ACLU in the Courts: Dickerson v. United States

On Writ of Certiorari to the
United States Court of Appeals
for the Fourth Circuit

Brief for Amicus Curiae
The American Civil Liberties Union
in support of petitioner

The American Civil Liberties Union is a nationwide, nonprofit, non-partisan organization with nearly 300,000 members that has been engaged in defense of the Bill of Rights for 80 years.^[1] Many of its efforts have focused on enforcing those portions of the Bill of Rights having to do with the administration of criminal justice, including participation as amicus curiae in Miranda v. Arizona, 384 U.S. 436 (1966). This brief will concentrate on the legal setting for this Courtís decision in Miranda, the factual circumstances of custodial interrogations today, and the evidence concerning Miranda's supposed "costs" to law enforcement. That limited focus, however, should not obscure our broader concern. The ACLU continues to believe, as we believed in 1966, that if the Fifth Amendment privilege is to remain an effective guarantor of our accusatorial system of criminal justice, some warning is necessary to assure that persons subjected to custodial interrogation know they are not obliged to respond to questioning by the police and that the police will respect their right to remain silent.

This Court's landmark decision in Miranda v. Arizona was a necessary response to the failure of the due process voluntariness test to protect adequately the core values of our accusatorial system of justice, embodied in the Fifth Amendment's prohibition against compelled self-incrimination. The per se nature of the holding in Miranda was entirely consistent with the law this Court had developed in other contexts to determine whether a particular government action amounted to unlawful compulsion under the Fifth Amendment.

In the past thirty-four years, Miranda's requirements have been crafted into a series of workable rules that fully respect the legitimate needs of law enforcement, and Miranda has become perhaps the most widely recognized constitutional protection in our criminal justice system. Stare decisis strongly counsels against wholesale changes in these rules "at this late date." Rhode Island v. Innis, 446 U.S. 291, 304 (1980) (Burger, C.J., concurring in the judgment). Nor is any such change justified.

Custodial interrogations today involve the same inherent compulsion that they did when Miranda was decided. Police manuals and reported cases continue to detail the sometimes illegal lengths to which law enforcement officials will go, even with the protections of Miranda, to get a confession from a suspect. These sources provide a disturbing glimpse of what police practices could become in a world without Miranda - where police need not tell citizens of the right to remain silent, and need not respect that right once invoked.

Some contend that warning citizens about their constitutional rights has imposed great "costs" to law enforcement. In fact, the evidence is directly to the contrary - in the vast majority of cases, police are able to conform their practices to the dictates of Miranda without difficulty. In the end, Miranda has established a system that works, both for police and for the citizens they chose to interrogate.

The Fifth Amendment protection against compelled self-incrimination embodies core values of our accusatorial system of criminal justice. The Miranda decision was a necessary response to the failure of the due process voluntariness test to adequately protect the privilege or guide courts or the police. Mirandaís per se requirement was consistent with prior due process voluntariness analysis, which also included some per se rules, and with this Court's approach to the distinct question of assessing compulsion under the Fifth Amendment. Miranda's holding, itself a compromise, has resulted in a series of rules that provide understandable and workable guidelines concerning custodial interrogations to courts and police alike.

Our accusatorial system of justice is based on the absolute right of citizens accused of crime not to confess, but instead to stand silent and put the prosecutor to his proof. The privilege against self-incrimination is the "essential mainstay" of that accusatorial system of prosecution. Malloy v. Hogan, 378 U.S. 1, 7 (1964).^[2] The privilege against self-incrimination protects fundamental notions of personal autonomy by giving citizens suspected of crime the right to choose whether to remain silent or to speak. See Miranda, 384 U.S. at 469; Moran v. Burbine, 475 U.S. 412, 426-27 (1986). By allowing people to choose not to speak with law enforcement, the privilege also reflects respect for personal privacy and individual dignity. See Couch v. United States, 409 U.S. 322, 327 (1973); Miranda, 384 U.S. at 460. Thus, the privilege embodies "principles of humanity and civil liberty, which had been secured in the mother country only after years of struggle, so as to implant them in our institutions in the fullness of their integrity, free from the possibilities of future legislative change." Bram v. United States, 168 U.S. 532, 544 (1897) (emphasis added).

In addition to protecting the accusatorial system of criminal justice, the privilege against self-incrimination plays an important role in protecting against abuses that may too easily taint any custodial interrogation. Custodial interrogation is designed to exploit the weaknesses of suspects, Miranda, 384 U.S. at 455, and this Court has found "a substantial risk that the police will inadvertently traverse the fine line between legitimate efforts to elicit admissions and constitutionally impermissible compulsion." Moran v. Burbine, 475 U.S. at 426 (citation omitted); see Escobedo v. Illinois, 378 U.S. 478, 489 (1964) ("The exercise of the power to extract answers begets a forgetfulness of the just limitations of that power. The simple and peaceful process of questioning breeds a readiness to resort to bullying and to physical force and torture.") (quoting John H. Wigmore, 8 Evidence ß 309 (3d ed. 1940)); Lawrence Herman, The Supreme Court, The Attorney General, and the Good Old Days of Police Interrogation, 48 Ohio St. L.J. 733, 752 (1987). The privilege against self-incrimination thus protects the reliability of our system of criminal justice. Johnson v. New Jersey, 384 U.S. 719, 730 (1966).

Finally, as this Court recently emphasized in Withrow v. Williams, 507 U.S. 680 (1993), the privilege "safeguards ëa fundamental trial right.' " Id. at 691 (quoting United States v. Verdugo-Urquidez, 494 U.S. 259, 264 (1990)). Without adequate protection during custodial interrogation, the privilege against self-incrimination invoked at trial would be an "empty formalit[y]." Miranda, 384 U.S. at 466.

It was not until Miranda that the Court came to terms with the compulsion inherent in every custodial interrogation. Prior to that, the Court only reviewed statements given in the course of custodial interrogations to determine whether they were voluntary pursuant to the due process clause, an inquiry that focused on whether a statement was a product of the suspect's overborne will. Miranda was, in large measure, a reaction to the inability of the due process voluntariness analysis to protect the core values reflected in the Fifth Amendment's prohibition against compelled self-incrimination. These shortcomings have been noted not only by supporters of Miranda, see, e.g., Yale Kamisar, Confessions, Search and Seizure and the Rehnquist Court, 34 Tulsa L.J. 465, 471-72 (1999), but also by its critics, see Joseph D. Grano, Voluntariness, Free Will, and the Law of Confessions, 65 Va. L. Rev. 859, 863 (1979) (noting "the intolerable uncertainty that characterized the thirty-year reign of the due process voluntariness doctrine"). See Miller v. Fenton, 474 U.S. 104, 114 n. 4 (1985) ("The voluntariness rubric has been variously condemned as useless, perplexing, and legal double talk.") (internal quotations and citations omitted).

