in England was consistently away from inquisitorial methods. The Court of Star Chamber, relying entirely on inquisitorial methods without a jury, was abolished, while the role of the jury in the common-law courts was enhanced. (See LEONARD W. LEVY, ORIGINS OF THE FIFTH AMENDMENT 38, 281-82 (1968); THEODORE T. PLUCKNETT, A CONCISE HISTORY OF THE COMMON LAW 134 (5th ed. 1956); FRANCIS H. HELLER, THE SIXTH AMENDMENT 10 (1969).) From and after the 1730s, counsel played an ever-growing role in examining and impeaching witnesses. (See White, supra, at 548.)
By the Revolutionary period, the Framers were familiar with the battle against inquisitorial procedures, seen as part of a long struggle to enhance citizen power relative to that of the Crown. (See Berger, supra, at 578-83.) Part of the colonists’ objection to the Stamp Act of 1765 was that it enlarged vice-admiralty courts’ jurisdiction, courts in which no juries sat and witnesses were examined in chambers upon interrogatories. (See id. at 579.) Lawyers active in America but trained in England were familiar with the confrontation trials and the commentary on them. (See id. at 580.)
Blackstone’s Commentaries, which played a major role in the development of the American legal system, also reflected thorough awareness of confrontation rights and of their sister guarantees of counsel, compulsory process, and the right to a trial by jury in the vicinage. (See LAWRENCE W. FRIEDMAN, A HISTORY OF AMERICAN LAW 102 (2d ed. 1985); Berger, supra, at 581-82.) Blackstone himself, though mentioning confrontation’s value in truth-promotion, stressed the importance of doing so via a method preferable to “private and secret examination taken down in writing before an officer, or his clerk,” borrowing the “practice from the civil law . . . where a witness may frequently depose that in private which he will be ashamed to testify in a public and solemn trial.” (4 WILLIAM BLACKSTONE, COMMENTARIES ON THE LAW OF ENGLAND 344-45 (University of Chicago Press 1979) (1765).) He continued: “[T]here [in private] an artful or careless scribe may make a witness speak what he never meant, by dressing up his depositions in his own forms and language. . . .” (Id.) Blackstone emphasized confrontation’s link to the jury, which “he and the framers of the Bill of Rights viewed as the principal safeguards of a people’s liberties,” but he otherwise was not concerned with the law of evidence in a period during which it was undergoing significant change. (Berger, supra, at 584.)
Though there is little in the way of legislative debates about the adoption of state constitutional equivalents to the Confrontation Clause or about that clause’s adoption itself, George Mason was concerned enough to insert a similar clause in the Virginia Declaration of Rights, the paradigm for bills in seven other colonies. (Mural A. Larkin, The Right of Confrontation: What Next? 1 TEX. TECH. L. REV. 67, 75-76 (1969).) Looking beyond the record of debates, moreover, reveals a deep concern about confrontation rights by Anti-Federalists, those opponents to the proposed federal Constitution who insisted on, and eventually got, a Bill of Rights. (See Berger, supra, at 585-86.) The Anti-Federalists viewed confrontation as inseparably linked to the jury trial right, and the jury as perhaps the central safeguard against governmental overbearing. (See AKHIL REED AMAR, THE BILL OF RIGHTS: CREATION AND RECONSTRUCTION 81-118 (1998).) An essayist, known as the Impartial Examiner, so explained in the Virginia Independent Chronicle, protesting the absence of adequate civil liberties guarantees in the proposed new federal constitution:

If you pass this new constitution, you will have a naked plan of government unlimited in its jurisdiction, which not only expunges your bill of rights by rendering ineffectual, all the state governments; but as proposed without any kind of stipulation for any of those natural rights, the security whereof ought to be the end of all governments. Such a stipulation is so necessary, that it is an absurdity to suppose any civil liberty can exist without it. . . . For instance, if Congress should pass a law that persons charged with capital crimes shall not have a right to demand the cause or nature of the accusation, shall not be confronted with the accusers or witnesses, or call for evidence in their own favor; and a question should arise respecting their authority therein, . . . can it be said that they have exceeded the limits of their jurisdiction, when that has no limits; when no provision has been made for such a right? . . . When no responsibility on the part of Congress has been required by the constitution?
(5 THE COMPLETE ANTI-FEDERALIST 185 (Herbert Storing ed., 1981) (emphasis added); see also White, supra, at 548-52 (tracing differences between the development of American and British criminal procedural rights, including the quicker development in America of the right to counsel but noting Americans’ perhaps even greater fear of government than was displayed by the English).)

