|
Using the innocent to scapegoat Miranda: another reply to Paul Cassell
by Richard A. Leo , Richard J. Ofshe I. INTRODUCTION In Protecting the Innocent from False Confessions and Lost Confessions--And From Miranda,(1) Paul Cassell advances several logically flawed and empirically erroneous propositions. These propositions appear to stem from Cassell's ideological commitments,(2) and they continue his seemingly empirical and virtually single-handed assault(3) on the Warren Court's ruling in Miranda v. Arizona.(4) Cassell's readers are accustomed to a steady stream of speculative accusations that Miranda causes tens of thousands of guilty suspects to escape conviction every year.(5) This time, however, Cassell attacks the constitutionally--based warnings on the grounds that they harm innocent suspects.(6) Not only is Cassell's thesis that "Miranda affirmatively harms the innocent"(7) unsupported by any evidence, it also flies in the face of reason. Protecting the Innocent advances three points in particular. First, Cassell suggests an "alternative" methodology for estimating the annual frequency of wrongful convictions arising from false confessions.(8) This proposal rests on empirically untenable assumptions, and it ignores the formidable methodological barriers to estimating the harms of improper police interrogation methods. Ultimately, Cassell's method for quantifying the frequency of wrongful convictions following from false confessions amounts to no more than grand speculation masquerading as a reasoned estimate of fact. Hence, it has no credible empirical foundation, and it should not be used as the basis for any factually-informed analysis. Second, Cassell claims that Miranda warning and waiver requirements "present more risks to the innocent [false confessor] than they would prevent."(9) Supposedly, Miranda harms the innocent because it inhibits police from gaining confessions from truly guilty suspects that would therefore exonerate innocents who have been wrongfully convicted. As with Cassell's first position, this claim too has no empirical foundation and should not be the basis for any policy-making jurisprudence. Finally, Cassell speculates that substituting the videotaping of police interrogation for the Miranda rules would "produce tens of thousands of truthful confessions that would help protect the innocent."(10) Although space limitations prevent us from fully discussing this claim, Cassell's proposal to scale back the Miranda warnings--and thereby undermine custodial suspects' abilities to resist the pressures of interrogation--might actually increase the number of false confessions police elicit from innocent suspects.(11) In this commentary, we focus primarily on Cassell's first two claims.(12) In Part II, we briefly explain why neither social scientists nor legal scholars presently possess the requisite information to make empirically sound quantitative estimates of the annual frequency of false confessions or the number of wrongful convictions they cause.(13) In Part III, we address Cassell's implausible claim that Miranda harms the innocent more than it helps them. Finally, in Part IV, we argue that Cassell's essay represents less of an effort to seriously explore optimal ways of protecting the innocent than a continuation of his scapegoating of Miranda for many of the real and imagined problems of the American criminal justice system.(14) Cassell's claim that Miranda harms the innocent not only strains credulity, but also is an unhelpful diversion from much needed reforms that would improve police investigation practices and protect the innocent from wrongful conviction. II. ESTIMATING THE FREQUENCY OF FALSE CONFESSIONS AND WRONGFUL CONVICTIONS(15) A. IS QUANTIFICATION POSSIBLE? IS IT NECESSARY? IS A HOUSEOF-CARDS REALLY BETTER THAN A HUMBLE "I DON'T KNOW?." To the extent that Cassell suggests that he is providing an "alternative" methodology to our procedures for quantifying the frequency of false confessions his analysis is misleading.(16) We have never suggested or advocated any particular methodology for estimating the frequency of false confessions. To the contrary, we have repeatedly pointed out that the methodological problems inherent in arriving at a sound estimate are formidable and unsolved, and we have concluded that no well-founded estimate has yet been published.(17) Although Cassell is well aware of our analysis of the barriers to estimating false confession rates, he suggests no solutions to these fundamental problems. Instead, he uses our paper (in this volume of The Journal of Criminal Law and Criminology) as another opportunity to engage in Miranda-bashing.