In its many voluntariness decisions prior to Miranda, the Court generally applied a "totality of the circumstances" test that gave no guidance to police and produced a welter of unpredictable results. In some cases, the "totality of the circumstances" included examination of the personal characteristics of the suspect in an effort to determine retrospectively whether the particular suspect had the ability to withstand interrogation without "breaking down." See, e.g., Reck v. Pate, 367 U.S. 433, 441-42 (1961); Spano v. New York, 360 U.S. 315, 321-22 and n.3 (1959); Fikes v. Alabama, 352 U.S. 191, 193 (1957). In other cases, the personal characteristics of the defendant were not so much as mentioned, with the analysis focusing on the conduct of the police, including whether the suspect had been allowed to consult with family members, friends, or counsel, or advised of his right to remain silent. See, e.g., Turner v. Pennsylvania, 338 U.S. 62 (1949); Harris v. South Carolina, 338 U.S. 68 (1949); Leyra v. Denno, 347 U.S. 556 (1954); Malinski v. New York, 324 U.S. 401 (1945).

In addition to its "totality of the circumstances" test, the Court also developed, as a matter of due process, a number of disjointed rules that held confessions inadmissible per se, regardless of other "circumstances." See Catherine Hancock, Due Process Before Miranda, 70 Tul. L. Rev. 2195 (1996). In the most obvious example, the Court held that a confession that was the product of physical force could not be admitted at trial. The Court explained that where there is "[p]hysical violence or threat of it by the custodian of a prisoner during detention," there "is no need to weigh or measure its effects on the will of the individual suspect," because such confessions are "too untrustworthy to be received as evidence of guilt." Stein v. New York, 346 U.S. 156, 182 (1953); see Brown v. Mississippi, 297 U.S. 278 (1936); White v. Texas, 310 U.S. 530 (1940).^[3] The Court also established a per se rule that a confession following continuous questioning of the suspect for thirty-six hours without rest or sleep was "so inherently coercive" that the resulting confession had to be excluded without inquiry into the remainder of the circumstances or their actual impact on the particular suspect. Ashcraft v. Tennessee, 322 U.S. 143, 154 (1944).

It is incorrect, therefore, to contend that prior to Miranda there was a single voluntariness test that governed the admissibility of confessions. Instead, during the period between Brown v. Mississippi and Miranda, there was a "cornucopia of Due Process tests." See Hancock, supra, 70 Tul. L. Rev. at 2237. The result was a body of law that gave little to no guidance to police or to lower courts about what practices or "circumstances" this Court would find consistent with due process and what would cross the line. See Kamisar, Confessions, supra, 34 Tulsa L.J. at 471-72; Herman, The Supreme Court, supra, 48 Ohio St. L.J. at 752.

It is a common misperceptionóshared by the court belowóthat Miranda's per se warning requirement represented a "new" analytical approach. See United States v. Dickerson, 166 F.3d 667, 685 (4th Cir. 1999). The only ground-breaking aspect of Miranda was its recognition that, like other governmental actions aimed at eliciting information from unwilling suspects, custodial interrogation raises Fifth Amendment issues of compulsion in addition to the due process question of voluntariness. Although the "compulsion" that violates the Fifth Amendment has occasionally been equated with the "involuntariness" or "overborne will" that violates due process,^[4] the two tests have always been distinct.^[5] In analyzing compulsion under the Fifth Amendment, this Court typically has not applied a "totality of the circumstances" test. Rather, the Court has held that many governmental actions intended to elicit information from unwilling suspects, or to impose a cost on asserting the privilege, are impermissibly "compelling" per se. Thus, Miranda's conclusion that unwarned custodial interrogation involves inherent compulsion fits comfortably within this Court's application of the Fifth Amendment in other contexts, both before Miranda and since.

The classic example of impermissible Fifth Amendment compulsion is the threat of a contempt citation. Such a threat has always been treated as compelling per se, see, e.g., Hoffman v. United States, 341 U.S. 479 (1951); New Jersey v. Portash, 440 U.S. 450 (1979), even though many witnesses can and do resist this threat and go to jail rather than testify. See, e.g., United States v. Wilson, 421 U.S. 309 (1975).

A threat to discharge a public employee for failure to testify is also compelling per se and therefore impermissible under the Fifth Amendment. See, e.g., Garrity v. New Jersey, 385 U.S. 493 (1967); Gardner v. Broderick, 392 U.S. 273 (1968). This Court has never suggested that the totality of circumstances must be assessed in such cases to determine if such a threat "overbears" a particular employee's will depending on his economic circumstances. See Lefkowitz v. Turley, 414 U.S. 70, 83 (1973) (Fifth Amendment compulsion occurs per se when refusal to waive privilege results in disqualification from contracting with public agencies); accord Lefkowitz v. Cunningham, 431 U.S. 801, 806 (1977). Similarly, a comment on a defendant's failure to testify is impermissible per se regardless of the "circumstances," because it imposes a cost on assertion of the privilege. See, e.g., Carter v. Kentucky, 450 U.S. 288 (1981); Brooks v. Tennessee, 406 U.S. 605 (1972); Griffin v. California, 380 U.S. 609, 614 (1965).^[6]

Thus, in holding that custodial interrogation of an unwilling or unwarned suspect involves inherent compulsion, Miranda applied to police questioning the same per se approach that has been a familiar staple of this Court's Fifth Amendment jurisprudence. Indeed, it is ß 3501, not Miranda, that departs from the traditional Fifth Amendment concept of compulsion. Section 3501 provides that all incriminating statements "shall be admissible" if found to be "voluntarily made . . . tak[ing] into consideration all the circumstances." 18 U.S.C. ß 3501(a), (b) (emphasis supplied).^[7] The totality-of-circumstances approach that would be mandated by ß 3501 is irreconcilable with this Court's conclusion in Hoffman, Portash, Gardner, Griffin, and their progeny that certain government action (or inaction, as in Carter) involve impermissible compulsion per se, without regard to the defendant's particular circumstances.^[8]

The Miranda decision was a compromise. This Court could have held that any statement made in the course of custodial interrogation by the police was inadmissible; or it could have adopted a rule consistent with the position then advocated by this amicus that the presence of a lawyer was necessary at custodial interrogations to protect adequately a suspect's Fifth Amendment rights. See Moran v. Burbine, 475 U.S. at 426. Instead, this Court held that warnings given by the police were adequate to dispel the compulsion inherent in custodial interrogations.