Greater accuracy in fact-finding was one desired and expected outcome of a vigorous right to confrontation but not its goal. (See Amicus Brief, supra, at 11 (“This [historical] account has not mentioned reliability. Though one of the advantages perceived for confrontation was its contribution to truth-determination, the confrontation right was not considered contingent, inapplicable upon a judicial determination that the particular testimony was reliable.”).) The central goal was restraining government overreaching in the secret creation of evidence.
The Confrontation Clause’s scope. This conclusion does not require, however, limiting the Confrontation Clause’s scope to situations in which the government is directly and purposefully involved in the creation of evidence. Friedman made this point by asking his readers to imagine that the judicial system publicly advertised the following:

If you want to make a criminal accusation against a person, make the statement however you wish and present it to us in a way that we can pass on to the fact-finder. If you want, you can make it in person to the fact-finder, but you don’t have to. You can make it on audio or video tape, you can make it in writing (no need for a signature), you can make it by telephone (we’ve set up a special number, 911, for just that purpose), or you can make it to any person you want, with the request that he or she pass it on to us. And you don’t have to take an oath. In fact, if you want to do the whole thing anonymously, that’s OK too. We can use the statement at trial however you make it.
(Friedman & McCormick, supra, at 1247.)

This hypothetical makes the point: if the Confrontation Clause’s protection does not extend to declarants’ statements that they reasonably expect to be used by the state, then “[t]he government, or others interested in the creation of prosecution evidence, would have an incentive to encourage the making of statements—such as 911 calls—lacking formalities such as the oath, because the avoidance of such formalities would ensure that the statement would not be covered by the Confrontation Clause.” (Id.) Not all thinkers would accept this focus on declarants’ intentions, reasonable or otherwise, insightfully arguing, as does Professor Robert Mosteller in this symposium, that such an approach is unduly complex and not always consistent in application with the clause’s fundamental goal. (See Robert Mosteller’s article, “Testimonial” and the Formalistic Definition—The Case for an “Accusatorial” Fix, on page 14 of this issue, where he suggests that his approach, emphasizing as central to the testimonial inquiry whether the statements are “accusatorial,” might lead to broader confrontation protection than does Friedman’s.) What matters, however, is not whether Friedman’s or Mosteller’s precise formulation is correct but rather that an interpretation of the Confrontation Clause that gives restraining the state a central role need not necessarily lead to a narrow scope for that clause nor require reintroducing reliability as, in itself, a touchstone for the analysis. Reliability inquiries belong elsewhere, in the fundamental due process guarantees of a fair trial.