(18) There are at least three major reasons why it is not presently feasible to estimate validly the incidence of police-induced confessions or the number of wrongful convictions they cause. First, American police typically do not record interrogations in their entirety. Therefore, it is not possible to ascertain the ground troth of the interrogation (i.e., what really happened) or the validit of certainty. Second, because no criminal .justice agency keeps records or collects statistics on the number or frequency of interrogations in America, no one knows how often suspects are interrogated or how often they confess, whether truthfully or falsely. Third, many, if not most, cases of false confession are likely to go entirely unreported by the media and therefore unacknowledged and unnoticed by researchers. For these reasons, it is not possible to reach a sound estimate of the incidence of false confession, or to estimate how often false confessions lead to wrongful convictions. Yet Cassell criticizes the false confession literature for failing to provide "a ballpark estimate of the frequency of false confessions,"(19) as if empirical researchers somehow bear this burden. However, one might view the absence of any such estimates as resulting from most researchers preference for an honest "I don't know" to the use of guesswork to arrive at specious estimates of real world facts. Until it becomes possible to draw a random sample of confession cases from a definable universe and accurately determine both the ground truth of the interrogation and the validity of the confession statement in each case, it will not be possible to arrive at a methodologically acceptable estimate of the annual frequency of false confessions or the number of wrongful convictions they cause. We reject not only Cassell's assertion that reasonable quantification is presently possible, but also his insistence that this is somehow necessary to make considered public policy decisions about the regulation of interrogation methods.(20) It is well established that psychologically-induced false confessions occur frequently enough to warrant the concern of criminal justice officials, legislators and the general public.(21) Whether or not Cassell wishes to acknowledge this(22)--and regardless of the advocacy numbers(23) he conjures up to speculate about the importance of other issues--police-induced false confessions, and the miscarriages of justice they spawn, remain serious policy problems in the American criminal justice system.(24) The project of quantification is Cassell's, not ours or any other researchers studying interrogation and false confessions. If Cassell were able to solve the methodological problems we have identified, or if he were able to explain why our methodological concerns are misplaced, then we would welcome quantification. If, however, Cassell's quantification scheme is merely a rhetorical device to permit him to argue for the superiority of his policy preferences, we must necessarily reject his conclusions as ideologically driven and the products of a commitment to the advocacy of a particular value position rather than to the empirical ascertainment of truth. Whenever Cassell attempts to give the appearance of quantitatively demonstrating the supposedly tragic consequences of Miranda, he uses a simple and formulaic approach. First, Cassell argues that the existing literature emphasizes one side of the problem(25) and thus a narrow set of policy solutions.(26) He then proposes that there are really two sides to the problem and, in an age of finite resources, policy decisions must be made in light of competing trade-offs.(27) Next, Cassell proclaims that researchers must quantify the dimensions of the problem in order to assess its importance and make informed policy judgments about competing tradeoffs.(28) Cassell then criticizes his opponents for failing to adhere to his standard.(29) He then conjures up severely flawed or simplistic estimates of the effects of the policy alternatives through a series of arcane numerical manipulations that supposedly measure the costs and benefits of competing policy proposals on a common metric.(30) Finally, Cassell concludes that the social costs of Miranda far outweigh its benefits, and that it should therefore be overruled.(31) Though Miranda relies on a series of highly problematic assumptions and extrapolations, Cassell's quantitative analysis permits him rhetorically to claim to be able to compare otherwise incommensurable policy trade-offs and then to declare a clear winner--the abolition of Miranda. The problem with Cassell's impulse to quantification, however, is that it oversimplifies complicated issues and inevitably presents speculation as fact in the service of a larger ideological agenda: abolishing Miranda and promoting of state power in the criminal justice system. While Cassell attempts to preempt this criticism by approvingly quoting Professor Schulhofer's statement that the "`size of a legal problem does matter, and we cannot avoid thinking about it, or rely only on our intuitions, just because a perfect study has yet to be done,'"(32) Cassell fails to mention that elsewhere Schulhofer characterizes Cassell's particular extrapolations as "simply rhetoric," and "not a serious foundation for assessing social policy."(33) B. CASE EXAMPLES OF FALSE CONFESSIONS Prior to explicating his estimation methodology, Cassell arbitrarily eliminates from consideration individuals who have been wrongfully arrested and/or wrongfully prosecuted based on false confessions, but not wrongfully convicted.