The requirements of Miranda are particularly crafted to protect the privilege. Miranda requires that the suspect be warned of his rights, not only to inform the suspect so he may make an intelligent and informed choice as to whether to invoke the privilege or waive his rights,^[9] but also to "show the individual that his interrogators are prepared to recognize his privilege should he choose to exercise it." Miranda, 384 U.S. at 436; see Minnesota v. Murphy, 465 U.S. 420, 433 (1984) ("the coercion inherent in custodial interrogation derives in large measure from an interrogator's insinuations that the interrogation will continue until a confession is obtained."). Miranda also requires that once the suspect invokes his right to remain silent or to have an attorney present, the interrogation must cease. Miranda, 384 U.S. at 473-74. In this way the suspect can exercise some degree of control over the duration of the interrogation and the subjects discussed. See Michigan v. Mosley, 423 U.S. 96, 104 (1975). Finally, Miranda requires a clear waiver of rights by the suspect before a statement will be admissible. Miranda, 384 U.S. at 475-76. This requirement again protects the right of the suspect to make an affirmative choice as to whether to speak or remain silent. The burden of proving a valid waiver is on the State because the isolated circumstances of custodial interrogation are its creation, and the State itself has the "only means of making available corroborated evidence of warnings." Id. at 475.

Miranda thus represents a "carefully drawn approach." Moran v. Burbine, 475 U.S. at 427. As this Court has recognized, a "principal advantage" of the rules announced in Miranda is their "ease and clarity of application." Id. at 425 (citations omitted). Miranda's requirements have "the virtue of informing police and prosecutors with specificity as to what they may do in conducting custodial interrogation, and of informing courts under what circumstances statements obtained during such interrogation are not admissible." Fare v. Michael C., 442 U.S. 707, 718 (1979). Furthermore, in the thirty-four years since Miranda, this Court has amplified and clarified its rules to provide workable guidance on the limits of custodial interrogation. On the one hand, this Court has applied Miranda narrowly in a range of circumstances in recognition of law enforcement needs.^[10] On the other hand, Miranda has been applied more broadly to protect suspects who request counsel during a custodial interrogation. See Edwards v. Arizona, 451 U.S. 477 (1981); Minnick v. Mississippi, 498 U.S. 146 (1990).

The success of Miranda's requirements lies not only in their clarity. Miranda represents perhaps the most widely recognized constitutional safeguard in our criminal justice system.^[11] A generation of Americans has grown up since 1966 confident that if brought to the police station for questioning, we have the right to remain silent, that the police will warn us of that right and, above all, that they will respect its exercise. See Richard A. Leo, The Impact of Miranda Revisited, 86 J. Crim. L. & Criminology 621, 671-72 (1996). This is not to suggest that longstanding constitutional rules are beyond re-examination, Brown v. Board of Education, 347 U.S. 483 (1954), but only to suggest that this Court has always acted with appropriate hesitancy before overruling "watershed decision[s]" designed to protect individual rights. Planned Parenthood v. Casey, 505 U.S. 833, 867 (1992). To measure the public's reliance on Miranda may be impossible, but that does not make it irrelevant. To paraphrase Casey, "[t]he Constitution serves human values, and while the effect of reliance on [Miranda] cannot be exactly measured, neither can the certain cost of overruling [Miranda]" for people who have come to view it as an unquestioned part of our constitutional heritage. Id. at 856. Thus, even if the question were closer than we believe it is, stare decisis weighs heavily in favor of retaining Miranda.

The decision in Miranda was based in large part on what the Court concluded was the compulsion inherent in all custodial interrogations. In the more than thirty years since, studies have confirmed that conclusion. Moreover, police manuals and case law of the kind relied on in Miranda continue to reveal a police force that, while better guided in the contours of suspects' Fifth Amendment rights, remains willing to go to questionable and in many cases unlawful lengths to obtain confessions.

Central to this Court's decision in Miranda was its conclusion that the very fact of custodial interrogation involves inherent compulsion, Miranda, 384 U.S. at 467, a view shared even by Miranda's critics. See Joseph D. Grano, Selling the Idea to Tell the Truth: The Professional Interrogator and Modern Confessions Law, 84 Mich. L. Rev. 662, 674 (1986) ("The ëinherent compulsion' of custodial interrogation would be present if an untrained, uniformed officer questioned the suspect in the stationhouse receiving room."); see also Ashcraft v. Tennessee, 322 U.S. at 161 ("[t]o speak of any confessions of crime made after arrest as being ëvoluntary' or ëuncoerced' is somewhat inaccurate, although traditional") (Jackson, J., dissenting).

According to the Miranda Court, several aspects of custodial interrogation coincide effectively to compel statements from suspects against their will. Custodial interrogations are done in private, secretively, with the police holding the suspect "incommunicado" until a confession is made. Miranda, 384 U.S. at 448-50. In addition, the "patience and perseverance" of the interrogators can result in "relentless questioning." Id. at 450-55. These essential attributes of custodial interrogation remain as inherently compelling today as they were in 1966, "exact[ing] a heavy toll on individual liberty and trad[ing] on the weakness of individuals." Id. at 455.

The essence of the interrogation room is still dominance and control, exerted first through isolation. As the late Fred E. Inbau, author of the leading treatise on police interrogations, still put it twenty years after Miranda, the "principal psychological factor contributing to a successful interrogation is privacy." Fred E. Inbau, et al., Criminal Interrogation and Confessions 24 (3d ed. 1986).^[12] See Arthur S. Aubry, Jr. & Rudolph R. Caputo, Criminal Interrogation 199 (2d ed. 1972) (emphasis added) ("Privacy serves the useful purpose of helping to withdraw the environment from the suspect, and in helping create and heighten the illusion that the suspect is now, and will continue to be, completely on his own resources after he has entered the interrogation area."). Denials of guilt are simply put aside or even prevented to avoid distracting the interrogator from the objective of obtaining a confession. See, e.g., Inbau, Criminal Interrogation, supra, at 142 (recommending that interrogators prevent suspects from making denials); David E. Zulawski & Douglas E. Wicklander, Practical Aspects of Interview and Interrogation 227-42 (1993) (providing several methods for preventing suspects from expressing denials of guilt).