Freestanding due process
But the same cannot be said about “freestanding” due process. Due process affects criminal trials in state courts in two ways: (1) by incorporating most of the criminal procedure protections of the Bill of Rights against the states; and (2) separate and apart from any specific Bill of Rights provision (thus the “freestanding” moniker), by imposing additional procedures required by fundamental fairness. (See Jerold H. Israel, Free-Standing Due Process and Criminal Procedure: The Supreme Court’s Search for Interpretive Guidelines, 45 ST. LOUIS L.J. 303, 303-04 (2001).) Examples of freestanding due process doctrines are the rules excluding involuntary confessions and certain tainted eyewitness identifications from criminal trials. (See TASLITZ & PARIS, supra, at 590-615, 807-13.) Due process prohibits the intentional destruction of, or failure to preserve, exculpatory evidence and the use of police practices that “shock the conscience,” while ensuring defense receipt of certain reciprocal discovery, creating a prosecutorial obligation to produce material exculpatory evidence within its possession or control for the defense and granting indigents access to experts needed to evaluate defenses resting on scientific expertise. (See Israel, supra, at 389-90.) Due process likewise limits prosecution use of material testimony known to be false or prosecution failure to advise the court or the defense of evidence within its possession or control undercutting its key witnesses. (See id.) In short, despite occasional references by the Court to disfavoring freestanding due process, the legal rules it generated by the doctrine are pervasive in the criminal trial. (See id. at 395-98.)
Yet the primary, though again not sole, function of freestanding due process is to promote reliable fact-finding. By “reliable” here, I mean leading to what are, at least in theory, measurably accurate outcomes, such as convicting the killer who wielded the knife and not the innocent bystander falsely caught in a web of flawed circumstantial evidence. (See id. at 397 (“[F]ree-standing due process rulings might be characterized as ‘narrow’ in the sense that they tend to focus on the value of adjudicatory fairness, looking primarily to protect against the conviction of the innocent,” though there are strands concerned with “dignitary” values as well (emphasis added)).)
Reliability, as used here, can thus be served in at least two ways: (1) by promoting procedures enhancing the likelihood that admissible evidence correctly reflects reality; and (2) by improving the fact finder’s ability to assess the accuracy of, and weight to be accorded to, admitted evidence. The first type of method does not mean that the trial judge, rather than the jury, decides credibility questions. But it does mean that no fact finder, whether judge or jury, should be free to rely on evidence lacking some significant guarantee of trustworthiness. The precise level of that guarantee may vary with the size of what is at stake, the risk that the fact finder may have trouble assessing the evidence even with improved procedures, and the record of error from previous reliance on this category of proof. The second method of promoting reliability—seeking to augment the jury’s ability to evaluate effectively the evidence before it—relies primarily on information enhancement: getting more information about the value of the evidence to the jury and giving it more guidance concerning how to go about its task.
Courts have indeed on occasion long specifically turned to due process, though often in dicta, as a way of improving the reliability of trial evidence. The Supreme Court in a preincorporation era case, West v. Louisiana, 194 U.S. 258 (1904), thus recognized that due process might prohibit the use of depositions as substantive evidence at trial, though the Court found no such violation on the facts before it. More recently, in United States v. Enjady, 134 F.3d 1427 (10th Cir.), cert. denied, 525 U.S. 887 (1998), the appellate court considered a defendant’s claim that Federal Rule of Evidence 413—permitting use against a defendant of his prior offenses of sexual assault at his current trial for such an offense—violated the Due Process Clause of the Fourteenth Amendment. The defendant alleged that Rule 413 violated due process in three ways: (1) by ignoring the “settled usage” of the prohibition on character propensity evidence, (2) by creating a presumption of guilt, and (3) by licensing the jury to punish a defendant for past wrongs rather than current ones.
The Enjady court rejected the idea that because a “practice is ancient . . . it is embodied in the constitution.” Moreover, the court seemed to reject the argument that Rule 413 created a presumption of guilt, describing the rule instead as a way to rebalance scales illogically tipped in the accused’s favor by countering jurors’ tendency in sexual assault cases to judge the victim’s “guilt” rather than that of the accused.
However, the court apparently agreed in principle with the concern that the jury might convict the defendant despite inadequate evidence of his guilt of the currently charged crime, for the court concluded that “without the safeguards embodied in Rule 403 we would hold the rule [i.e., Rule 413] unconstitutional.” (Id. at 1433.) As the court read Rule 413, it was not a per se rule of automatic admissibility but one that required subjection first to Rule 403’s balancing of unfair prejudice against probative value. In the sexual assault context, concluded the court, Rule 403 balancing must consider the clarity of proof of the prior act’s occurrence, its probative value in the specific case, the degree of dispute over the material fact involved, whether the government has less prejudicial evidence available, the likelihood that the proffered evidence will contribute to an improper verdict, the extent to which it will distract the jury from central issues in the case, and how time-consuming will be proof of the prior conduct. All these factors but the last concern the reliability of the testimony. (See id. at 1433-35.)
Moreover, the court further concluded that there must be “sufficient evidence to support a finding by the jury that the defendant committed the similar act,” another safeguard against too hasty jury decision making. (Id. at 1433.) In other words, the court at least implicitly applied the interpretive maxim to do all that reasonably can be done to interpret statutes and rules in ways consistent with constitutional principles, in this case due process. The Enjady court’s approach seems similar to that described by legal scholar Peter Westen: “The due process clause prohibits the state (1) from convicting a defendant unless its evidence, taken as a whole, is sufficiently probative to permit the trier of fact to find him guilty beyond a reasonable doubt; and (2) from using any single item of evidence against a defendant which is inherently too unreliable for rational evaluation by the jury.” (Peter Westen, Confrontation and Compulsory Process: A Unified Theory of Evidence for Criminal Cases, 91 HARV. L. REV. 567, 598 (1978).)
Nevertheless, freestanding due process has often generated specific doctrines too weak to serve the goal of truth-seeking. There are numerous possible reasons for this, but commentators have noted several that stand out. First, although there are due process doctrines crafted in terms of relatively clear rules, too often the Court has articulated flexible utilitarian balancing tests, seeing due process as primarily about doing justice under the unique circumstances of a particular case. (See Andrew E. Taslitz, Wrongful Rights, 18 CRIM. JUST. 4, 9-10 (Spring 2003) (making this point as to both due process and as to aspects of other constitutional doctrines).) Second, the Court often makes little use of social science—which is sometimes relatively uncontradicted—or uses it inadequately, in a way that makes it too easy to weigh state interests over individual ones. (See id.; Tracey Meares & Bernard E. Harcourt, Transparent Adjudication and Social Science Research in Constitutional Criminal Procedure, 70 J. CRIM. L. & CRIMINOLOGY 733 (2000) (on Court’s general failures in using social science in constitutional criminal procedure.) Third, the Court may take too cautious an approach to creating incentives for the generation of more and better information to help the jury. (Cf. Dale Nance, The Best Evidence Principle, 73 IOWA L. REV. 227 (1988) (on importance of such incentives in evidence law more generally).)
These weaknesses have particularly been highlighted by the innocence movement. That movement has documented significant numbers of cases in which supposedly voluntary confessions resulted in convicting the wrong man, cross-examination concerning the weaknesses in lineup and photospread identifications proved fruitless, “scientific” tests were mistaken, and informants’ credibility was misjudged by fact finders. (See Taslitz, supra, at 4-12.) Toothless due process rules did little to exclude this flawed evidence from trial or to enhance the jury’s ability to avoid error, as has been well documented in the pages of Criminal Justice on several occasions over the last few years. (See, e.g., symposium on Wrongful Convictions, 18 CRIM. JUST. (Spring 2003).)
The errors brought to light by the innocence movement have led numerous organizations, including the American Bar Association (ABA), to call for reform in both pretrial and trial procedures and have led several commentators to call for a reinvigorated due process methodology designed to prevent exposing individuals to unacceptably high risks of mistaken conviction. (See generally BRIAN FORST, ERRORS OF JUSTICE: NATURE, SOURCES, AND REMEDIES (2004).) Professor Donald Dripps, in crafting an extended argument for a more muscular procedural due process of this nature, has embraced continuing use of balancing tests but stripped of “the concerns with founding-era common law practice and the delicate attitude toward state prerogatives” that “are entirely appropriate in general substantive due process analysis, but . . . [that] are quite inapposite in procedural due process analysis,” for these limitations ignore the lessons of experience and of social science. (DRIPPS, supra, at 141.) On the other hand, where possible, he would favor using these balancing tests to craft more specific rules to give clear guidance in future cases. (See id. at 155-59.) Other scholars, like Professor Welsh White, again relying on social science and experience, favor wider use of specific rules over balancing tests. Thus White would exclude confessions resulting from any interrogation lasting over six hours; given by mentally handicapped defendants; or induced by threats of punishment, promises of leniency, threats of adverse consequences to a friend or loved one, or misrepresentations of the nature and scope of the evidence against a loved one. (See WELSH WHITE, MIRANDA’S WANING PROTECTIONS: POLICE INTERROGATION PRACTICES AFTER DICKERSON 196-215 (2001).) These bright-line rules help to overcome too-easy deference to the state and to give concrete meaning to the quest for accuracy.
The wisdom of relying on a strengthened due process doctrine in controlling the admissibility of hearsay has at least been impliedly challenged by some commentators who argue that hearsay is not necessarily less probative than live, in-court testimony and that jurors are well equipped to give hearsay its proper weight. (See, e.g., Nance, Rethinking Confrontation, supra, at 1, 12-13 (viewing a reliability-based approach to the exclusion of hearsay as reflecting unwarranted distrust of the jury).) This argument might be understood primarily as one for elimination of the general prohibition against hearsay, something that courts and legislatures in this country have so far been loath to do, though a hearsay bar can still be justified on other grounds. (See id. at 13 (exclusion might encourage production of the best evidence available).) Moreover, it relies on a body of research that is too small and too young to support any firm conclusions. (See, e.g., Roger Park, Visions of Applying the Scientific Method to the Hearsay Rule, 2003 MICH. ST. L. REV. 1149, 1170 (empirical research is too young to justify abolishing the hearsay rule on those grounds, and it is likely that hearsay’s value varies with the type of hearsay and the situation).) Some of the research, furthermore, instead supports the view that some hearsay in fact poses the classic sort of risks of promoting fact finder error that prompted the hearsay rule in the first place. (See, e.g., William C. Thompson & Maithilee K. Pathak, Empirical Study of Hearsay Rules: Bridging the Gap between Psychology and Law, 5 PSYCH., PUB. POL’Y & L. 456, 469 (1999) (some research suggests jurors may overvalue hearsay or be insensitive to reliability variations in the declarant or witness under certain circumstances); Amye R. Warren & Cara E. Woodall, The Reliability of Hearsay Testimony: How Well Do Interviewers Recall Their Interviews with Children? 5 PSYCH., PUB. POL’Y & L. 355, 369 (1999) (although more research is needed, their results “suggest that the hearsay testimony of children’s interviewers is degraded”).) Some of the critics have likewise conceded that at least some hearsay is of so little probative value and such great jury impact that it merits great skepticism. (See Park, supra; Thompson & Pathak, supra; Warren & Woodall, supra.) Finally, the examples enumerated below illustrate how very dangerous some hearsay can be.
A revitalized hearsay due process would not render pre-Crawford Confrontation Clause cases irrelevant but would require viewing them in a new light. As Professor Jerold Israel has explained, the Sixth Amendment right to confrontation, as it existed in the pre-Crawford world, might “readily be described as focusing primarily on the value of adjudicatory fairness,” much in the way that the Due Process Clause does. (See Israel, supra, at 398.) The earlier Confrontation Clause case law should, therefore, in the post-Crawford world, be recast as due process cases, but their reasoning and holdings should be modified somewhat by the lessons of the innocence movement. That would mean, for example, deep-sixing Roberts’s “firmly rooted” notion because that encourages adherence to ancient doctrines proven to lead too often to mistakes. It would also mean a quest for more bright-line rules, where appropriate; to close examination of scientific research, where available; and to increased sensitivity to improving jurors’ performance, as well as that of the police.
In the remainder of this article, I draw on one extended example stemming directly from the innocence movement—namely, introduction at trial of a witness’s out-of-court identification of the offender—to illustrate the sort of methodology to be used in a vigorous hearsay due process doctrine. I conclude with some very brief speculations on how the lessons of that example may be applied to a wide array of other hearsay problems.