(34) However, the harms that the criminal justice system inflicts on false confessors are not limited to wrongful incarceration post-conviction, but also include wrongful (and sometimes lengthy) pre-trial deprivation of liberty, the stigma associated with criminal charges, the irrevocable loss of reputation, the stresses of standing trial and the sometimes bankrupting financial burdens of defending oneself in costly and drawn out proceedings against the state. We believe it is therefore appropriate to view the amount of harm that the system inflicts on innocent individuals, not merely the fact of their conviction, as the proper measure of impact in the study of miscarriages of justice. The amount of harm can thus be treated as a continuous variable, and its victims can be indexed along a continuum ranging from wrongful arrest, prosecution and pre-trial deprivation of liberty to wrongful conviction and incarceration to wrongful execution. The end category (the execution of the innocent), not the middle one that Cassell selects (wrongful conviction), is "the ultimate miscarriage of justice."(35) The decision to focus only on those cases that result in wrongful conviction and post-trial incarceration arbitrarily minimizes all of Cassell's estimates of the harms of police-induced false confessions. For example, this decision reduces Cassell's focus from the sixty cases we report to only those twenty-nine that resulted in wrongful conviction. Cassell then suggests that these twenty-nine cases represent only "a few drops in this very large bucket"(36) because, according to his extrapolations, approximately 368,000 homicide interrogations occurred during the time period of these twenty-nine cases (from 1973 to 1996).(37) However, Cassell's decision to treat either our sixty case examples or the twenty-nine convictions as though they constitute what we believe to be the entire relevant population of false confessions resulting in wrongful convictions from 1973 to 1996 is both fallacious and ideology-serving. We were able to investigate only a small fraction of the disputed police interrogations that occurred in this twenty-three year interval. Cassell's implication that only twenty-nine wrongful convictions from false confessions occurred during this time period is therefore misleading,(38) As we made clear in our article, our descriptive statistics summarize variation in the case outcomes of the set of false confessions we studied.(39) We have no idea what proportion of the false confessions occurring during this twenty-three year period we have discovered. Cassell attempts to reduce the significance of the false confession problem further by mis-reviewing the import of previous studies of police interrogation in America and England and concluding that they represent a "dry well for false confessions."(40) For example, it is misleading to hold out Richard Leo's doctoral dissertation to support the position that false confession rarely happens since Leo made no effort to study this issue in that research. The purpose of Leo's observational research was to study routine interrogation practices, not false confessions.(41) Accordingly, Leo never attempted to determine whether any of the 182 interrogations that he observed led to false confessions. As with Leo's dissertation research, the purpose of the well-known 1967 Yale Study also was to observe routine interrogation practices and evaluate the impact of Miranda in a single jurisdiction, not to discover instances of false confession. The fact that neither these post-Miranda impact studies, nor the others that Cassell cites,(42) study the phenomenon of false confessions simply cannot be construed as support for Cassell's assumption that false confessions are rare. Cassell's conclusion that "[t]he frequency of false confessions in America appears to be so low as to have escaped detection in the various samples that have been drawn"(43) is therefore unjustified. In his effort to negate the claim that wrongful convictions from false confessions are common, Cassell also mentions Dr. Gisli Gudjonsson, who is widely recognized as one of the leading authorities on false confessions(44) and whose primary research in the last decade and a half has focused on interrogation and confession in England and Iceland.(45) However, it is surprising to cite Gudjonsson's research for the proposition that false confessions rarely occur, since Gudjonsson himself generally makes the opposite point: that false confessions occur more frequently than most people believe.(46) In his treatise on the psychology of interrogation and confession, Gudjonsson devotes three full chapters to the problem of false confession and miscarriages of justice,(47) describes numerous documented cases of false confession,(48) and concludes that, "[f]alse confessions, which most commonly result from psychological coercion during police interrogation, are known to be the cause of wrongful conviction in a sizeable proportion of all cases where miscarriages of justice have occurred."(49) However, Cassell does not substantively discuss Dr. Gudjonsson's seminal book in the main text of his article. Instead, he focuses on three studies of Icelandic prisoners by Gudjonsson and his colleagues.(50) Cassell reports how often these prisoners claimed to have falsely confessed to the crimes for which they were incarcerated, but he relegates to a footnote the more important issue of how often they claimed to have made false confessions.(51) Twelve percent (N=27) of the Icelandic prisoners interviewed in Sigurdsson and Gudjonsson's 1994 study reported having in the past made a false confession.(52) In addition, Sigurdsson and Gudjonsson cite an earlier English study of juveniles in which 23% (N=60) of those interviewed reported having made a false confession to police.