Several psychological studies conducted since 1966 have confirmed that the tactics still endorsed in the interrogation manuals tend to promote false or untrustworthy confessions by implanting notions of guilt or implying leniency. See Saul M. Kassin & Katherine L. Kiechel, The Social Psychology of False Confessions: Compliance, Internalization, and Confabulation, 7 Psychol. Sci. 125, 126-27 (1996); Saul M. Kassin & Karlyn McNall, Police Interrogations and Confessions: Communicating Promises by Pragmatic Implication, 15 Law & Hum. Behav. 233, 236-41 (1991); see also Richard A. Leo & Richard J. Ofshe, The Consequences of False Confessions: Deprivations of Liberty and Miscarriages of Justice in the Age of Psychological Interrogation, 88 J. Crim. L. & Criminology 429 (1998) (in a review of 60 post-Miranda confessions, 34 cases of "proven" false confessions, 18 "highly probable" false confessions, and 8 "probable" false confessions). Even drawing from the estimates of Professor Paul Cassell, in the years 1966 to 1998, "somewhere between 320 and 12,608 individuals have been wrongly convicted from false confessions alone." Richard A. Leo & Richard J. Ofshe, Using the Innocent to Scapegoat Miranda: Another Reply to Paul Cassell, 88 J. Crim. L. & Criminology 557, 562 n. 24 (1998). The tendencies toward false confessions through standard interrogation practices that these studies presentówhile alarming enough in themselvesóonly serve to emphasize the atmosphere of inherent compulsion that continues to characterize custodial interrogations today.

Perhaps the most troubling trend in the interrogation room of today is the growing police practice of ignoring Miranda altogether. In Oregon v. Hass, 420 U.S. 714 (1975), Justice Brennan warned that allowing use of unwarned statements for impeachment would send police the message that they were free to violate suspects' Constitutional rights in pursuit of confessions. Id. at 725 (Brennan, J., dissenting). The Court dismissed this suggestion as merely a "speculative possibility," id. at 723, but today, that "speculative possibility" has become a distinct reality.

Cases from thirty-eight states provide instances of police deliberately ignoring suspects' invocations of silence or right to counsel in attempts to obtain confessions. See Charles D. Weisselberg, Saving Miranda, 84 Cornell L. Rev. 109, 137-38 & n.152 (1998).^[13] In one recent empirical study, police continued to question suspects despite the suspects' invocation of their right to silence or to consult with an attorney in nearly one out of every five cases in which suspects invoked these rights. See Richard A. Leo, Inside the Interrogation Room, 86 J. Crim. L. & Criminology 266, 275-76 (1996). These cases illustrate that in custodial interrogations today, interrogators can and do intentionally violate suspects' rights to obtain confessions, even when they know the resulting confession cannot be used in the prosecutor's case in chief.

Cooper v. Dupnik, 963 F.2d 1220 (9th Cir. 1992) (en banc), provides a stark example. In that case, brought under 42 U.S.C. ß 1983, law enforcement authorities had followed a "preexisting interrogation plan" whereby they systematically "ignored Cooper's repeated requests to speak with an attorney, deliberately infringed on his Constitutional right to remain silent, and relentlessly interrogated him in an attempt to extract a confession." 963 F.2d at 1223.^[14] Following their plan of deliberately ignoring Cooper's repeated requests for counsel, the police interrogators questioned Cooper for four hours regarding, among other things, his Jewish upbringing and his sexual relations with his wife. Id. at 1229-31. "With his requests to see a lawyer disregarded, Cooper was a prisoner in a totalitarian nightmare, where the police no longer obeyed the Constitution, but instead followed their own judgment, treating suspects according to their whims." Id. at 1243. Cooper was later cleared of all suspicion in the matter. Id. at 1234.

This practice of questioning "outside Miranda," i.e., in deliberate violation of a suspect's request for counsel or invocation of silence, is far from rare. To the contrary, it has become an institutionalized feature of several major police departments. See California Attorneys for Criminal Justice v. Butts, 195 F.3d 1039 (9th Cir. 1999). The training manuals and videotapes at issue in the California Attorneys caseówhich feature a California deputy district attorney recommending that police officers deliberately ignore requests for counsel or invocation of the right to silenceóreceived statewide distribution and were followed elsewhere. See Weisselberg, supra, 84 Cornell L. Rev. at 136-37.

In a world "without Miranda," with an after-the-fact determination of voluntariness as the only guide, the state of police interrogation would be very much worse indeed. As the former chief of several major metropolitan police departments has said, "I know cops. And one of the things I know about cops is you have to control them. If you don't control them, they will go out of control . . . And Miranda is one of those controls that keeps them in line." 60 Minutes (CBS television broadcast, Dec. 5, 1999) (William Bratton, spokesperson for the National Fraternal Order of Police); See Richard A. Leo & Welsh S. White, Adapting to Miranda: Modern Interrogators' Strategies for Dealing with the Obstacles Posed by Miranda, 84 Minn. L. Rev. 397 (1999) (describing what interrogation practices might look like in a world "without Miranda.") Without Miranda, the police would lack the guidance or boundaries necessary to protect the Fifth Amendment privilege against such abuses.

Not only does Miranda work to guard against compulsion in the interrogation room, but over thirty years of accumulated evidence and the direct experience of police professionals confirms that it does so without significantly burdening law enforcement. A substantial number of police chiefs and prosecutors agree that the Miranda requirements are beneficial, not burdensome. In addition, the empirical data are clear and, with the exception of conclusions by Professor Cassell, strikingly consistent in finding no net cost to law enforcement attributable to Miranda.^[15] Before addressing empirical details, we note that the prevailing judgment about Miranda's beneficial effect is not as surprising as might at first appear. Miranda's impact must be assessed as a net result of its "costs" and its offsetting gains for law enforcement.

In hundreds of thousands of custodial interrogations each year,^[16] Miranda helps police in two ways. It makes clear what an officer must do to protect a suspect's rights and thereby to ensure that his statement will be admissible. See Michigan v. Tucker, 417 U.S. at 433. In addition, Miranda warnings tend to put suspects at ease by promoting confidence that their right to cut off questioning will protect them from having to make incriminating statements. Studies consistently show that because of Miranda, and public awareness that courts will enforce it, suspects tend to be less wary, let down their guard and then make damaging admissions.^[17]Interrogators have learned that Miranda often produces confessions that they would not otherwise obtain.^[18]

Moreover, to focus only on elusive possibilities of change in the number of confessions obscures the more important and universally acknowledged qualitative changes attributable to Miranda.^[19] These include not only enhanced protection for suspects and enhanced legitimacy for police but also enhanced reliability of the confessions themselves. Miranda has altered interrogation techniques, discouraged tacit intimidation, and forced interrogators to work from a suspect's overconfidence and misplaced trust.^[20] There can be little doubt that confessions elicited in this manner are more reliable than those prompted by fear.

These offsetting tendencies help explain why Miranda's requirements cause little or no net loss of confessions.