Prior statements of identification of a person
Under Federal Rule of Evidence 801(d)(1)(C), a statement of identification of a person made after perceiving that person—such as one made at a lineup or during a photo spread or in a showup—is excluded from the hearsay rule if the declarant testifies at the trial or hearing and is subject to cross-examination concerning the statement. Because this hearsay exclusion requires the hearsay declarant to be available for cross-examination at trial, the dictates of the Crawford rule are probably met. (Cf. Mosteller, Ensuring Confrontation, supra, at 578-86 (arguing that the witness must actually testify at trial rather than simply being “available” to do so).) Such hearsay has a powerful effect on juries, that see out-of-court identifications made shortly after the crime as of far greater probative value than are in-court ones where the defendant can easily be picked out (he sits at defense counsel’s table) and where the witness testifies months or years after the crime. (See AMERICAN BAR ASSOCIATION, REPORT IN SUPPORT OF THE CRIMINAL JUSTICE SECTION RESOLUTION ON EYEWITNESS TESTIMONY 12-13, available at (2004); United States v. Wade, 388 U.S. 218, 232 (1967) (“The trial which might determine the accused’s fate may well not be that in the courtroom but that at the pretrial confrontation, with the State aligned against the accused, the witness the sole jury, and the accused unprotected against the overreaching, intentional or unintentional, and with little or no effective appeal from the judgment there rendered by the witness—‘that’s the man.’ ”).)Yet, as the innocence movement has demonstrated, such evidence is often of questionable reliability. (See generally ABA REPORT, supra.)
Current due process doctrine concerning the admissibility of pretrial identification procedure is insufficient to address this problem for several reasons. First, it too is as concerned with affirmative police wrongdoing as much as reliability. Second, its reliability test is uninformed by the current teachings of the best social science. Third, it relies on flexible case-specific balancing tests rather than per se rules. A rule-informed hearsay due process focused on reliability in itself, to the exclusion of any other concerns, would do far better. (Cf. Taslitz, supra, at 6 (arguing mere admission of unreliable evidence meets the state action requirement).) Moreover, pre-Crawford confrontation jurisprudence, while generally purporting to create an opportunity for effective cross-examination, has in practice reached results more consistent with requiring that the defense merely have the opportunity to pose questions, regardless of how useful that questioning might be in fully informing the jury. That is an inadequate safeguard against error. For example, in United States v. Owens, 484 U.S. 554 (1988), an assault victim had been hospitalized because of a skull fracture suffered in an attack. While in the hospital, the victim identified the defendant in a photo array. At trial, however, the victim could not remember that he had seen his assailant. Furthermore, although the victim remembered identifying the defendant in the photo array, the victim could not recall whether any of his hospital visitors had suggested, before the array, that the defendant was the assailant. The Court considered whether, despite the victim’s memory lapse, the defendant had an adequate opportunity for effective cross-examination of the victim.
The Court found that the defendant indeed had an opportunity for effective cross-examination because he had “realistic weapons” for cross-examination. This was demonstrated by defense counsel’s summation, which emphasized the witness’s memory loss and argued that the witness had identified the defendant solely because of the suggestions of visitors to the hospital. The Court rejected the argument that, given the grave dangers of identification testimony, the out-of-court identification was so inherently unreliable that the tools used by defense counsel were inadequate means of impugning the victim’s testimony in this particular case. The Court emphasized, however, that it based this final conclusion on the absence of any argument that the photo array was conducted in a suggestive manner. But, given the findings of the innocence movement and of supporting psychological research, the questioning in Owens was unlikely to create any realistic chance of overcoming the de facto presumption of guilt arising from the mere fact of the out-of-court identification. (See generally Taslitz, supra (summarizing research).) Furthermore, the Court seemed implicitly to conclude that the defense must produce evidence of some affirmative highlighting of the defendant as the offender to constitute “suggestion.” Given the easy accessibility to lay audiences today of data on how to conduct fair lineups and the risks of error from not doing so, the mere absence of sound procedures should be considered suggestive, creating the corresponding need to screen unreliable evidence from the jury or to enhance its ability to make a fully informed evaluation of the evidence’s quality. (See ABA REPORT, supra (summarizing readily accessible sources).)
A better approach would be one more consistent with the Court’s emphasis in right to counsel cases on the need to provide the defense with effective tools, not simple questioning, to preserve fundamental confrontation values. Thus, in United States v. Wade, 388 U.S. 218 (1967), the Court declared that the Sixth Amendment right to counsel extended to postindictment lineups because the “inability effectively to reconstruct at trial any unfairness that occurred at the lineup may deprive [the defendant] of his only opportunity meaningfully to attack the credibility of the witness’ courtroom identification.” (Id. at 232.) The Court continued:
Insofar as the accused’s conviction may rest on a courtroom identification in fact the fruit of a suspect pretrial identification which the accused is helpless to subject to effective scrutiny at trial, the accused is deprived of that right of cross-examination which is an essential safeguard to his right to confront the witnesses against him.
(Id. at 235.)

The Wade Court found that a right to counsel was necessary at a lineup because of the mere possibility that, without counsel, it would be impossible effectively to disclose at trial whether the lineup was “suspect,” that is, conducted in a suggestive manner. Similarly, sound confrontation principles would require, for example, videotaping of lineups to preserve the chances for effective cross at trial and proper concerns about fairness would require close attention to the reliability of the lineup procedures that may ultimately be what determine the offender’s guilt or innocence. (See ABA REPORT, supra.)
Professor John Douglass has made an analogous pre-Crawford argument for “a right to real Confrontation” under the Confrontation Clause, including (1) a right to call a hearsay declarant as a witness, and the state’s assistance in doing so; (2) broad discovery rights to enable effective impeachment of a present or absent hearsay declarant; (3) broad latitude at trial to impeach such a declarant, for example, by using prior inconsistent statements; and (4) government efforts to preserve useful ways to challenge the hearsay declarant, such as via depositions of the declarant videotaped for use at trial. (See John G. Douglass, supra, at 241-70.) Although Douglass was more concerned with what he called “real confrontation” than with reliability, and although his suggestions can be justified under the Confrontation Clause too in a post-Crawford world (because discovery and evidence preparation aid in disclosing governmental abuses in the creation of evidence), his recommendations also aid in improving reliability in the sense of the jury’s capacity to evaluate fairly the weight of the evidence. Moreover, given the Court’s apparent de facto unwillingness thus far to adopt the sort of procedures that Douglass suggests as matters of the law of confrontation, due process might offer a more promising avenue for redress. (Cf. Pennsylvania v. Ritchie, 480 U.S. 39 (1987) (plurality suggesting that the mere opportunity to pose questions is sufficient under the Confrontation Clause, but Justice Blackmun, in concurrence, arguing that wide pretrial discovery may sometimes be essential to protecting the opportunity for effective cross-examination at trial); Eleanor Swift, A Foundation Fact Approach to Hearsay, 75 CAL. L. REV. 1339 (1987) (given that most categorical hearsay exceptions do little to ensure evidentiary accuracy, a gloss should be added to those exceptions by which a “foundation fact witness” must produce for the jury adequate evidence about the circumstances under which the hearsay statement was made and its nature to enable the jury to make a reasonable assessment of its likelihood of being true).)