Law enforcement officials widely share the perception that Miranda does not hinder, and in fact benefits, the police. Though many officials would prefer to escape Miranda's constraints, many others acknowledge that Miranda is not burdensome. Justice Clark, a former Attorney General, and Justice White, a former Deputy Attorney General, both dissented in Miranda but later supported the decision,^[21] as did Chief Justice Burger. See Rhode Island v. Innis, 446 U.S. at 304 (Burger, C.J., concurring in the judgment). Leading police chiefs and prosecutors who initially opposed Miranda likewise changed their position,^[22] and wider surveys confirm law enforcement support for Miranda.^[23]

When a division of the Department of Justice suggested challenging Miranda in 1986, Solicitor General Charles Fried noted that "[m]ost experienced federal prosecutors in and out of my office were opposed to this project, as was I." Charles Fried, Order and Law 46 (1990). A survey conducted for the Department revealed similar sentiment among police executives,^[24] and other reports at the time showed emphatic support for Miranda among police chiefs and line officers alike.^[25] This view remains widely held. William Bratton, who has headed five police departments including New York's, states that Miranda "works fine" and is "an effective control on abuses." Roger Parloff, Miranda on the Hot Seat, New York Times Magazine, Sept. 26, 1999, pp. 84, 87.

Police support for Miranda suggests that many chiefs would impose similar requirements even if Miranda were overruled. Nonetheless, police officials widely endorse federal enforcement, in order to insure that warning and waiver requirements remain effective.^[26] Moreover, a decision ending Miranda's constitutional status in our national culture would undermine suspects' confidence that their rights will be respected in interrogations. Respect for Miranda's requirements "is sustained in no small part by the existence of [federal] review." Withrow, 507 U.S. at 695.

Notwithstanding this background of widespread support by the law enforcement community, the existing state of custodial interrogation and the significant number of officials urging this Court to affirm the result below leave little doubt that freedom to depart from Miranda would not go unused. The resulting local variations would quickly erode public confidence that clear-cut safeguards still apply. Even in communities that adopted Miranda-like rules voluntarily, the absence of a constitutional imprimatur would undercut the public awareness, legitimacy and trust that Miranda's nationwide acceptance now confers. Leo, The Impact of Miranda, supra, 86 J. Crim. L. & Criminology at 679-80.

In the wake of Miranda, many studies attempted to assess its effect. Though crude, the studies generally pointed in the same directionówith isolated exceptions, Miranda's initial impact on confession and conviction rates appeared small or non-existent.^[27] Two 1996 studies provide contemporary data. In the first, covering three California cities, police obtained incriminating statements in 64% of their interrogations.^[28] In Salt Lake City, incriminating statements were obtained in 54% of the interrogations.^[29]Both rates were well within the range reported in various cities prior to Miranda.^[30] In a 1998 report on Washington, D.C., the interrogation success rate, though not quantified, was so high that Miranda's impact was described as "approximately zilch."^[31]

Scholars are virtually unanimous in noting the absence of empirical evidence that Miranda impedes law enforcement.^[32]And nearly all scholars concur in dismissing the contrary claims of Professor Cassell, whose arguments we now consider.

In his effort to demonstrate that Miranda impedes law enforcement, Professor Cassell has used four kinds of data: before-and-after studies of Miranda's initial impact, data from contemporary interrogation in Salt Lake City, the trend of clearance rates for four years after Miranda, and a regression analysis of clearance rates over a more extended period. Several of Cassell's studies have drawn unusually sharp criticism for their failure to meet standard scholarly norms. See, e.g., Leo, Using the Innocent to Scapegoat Miranda, supra, 88 J. Crim. L. & Criminology at 557 n.2 (1998). Each study by Cassell suffers from major defects that render its claims untenable.

Cassell's first set of data examined eleven prior studies comparing confession rates immediately before and after Miranda. He excluded three studies that he deemed flawed; in the remainder, confessions were said to drop on average by 16%, and convictions by an estimated 3.8%. Cassell, Miranda's Social Costs, supra, 90 Nw. U. L. Rev. at 437-38. As these were crude studies from a period before police had adapted to Miranda, the claimed 3.8% decline in conviction rates is not necessarily alarming. Yet examination of Cassell's procedures showed that inconsistent methods and misreading of sources had significantly inflated the claimed drop in confessions,^[33] and had even inflated the case attrition rate to the 3.8% level; on reanalysis, the properly adjusted case attrition rate was less than 1%,^[34] with reason to think the net impact would be even less today.^[35]

Cassell's second study examined contemporary interrogation in Salt Lake City. Cassell reported that only 42% of interrogated suspects gave incriminating statements, compared to pre-Miranda figures for other cities that he put in the 55-60% range.^[36]Reanalysis of the data showed that Cassell had excluded confessions he deemed "volunteered" and useful statements in the form of incriminating "denials with explanation." When these were included, as any valid comparison to pre-Miranda figures requires, the adjusted confession rate (54%) was indistinguishable from Cassell's pre-Miranda comparison figure (55-60%). Thomas, Plain Talk, supra, 43 UCLA L. Rev. at 935-36, 946-52, 953. In earlier work, moreover, Cassell reported that pre-Miranda confession rates varied over a much wider range (31%-88%), with eleven of the twenty cities below 50%. See id. at 936; Cassell, Miranda's Social Costs, supra, 90 Nw. U. L. Rev. at 459. Cassell's own figure for Salt Lake City (42%) falls well within the wider pre-Miranda range, and the properly adjusted figure (54%) is above the median of the pre-Miranda reports. The Salt Lake City data offer no reason to think that interrogation is less successful today than it was before Miranda.

Cassell's third set of data shows that clearance rates (the percentage of reported crime that police solve by arrest) remained at about 60% from 1950-1965, but dropped steadily during the four years following Miranda, and then leveled off at about 45% thereafter. Cassell suggests that no other event, only Miranda, could explain this mid-1960's drop in clearance rates. Paul G. Cassell, All Benefits, No Costs: The Grand Illusion of Miranda's Defenders, 90 Nw. U. L. Rev. 1084, 1089-91 (1996). But when the clearance-rate trend is compared to the trend in police clearance capacity (the number of officers available relative to the number of crimes to be solved), the two trends closely match.^[37]Crime rates had been skyrocketing since 1960, largely for social and demographic reasons. See James Q. Wilson, Thinking About Crime 4-20 (1975). Because police resources failed to keep up, the number of officers per 100 violent crimes fell sharply, from 115 in 1960 and 82 in 1965 to only 45 in 1970. See, Schulhofer, Clearance Rates, supra, 91 Nw. U. L. Rev. at 284, 285 n.16. Given the collapse of investigative capacity during this period, the predictable drop in clearance rates provides no reason to think that Miranda made police less efficient. In fact, the number of crimes cleared per officer steadily rose during this period. Id. at 286-87.