Manson’s impossible balancing test
The Court’s current due process test renders out-of-court identifications inadmissible if they are so unnecessarily suggestive as to create a very substantial likelihood of irreparable misidentification. (See TASLITZ & PARIS, supra, at 807-09.) “Necessity” seems similar to “urgency” and focuses, therefore, on the police’s need to act heedless of accuracy, though also seemingly involving a cost/benefit analysis of how much effort we can expect the police to expend to get the right man. (See Taslitz, supra, at 8.) This test thus tolerates significant risks of error where there is no affirmative police wrongdoing rendering suggestion “necessary.” (But see Neil v. Biggers, 409 U.S. 188 (1972) (a showup was “unnecessarily” suggestive, despite the police finding no one at either the city jail or the juvenile home on the date of the identification procedure who fit a physical description comparable to the defendant’s).) This doctrine also relies on a weighing of a commonsense set of circumstances rather than the teachings of experimental research. Once suggestion has been shown, a pretrial identification is admissible if it is nevertheless likely to be reliable based upon weighing the degree of corruption resulting from the suggestion against these five factors: (1) the opportunity of the witness to view the offender at the time of the crime, (2) the witness’s degree of attention, (3) the accuracy of the witness’s own description, (4) the witness’s degree of certainty, and (5) the tie between the crime and the identification. (Manson v. Braithwaite, 432 U.S. 98, 114 (1977).) As research psychologist Gary Wells points out in an important forthcoming paper, the first four of these factors rely on “self-reports” by the witness, that is, the witness’s own assessment of how attentive or certain he was or of how much chance he had to observe. (See Gary L. Wells, What Is Wrong with the Manson v. Braithwaite Test of Eyewitness Identification Accuracy? 1, 7 (unpublished draft, Oct. 2004).) But “[v]olumes of research in social and cognitive psychology indicate that self-reports can be notoriously unreliable.” (Id.) This observation is especially true of eyewitness reports, which notably overestimate the time and opportunity for observation and are especially skewed by the effects of stress or anxiety at the time of observation. (See id.)
Moreover, some of these factors are simply not good predictors of accuracy, there, for example, being no “close correspondence between the description given by the eyewitness and the likelihood that the identification is accurate.” (Id. at 7-8.) Furthermore, and perhaps most worrisome, these factors are not independent of suggestion, that is, once a witness has made an identification in a suggestive procedure, the witness’s self-reports change from those made before the procedure, the witness giving higher estimates of how good a view he had of the wrongdoer, how well he was able to make out the criminal’s face, and how certain he was of the identification at the time that he made it. (See id. at 8.) Yet such a witness is likely confidently to deny that the suggestion affected any of his self-reports and will suddenly “remember” new details that are miraculously consistent with his suggestive identification but that escaped his memory before that time. (See id. at 8-9.) Later fair procedures will not cure any error, as Wells again explains:

If an eyewitness mistakenly identifies a person from a photo lineup, a later identification test involving that person is likely to produce the same result. In the original demonstration of this . . . , some witnesses to a staged crime were shown a 12- person photospread and others (a control condition) were not. The 12-person photospread did not include the staged-crime culprit, but suggestive instructions led all of the witnesses who viewed the photospread to select someone. Four days later, all witnesses were shown a fair six-person photo-lineup that included the person picked earlier and also included the actual culprit from the staged crime. Although the control condition witnesses were able to correctly identify the culprit 39% of the time, those who had been shown the suggestive photo-spread earlier could not identify the culprit and instead tended to choose the same person they had erroneously chosen earlier. Similar results have been reported in other published experiments using variations on these procedures . . . . The main point is that a suggestive identification procedure taints all later identification procedures involving that same witness and suspect, regardless of how fair any later procedure might appear to be on its own.
(Id. at 5 (emphasis added).)