Cassell's fourth study used regression analysis to test whether clearance rates were higher before Miranda (1950-65) than after (1966-94). Controlling for all quantifiable variables, the study asserted that Miranda reduced clearance rates for property crimes, robbery and total violent crimes, but not for murder, rape and assault. Paul G. Cassell & Richard Fowles, Handcuffing the Cops? A Thirty-Year Perspective on Miranda's Harmful Effects on Law Enforcement, 50 Stan L. Rev. 1055, 1086-88 (1998).^[38] A reanalysis with different statistical models confirmed only some of Cassell's results.^[39]

Whichever the preferred model, there is no basis for Cassell's speculative claim that Miranda was the likely cause of any statistical differences between the 1950-65 and 1966-94 periods. A host of other factors could explain the results, including defects in the data and the many nonquantifiable factors that differed in the two periods.^[40] Potential causes not controlled in his study include social changes that make crime harder to solve (e.g., the increasing proportion of stranger crimes), legal changes unrelated to Miranda (e.g., implementation of the right to counsel, more widespread compliance with the Fourth Amendment), and changes in police organization, including more objective crime-recording practices. Given vast changes in society and law enforcement between 1950-65 and 1966-94, there is no reason to assume that Miranda must have caused the clearance-rate drop or any part of it.^[41]

Cassell argues that such factors could not have been important because none of them changed dramatically right after Miranda, as clearance rates did. Cassell & Fowles, Handcuffing the Cops, supra, 50 Stan. L. Rev. at 1108, 1113, 1117. There are two large flaws in this response. First, major record-keeping improvements were introduced in the late 1960s, so their principal effect could well have occurred in that period. Donohue, Did Miranda Diminish, supra, 50 Stan L. Rev. at 1172 n.100. Second, as discussed above, a precipitous decline in clearance capacity also occurred in the late 1960s, in a pattern that fits the clearance-rate drop. Cassell's regression analysis controls for this factor, but only by measuring an average clearance-rate effect over the entire 1966-94 period. The differences identified by that analysis could have been caused by any of the many social and legal factors that were different on average during the 1966-94 period than they were before, even if those factors did not change abruptly right after 1966.^[42]On examination, the regression analysis, like the other Cassell studies, provides no evidence that Miranda, on balance, caused a net harmful effect.

At bottom, of course, any potential "cost" to law enforcement due to enforcement of the Fifth Amendment through Miranda is beside the point; the Fifth Amendment was unquestionably intended to restrain law enforcement. Given the reliance by counsel supporting the decision below on such claims of law enforcement "costs," however, this Court should have an accurate view of what those costs are. There is no reason to question the widely held view that Miranda does not significantly burden law enforcement.

^[1] All parties have consented to the filing of this brief; letters of consent have been lodged with the Clerk. No counsel for a party authored this brief in whole or in part, and no person other than the amicus, its members, and its counsel made a monetary contribution to the preparation or submission of the brief.

^[2] See Murphy v. Waterfront Comm'n, 378 U.S. 52, 55 (1964) (privilege against self-incrimination reflects "our unwillingness to subject those suspected of crime to the cruel trilemma of self-accusation, perjury or contempt; [and] our preference for an accusatorial rather than an inquisitorial system of criminal justice") (citations and internal quotations omitted); Watts v. Indiana, 338 U.S. 49, 55 (1949) ("Protracted, systematic and uncontrolled subjection of an accused to interrogation by the police for the purpose of eliciting disclosures or confessions is subversive of the accusatorial system.").

^[3] Despite this rule, the Courtópursuant to its "totality" caselawófrequently allowed admission of confessions where there was substantial evidence of physical violence, refusing to consider that evidence unless it was undisputed by the police. See, e.g., Stein, 346 U.S. at 169-70; Crooker v. California, 357 U.S. 433, 437-38 n.3 (1958); Stroble v. California, 343 U.S. 181, 185-86 (1952).

^[4] See, e.g., Paul G. Cassell, The Costs of the Miranda Mandate: A Lesson in the Dangers of Inflexible, "Prophylactic" Supreme Court Inventions, 28 Ariz. ST. L.J. 299, 302 (1996).

^[5] The distinction between compelled and coerced statements is fundamental to our system of civil and criminal justice. While coerced statements are never admissible for any purpose, Mincey v. Arizona, 437 U.S. 385 (1978), witnesses are routinely compelled to give statements under subpoena. Absent a valid claim of privilege, such compelled statements are clearly admissible, Kastigar v. United States, 406 U.S. 441 (1972), and are the bread and butter of virtually all judicial proceedings. United States v. Nixon, 418 U.S. 683, 709 (1974) ("it is imperative to the function of courts that compulsory process be available"). Indeed, the Sixth Amendment provides the accused with the right to compulsory process to call witnesses in his favor; again, those witnesses are compelled to testify absent an applicable privilege. See Washington v. Texas, 388 U.S. 14 (1967). The only such privilege guaranteed by the Constitution is that against compelled self-incrimination, for that privilege goes to the heart of our assurance that people are innocent until a prosecutor proves them guilty.

^[6] Miranda's requirement that police take affirmative steps to dispel inherent compulsion, by reciting warnings, finds a close parallel in the Fifth Amendment requirement that a trial judge take affirmative steps to warn the jury not to draw an adverse inference from a defendant's failure to testify. This Court noted in Carter v. Kentucky that "if left to roam at large with only its untutored instincts to guide it," an unwarned jury might draw inferences that would jeopardize the Fifth Amendment privilege. Carter, 450 U.S. at 301. But, the Court continued, "[a] trial judge has a powerful tool at his disposal to protect the constitutional privilege ’ the jury instruction ’ and he has an affirmative obligation to use that tool." Id. at 303 (emphasis added). Carter's affirmative warning requirement, triggered in the trial setting by a defense counsel's request, is both per se and explicitly prophylactic, much like Miranda's affirmative warning requirement triggered by custodial interrogation.

^[7] To be sure, Miranda recognized that Congress or the States might adopt alternative means of protecting the Fifth Amendment privilege, but made clear that such alternative measures must be "at least as effective [as Miranda] in apprising accused persons of their right of silence and in assuring a continuous opportunity to exercise it." Miranda, 384 U.S. at 467. Section 3501 does not require that any warnings be given and merely returns protection of the Fifth Amendment privilege to the same retrospective "totality" analysis rejected by Miranda. Cf. Smith v. Robbins, No. 98-1037, slip op. at 19 (Jan. 19, 2000) (California procedure for appointed counsel to file brief on behalf of indigent upheld where procedure, while not identical to prophylactic requirements of Anders v. California, 386 U.S. 738 (1967), "far exceed[s] those procedures that we have found invalid").