Current safeguards are unlikely to correct for these problems. Thus, the right to counsel at an identification procedure does not apply at photo spreads, which today probably constitute 80 percent of all pretrial initial identifications (id. at 6), and also does not apply precomplaint or preindictment. (See TASLITZ & PARIS, supra, at 807-08.) Counsel are themselves too often ignorant of the relevant science (Wells, supra, at 6), a state of affairs likely to change if courts recognize due process grounds for potential suppression where safeguarding procedures are absent, for then counsel will have a greater incentive to learn the science in order to prevail at trial. Nor will ordinary cross-examination techniques generally suffice because juries give tremendous weight to witness identifications (see ABA REPORT, supra, at 12-16), especially for very confident and persuasive but mistaken witnesses who are by definition unaware of their error:

The case of victim-witness Jennifer Thompson vividly illustrates this point. Thompson had mistakenly identified Ronald Cotton as her attacker. Later, when her actual attacker, Bobby Poole, was shown to her at a hearing, she said, “I have never seen this man before.” Surely there is no more powerful cross-examination tool than showing the eyewitness the actual culprit, and yet even that technique will not undo a mistaken identification.
(Wells, supra, at 6 (also discussing the weaknesses in voir dire procedures and too many courts’ reluctance to improve jury assessment of eyewitness accuracy via scientifically informed jury instructions and via experts on the science behind eyewitnessing).)

These are all good reasons for due process generally to require that the following safeguards be implemented at most pretrial identification procedures, safeguards endorsed by the ABA as a matter of sound policy but which I argue should also be mandated as constitutional principle:

1. The use of double-blind procedures in which neither the detective administering the lineup or photospread nor the witness has any idea who is the suspect, thus avoiding the risk of even unintended hints subconsciously being conveyed to the witness. (The ABA suggests this procedure only where “practicable.”)
2. Giving the witness instructions clearly conveying that the culprit may not be in the lineup or photospread; that they should not assume that the administrator knows who, if anyone, is the suspect; and that they should not feel compelled to identify anyone.
3. Using foils chosen for their similarity to the witness’s description of the perpetrator, without the suspect’s standing out in any way from the foils and without other factors drawing undue attention to the suspect.
4. Immediately after the identification procedure, requesting witnesses to indicate their level of confidence in any identification and ensuring that their response is accurately documented.
5. Avoiding at any time giving the witness feedback on whether he or she selected the “right man,” i.e., the person believed by law enforcement to be the culprit.
(American Bar Association Statement of Best Practices for Promoting the Accuracy of Eyewitness Identification Procedures, Dated Aug. 24, 2004.)

The ABA policy recommends a variety of additional best practices, but I have listed above only that subset of practices that are so strongly supported by the scientific research and so essential to avoiding mistaken identifications that ignoring any one of these requirements should presumptively constitute a due process violation. (See id.; American Bar Association Resolution on Eyewitness Identification (recounting ABA’s statement of additional eyewitness identification procedure best practices).) None of these procedures imposes an undue financial burden on even small police departments, given technological advances and the freedom to use photospreads rather than live lineups. (See ABA REPORT, supra, at 11-12.) Given the early stage of some portions of the research on showups, I would not find a due process violation in using that procedure near the time of the crime, when memories may be freshest, but would find such a violation where showups are used significantly after the time of the crime. (See id. at 10 & nn.55-56.)
Where these safeguards are missing, however, note that I have only said that use of the hearsay identification at trial would presumptively violate due process. There may be some unusual circumstances where, even in the absence of one or more of these requirements, there is strong reason to believe that the identification is nevertheless reliable: “Consider . . . an abduction case in which the victim-witness was held for days by the culprit, the culprit never wore a mask, there was good lighting, the culprit’s face was in full view for hours at a time, and the witness was able to describe in detail the size, shape, and location of unique scars on the culprit’s face.” (Wells, supra, n.10.) Any identification made at a lineup held only hours after the victim’s rescue but in which the only error was that the fillers did not fit the culprit’s description should still be admissible because of the significant likelihood of accuracy given little in the way of flaws when memories are so fresh. (See id.) Additionally, in any case in which there is a properly administered procedure followed by a suggestive one, statements of identification made before the flawed procedure should be admissible. (See id.) Professor Wells would go even further and, once suggestiveness (meaning the failure to follow scientifically informed procedures) is shown, would shift the burden of reliability to the state, a burden that could be met only by a combination of presuggestion statements and circumstances establishing that the same identification would have been made independently of any suggestion. This should, however, be a difficult burden to meet (id.), especially because it reintroduces the concerns that science raises about flawed self-reports and that experience raises about the malleability of case-specific balancing inquiries. Nevertheless, it would be an improvement over the current Manson test because a clear per se set of rules would be the jumping off point, any deviation from which must be justified or explained away in scientifically supportable terms and with a presumption against reliability. This approach to remedy is arguably analogous to that recently embraced by a majority of the Court in the area of sentencing and the right to trial by jury in United States v. Booker, 125 S. Ct. 738 (2000). Post-Booker, the Federal Sentencing Guidelines are apparently advisory, but intense appellate scrutiny will be given to any departures from those guidelines that are not thoroughly justified by adequately stated reasons. (See id.) Likewise, under this approach to the due process hearsay exclusion for pretrial eyewitness identifications, departure from the reliability guidelines would mean presumptive exclusion of the evidence absent a statement of scientifically informed, adequate reasons for departures. Police and prosecutors can, of course, avoid the risk of suppression entirely by doing the identifications correctly in the first place.
I would, moreover, add one piece to Wells’s solution to the Manson puzzle: if a suggestive identification is nevertheless admitted under the Federal Rule of Evidence 801(d)(c) hearsay exception as sufficiently reliable, that admission should be subject to procedures for improving the likely effectiveness of cross-examination in promoting truth-discovery. Here I am returning to the two senses in which I have been using the term “reliability.” The proposed safeguards help to maximize the likelihood that the identification is reliable in the sense of being accurate. The absence of even some of those safeguards, however, raises sufficient accuracy concerns that we need to be more concerned about reliability in its second sense: improving the jury’s ability adequately to evaluate how accurate potentially flawed identification evidence is likely to be, that is, how much weight it deserves. As I explained above, the mere opportunity to pose questions will ordinarily be inadequate to aiding juries in this task. The jurors need more help, which they can get from judges, lawyers, and legislators who recognize the need for cautionary instructions, the right to offer expert testimony on the factors affecting lineup accuracy, and the importance of expansive pretrial discovery.

Applying this method to other hearsay exceptions
Space does not allow me to address other hearsay exceptions in any depth. Nevertheless, there is ample reason to believe that this due process methodology may be useful with many other hearsay exceptions. Again, the basics of a hearsay due process methodology would be a single-minded focus on reliability, grounded in social science, ideally crafting per se rules over flexible guidelines with departures permitted only upon the prosecution’s offering scientifically supported evidence establishing adequate reason for doing so, and, where accuracy assurances are lacking, requiring at a minimum vigorous efforts to ensure that the jury is fully informed in a way that will enable it effectively to gauge the weight to which the evidence is entitled. (Cf. Rock v. Arkansas, 483 U.S. 44, 60 (1987) (suggesting that a guidelines approach to the admissibility of expert testimony under the Compulsory Process Clause might be wise); Maryland v. Craig, 497 U.S. 836, 840, 851 (1990) (approving, under the Confrontation Clause, a closed circuit television procedure involving a child witness where there were “safeguards of reliability and adversariness”).) But cf. Idaho v. Wright, 497 U.S. 805, 818-19 (1990) (rejecting reliance on a preconceived set of procedural safeguards in determining compliance with the Confrontation Clause, at least under the circumstances before the Court).) The sort of analysis engaged in above for the hearsay exemption for pretrial identifications can thus be replicated for other categories of hearsay.
For example, in jurisdictions permitting use of children’s hearsay in child abuse cases, some scholars have convincingly argued that certain procedural safeguards against children’s confabulating evidence—sincerely remembering as true things that never happened—are required. (See LUCY MCGOUGH, CHILD WITNESSES: FRAGILE VOICES IN THE AMERICAN LEGAL SYSTEM (1994).) Absent such safeguards, expert testimony on children’s memory may be necessary. (See John E.B. Myers et al., Hearsay Exceptions: Adjusting the Ratio of Intuitions to Psychological Science, 65 L. & CONTEMP. PROBS. 3 (2002).) Other scholars have sought to attack entire categories of hearsay exceptions as unreliable. One illustration is the argument that stress in fact corrodes the accuracy of witness perceptions and memory, calling into question the wisdom of a per se excited utterance exception. (See Aviva Orenstein, “‘My God!’: A Feminist Critique of the Excited Utterance Exception to the Hearsay Rule, 85 CAL. L. REV. 159, 178-83 (1977).) The sort of analysis argued for here might suggest the availability of depositions and document request procedures to uncover suggestive statements made to children or other reasons to doubt the trustworthiness of otherwise admissible hearsay. Furthermore, although the admission of hearsay under the business records and public records exceptions can already be challenged as untrustworthy in an individual case under the Federal Rules of Evidence, a due process version of that requirement in criminal cases would ensure the burden of trustworthiness’s being placed on the prosecution and of the statutory rules being applied in a firm fashion. (See FED. R. EVID. 803(6)(8).)
I take no position on the wisdom of these assertions here. My point is simply that some variant of a reinvigorated due process analysis that draws on the best available social science; that favors per se rules over case-specific balancing tests, where plausible; and that looks for assurances of hearsay accuracy but, where they are flawed, at least requires significant steps in augmenting the jury’s ability correctly to assess such accuracy should have wide applicability. In a post-Crawford world, it is clearer and simpler to leave to the Confrontation Clause the job of preventing governmental abuses in evidence-creation and to leave assurances of reliability to the Due Process Clause. What remains of reliability after Crawford? Plenty. You just have to know where to look.

Return to Table of Contents - Summer 2005

Return to Criminal Justice magazine home page