^[8] Similarly, in the context of Sixth Amendment right-to-counsel and confrontation claims, the ß 3501 mandate requiring federal courts to admit all "voluntary" statements would, if upheld, require overruling such cases as United States v. Henry, 447 U.S. 264 (1980), Bruton v. United States, 391 U.S. 123 (1968), and Massiah v. United States, 377 U.S. 201 (1964), and by implication such parallel state cases as Gray v. Maryland, 523 U.S. 185 (1998), Cruz v. New York, 481 U.S. 186 (1987), and Brewer v. Williams, 430 U.S. 387 (1977).

^[9] Requiring warnings before the Fifth Amendment privilege may be waived is necessary because the privilege protects the basic right to a trial, as well as the reliability of the truth-finding process. See Withrow, 507 U.S. at 691. Accordingly, because the Fifth Amendment privilege against self-incrimination "is a fundamental trial right," it "operates in a different manner" than the Fourth Amendment. United States v. Verdugo-Urquidez, 494 U.S. 259, 264 (1990); accord Schneckloth v. Bustamonte, 412 U.S. 218, 232, 240 (1973).

^[10] See, e.g., Michigan v. Tucker, 417 U.S. 433 (1974); Oregon v. Hass, 420 U.S. 714 (1975); New York v. Quarles, 467 U.S. 649 (1984); Oregon v. Elstad, 470 U.S. 298 (1985).

^[11] As more fully explained in the briefs for petitioner and the United States, there can be no doubt that the Miranda Court's requirement of minimum warnings in custodial interrogations was based on constitutional grounds. All nine Justices participating in Miranda understood that the Court's ruling stemmed from the requirements of the Constitution. See Miranda, 384 U.S. at 490-91 ("the issues presented are of constitutional dimensions . . . . Where rights secured by the Constitution are involved, there can be no rule making or legislation which would abrogate them." ) (emphasis added); id. at 500 (Clark, J., concurring in part and dissenting in part) (emphasis added) (observing that the Court had "fashion[ed] a constitutional rule"); id. at 505 (Harlan, J., dissenting); id. at 532 (White, J., dissenting). In addition, the application of Miranda to the Statesóboth on direct appeal and on habeas reviewódemonstrates that this Court has understood its warning requirements to be of constitutional stature. See id. at 491-94, 497-99; Withrow, 507 U.S. 680. The tainted fruits holdings of Oregon v. Elstad, 470 U.S. 298, and Michigan v. Tucker, 417 U.S. 433, and the public safety exception of New York v. Quarles, 467 U.S. 649, while imposing limits on the scope of Miranda's applicability, cannot support the sweeping conclusion that Miranda's warning requirement was without constitutional foundation.

^[12] See Miranda, 384 U.S. at 449 (quoting this language from the first edition of Inbau's treatise). Until his death in 1998, Inbau continued to enjoy the reputation as the foremost authority on police interrogation techniques. See, e.g., Gisli H. Gudjonsson, The Psychology of Interrogations, Confessions and Testimony 31 (1992).

^[13] Not only do police routinely ignore requests for counsel or invocations of silence, but courts also routinely admit the confessions so obtained as "voluntary" under the traditional due process test. See, e.g., People v. Peevy, 953 P.2d 1212, 1219 (Cal.), cert. denied, 119 S. Ct. 595 (1998); People v. Winsett, 606 N.E.2d 1186 (Ill. 1992), cert. denied, 510 U.S. 831 (1993); People v. Bradford, 929 P.2d 544 (Cal. 1997); State v. Favero, 331 N.W.2d 259, 261, 263 (Neb. 1983).

^[14] The police agencies in question had assembled a "Task Force" to find a suspect. The Task Force's plan was "to hold the suspect incommunicado, and to pressure and interrogate him until he confessed," including lying to the suspect about the evidence against him and ignoring any request for counsel or right to silence:

Although the officers knew any confession thus generated would not be admissible in evidence in a prosecutor's case in chief, they hoped it would be admissible for purposes of impeachment if the suspect ever went to trial. They expected that the confession would prevent the suspect from testifying he was innocent, and that it would hinder any possible insanity defense.

Cooper, 963 F.2d at 1224.

^[15] Even Professor Cassell has conceded that "warnings [the only Miranda requirement allegedly violated in this case] can be given to suspects without significantly harming the confession rate." Paul G. Cassell, Miranda's Social Costs: An Empirical Reassessment, 90 Nw. U. L. Rev. 387, 494 (1996).

^[16] Police annually make over 750,000 arrests for violent crime and over two million arrests for property crime. Id. at 440. Roughly 80% of these arrestees are questioned. Paul G. Cassell & Bret S. Hayman, Police Interrogation in the 1990s: An Empirical Study of the Effects of Miranda, 43 UCLA L. Rev. 839, 854 (1996).

^[17] See David Simon, Homicide 197-202 (1991); Richard A. Leo, From Coercion to Deception: The Changing Nature of Police Interrogation in America, 18 Crime, L. & Soc. Change 35 (1992); Peter Carlson, You Have the Right to Remain Silent, Washington Post, Sept. 13, 1998, at 6.

^[18] Carlson, supra at 11, 19; Richard A. Leo, Miranda's Revenge: Police Interrogation as a Confidence Game, 30 L. & Soc'y Rev. 259 (1996); cf. George C. Thomas, III, Is Miranda a Real-World Failure?: A Plea for More (and better) Empirical Evidence, 43 UCLA L. Rev. 821, 831 (1996).

^[19] In fact, even authors otherwise opposed to Miranda have praised it for helping to "dignify" police practices and to improve the credibility of confessions obtained by interrogation. See, e.g., Paul B. Weston & Kenneth M. Wells, Criminal Investigation: Basic Perspectives, 184 (2d ed. 1974). At least one commentator has warned that removing this bar would send a dangerous message to the law enforcement community. See Leo, The Impact of Miranda, supra, 86 J. Crim. L. & Criminology at 680.

^[20] See, e.g., Carlson, supra; Leo, From Coercion to Deception, supra; Leo, Miranda's Revenge, supra.

^[21] Justice Clark reported in 1968 that confessions were "still going stronger than ever," and stated, "I confess error in my appraisal of [Miranda's] effect." Tom C. Clark, Criminal Justice in America, 46 Tex. L. Rev. 742, 744-45 (1968). Justice White acknowledged "the coercion inherent in custodial interrogation," Minnesota v. Murphy, 465 U.S. at 433, and joined this Court in holding that Miranda claims must remain enforceable in the context of habeas corpus. Withrow, 507 U.S. 680.

^[22] Pittsburgh's Chief of Police "became convinced that his officers were doing better work as a result of Miranda." See Yale Kamisar, Overturning Miranda, (Detroit) Legal Advertiser, Feb. 26, 1987, at 4. Los Angeles District Attorney Evelle Younger and Philadelphia District Attorney Arlen Specter, both prominent critics of Miranda, later endorsed it, with Specter stating that "law enforcement has become accommodated to Miranda, and therefore I see no reason to turn the clock back." Quoted in Yale Kamisar, Landmark Ruling's Had No Detrimental Effect, Boston Sunday Globe, Feb. 1, 1987, at A27; Evelle Younger, Interrogation of Criminal Defendants, 35 Fordham L. Rev. 255 (1966).

^[23] A 1976 survey of 67 police chiefs, sheriffs, and state police commanders also found an "overwhelming" number (more than 90% in one state) favorable to Miranda. Stephen L. Wasby, Small Town Police and the Supreme Court (1976). A 1988 American Bar Association survey of more than 800 prosecutors, judges, and police officers found that "[a] very strong majority . . . agree that compliance with Miranda does not present serious problems for law enforcement." ABA Special Comm'n on Criminal Justice in a Free Society, Criminal Justice in Crisis 28 (1988).

^[24] See Stephen J. Schulhofer, Miranda's Practical Effect: Substantial Benefits and Vanishingly Small Social Costs, 90 Nw. U. L. Rev. 500, 507 n.23 (1996) (76% of those polled opposed any modification of Miranda or "prefer[red] modification to reversal;" only 24% supported overruling of Miranda).

^[25] Tom Gibbons & Jim Casey, Ed Meese's War on Miranda Draws Scant Support, Chi. Sun-Times, Feb. 17, 1987, at 41; Eduardo Paz-Martinez, Police Chiefs Defend Miranda Decision Against Meese Threats, Boston Globe, Feb. 5, 1987, at 25-29. The Executive Director of the International Association of Chiefs of Police, one of the few police leaders to voice support for reconsideration at that time, stated that Miranda "could be set aside, provided there are appropriate guidelines that deal very stringently with police officers that abuse them." Pete Yost, Police Executive Backs Miranda Ruling Review, Washington Post, Jan. 23, 1987, at A17.

^[26] One chief warned that even if communities imposed similar requirements on their own, individual officers would "probably exercise their own judgment" and "then you're going to have a mixture of a lot of things happening." Paz-Martinez, Police Chiefs, supra, at 29.

^[27] See Liva Baker, Miranda: Crime, Law and Politics, 180-81, 403-05 (1983); Leo, The Impact of Miranda, supra, 86 J. Crim. L. & Criminology at 632-45. In the cautious language of social science, "[t]here is insufficient evidence that Miranda has a statistically significant effect on the percentage of suspects who incriminate themselves." Thomas, Is Miranda a Real-World Failure?, supra, 43 UCLA L. Rev. at 830-31.

^[28] Leo, Inside the Interrogation Room, supra, 86 J. Crim. L. & Criminology at 280.

^[29] See George C. Thomas, III, Plain Talk About the Miranda Empirical Debate; A "Steady-State" Theory of Confessions, 43 UCLA L. Rev. 933, 935-36, 952 (1996).

^[30] Id., at 953, 955; Schulhofer, Miranda's Practical Effects, supra, 90 Nw. U. L. Rev. at 509 n.28.

^[32] The principal scholarly disagreement is on the subtle question whether the data affirmatively show that Miranda causes no harm or merely fail to show that Miranda does cause harm. And even the more cautious scholars concede that an absence of evidence is telling: "Reaching any decision on the status of Miranda without better empirical evidence is roughly like having to decide whether you need surgery before knowing whether you are sick." Thomas, Plain Talk, supra, 43 UCLA L. Rev. at 958.

^[33] Schulhofer, Miranda's Practical Effect, supra, 90 Nw. U. L. Rev. at 502, 538-39. For illustration, we mention one of the many flaws in this study. Cassell included in his averages substantial confession-rate drops reported in Brooklyn and New Orleans, though in both cases the confession rate in the "before" period was not determined from an actual count but solely from the subjective estimate of an official critical of Miranda. Id. at 531-33. Yet, inconsistently, Cassell excluded a Los Angeles study finding that confessions had increased by 10%, doing so on the basis of strained speculation that there might have been a difference in the kinds of confessions counted before and after. Id. If such a discrepancy might have affected the Los Angeles study, an actual count, there is at least equal reason for concern about inaccuracy in the before-period "guestimates" that Cassell deemed reliable in the Brooklyn and New Orleans cases. Other similarly serious flaws further skew the conclusions of Cassell's before-and-after study. See, e.g., id. at 516-30.

^[34] Id. at 539, 544-47. Among the usable studies, change in confession rates varied widely, from a 16% decrease to a 10% increase. Id. at 539. Such diverse results suggest that Miranda itself probably did not cause the fluctuations.

^[36] Cassell & Hayman, Police Interrogation, supra, 43 UCLA L. Rev. at 871-72.

^[37] See Stephen J. Schulhofer, Miranda and Clearance Rates, 91 Nw. U. L. Rev. 278, 283-84 (1996).

^[38] It should be noted that the variable labeled "Miranda" throughout this article is merely a before-after "dummy" variable centered on 1966, not necessarily a measure of Miranda's causal effect. See John J. Donohue, III, Did Miranda Diminish Police Effectiveness?, 50 Stan. L. Rev. 1147, 1161, 1163, 1172 (1998).

^[39] See, Donohue, Did Miranda Diminish, supra, 50 Stan. L. Rev. at 1168-70, finding a significant difference for larceny but not for other property crimes, and for total violent crimes, but not for the individual crimes that comprise this category (murder, rape, robbery and assault).

^[41] One particularly important factor makes clear why it is so speculative to attribute clearance-rate differences to Miranda. It is well-established that during the 1950-65 period, many police departments artificially inflated their clearance rates, for both internal and public-relations purposes. See Jerome Skolnick Justice Without Trial, 164-81 (1966). More professional record-keeping standards introduced in the late 1960s unquestionably reduced reported clearance rates, not because fewer crimes were solved but because bogus clearance claims common before were no longer included. Donohue, Did Miranda Diminish, supra, 50 Stan. L. Rev. at 1152-55, 1172.

^[42] Id. at 1172 n.100. Professor Donohue concludes that "one is left with unbridgeable uncertainty about how much confidence to repose in any of the statistical results" and that even if there were a statistical change in clearance rates, "causal attribution to Miranda is difficult in light of the vast systemic changes both in crime and criminal justice that were occurring and have persisted since the mid-1960s." Id. at 1172.

Copyright 2000, The American Civil Liberties Union