(53) In Gudjonsson and Sigurdsson's more recent study, once again 12% (N=62) of the Icelandic prisoners interviewed reported making a false confession to police some time in their lives.(54) Contrary to Cassell's reading of Gudjonsson's research, the empirical record does not cast doubt on the likely occurrence of false confessions. Moreover, Cassell fails to mention the conditions that limit the generalizability of Gudjonsson's research on Icelandic and English interrogation practices to contemporary American interrogation practices. Unlike their American counterparts, Icelandic and English police are not permitted to trick or deceive their suspects.(55) This difference is extremely important because our field research identifies the tactic of creating a sense of hopelessness by overstating or inventing evidence--which is not available to Icelandic or English interrogators--as one of the major causes of police-induced false confessions in America.(56) C. ESTIMATING ERROR RATES BY AGGREGATING IGNORANCE The problem with Cassell's approach to estimating the rate of wrongful conviction from false confessions is easily recognizable to anyone familiar with the intellectual game of computer model building and computer simulation. When a great deal of data is available about the operations of a complex system, computer simulations of the system's behavior can be helpful and reveal properties that are not immediately obvious. However, as the amount of reliable information declines and more assumptions have to be added to fill gaps in actual knowledge, the entire enterprise becomes questionable because the system's predictions become determined by the assumptions that were introduced in the beginning. In the end, the model turns out to be nothing more than a structure that transmutes an initial assumption into a consequent of the analysis. Cassell's numerical estimate of the harms of Miranda illustrates this problem because his conclusions all depend on nothing but speculation about the annual rate of criminal justice system error. Cassell arrives at his estimate of wrongful convictions from false confessions by multiplying the number of annual convictions (variable number 1) by the error rate in the criminal justice system (variable number 2) by the proportion of errors attributable to false confessions (variable number 3).(57) The fatal flaw in this calculation, however, is that no one knows--or even has a reasonably good idea of--the error rate in the system (variable number 2). Because there are no valid estimates of the rate of system error to be found anywhere in the literature on miscarriages of justice, all of Cassell's extrapolations are utterly speculative. In the end, Cassell's estimates can never be any better than the assumptions he makes about this error term. Cassell attempts to maneuver around this glaring problem by drawing upper-end and lower-end estimates of the error rate in the system from the work of Ronald Huff and his colleagues.(58) The upper-end estimate upon which Cassell draws puts the rate of system error at .05% (1 out of every 200 convictions). However, this estimate is nothing more than the collective speculation of some criminal justice officials in Ohio and some state attorney generals: it has absolutely no basis in observation or fact. Huff et al. arrived at this figure by asking 55 Ohio county judges, 53 prosecutors, 21 public defenders, 59 sheriffs and chiefs of police, and 41 state attorney generals to guess into which of four categories the error rate in the system most likely fell ("never," "less than 1%," "1-5%," and "6-10%"). Huff et al. then arbitrarily assigned the final number as the mid-point between the two most frequently chosen categories--"never" and "less than 1%."(59) Because this estimate is based on the aggregated ignorance of some criminal justice officials, it is empirically worthless as a valid measure of the tree error rate in the criminal justice system.(60) Even more remarkably, Cassell bases his lower-end estimate of the error rate in the entire American criminal justice system on the opinion of one Ohio county judge, who responded to the Huff et al. survey by stating that he has "a strong suspicion that each year in Ohio, at least one or two dozen persons are convicted of crimes of which they are innocent."(61) From this single statement, Cassell extrapolates a system error rate of .035%,(62) which he then calls a "reasonable estimate."(63) This method of estimating a real world phenomenon is patently absurd. Cassell's willingness to base his policy analysis and recommendations for profound changes in the societal regulation of police interrogation on such flimsy guesswork lends support to Professor Schulhofer's observation that Cassell's extrapolations are often "simply rhetoric,"(64) and "not a serious foundation for assessing social policy."(65) Relying on the collective speculations found in the Huff et al. study to estimate the error rate in the American criminal justice system makes about as much sense as relying on biblical writings to estimate the age of the earth.(66) III. DOES MIRANDA HARM THE INNOCENT? Cassell has advanced a plainly counter-intuitive and entirely speculative argument: that "Miranda has harmed innumerable innocents by preventing the police from obtaining confessions from the actual perpetrators of crimes...."(67) The typical conservative attack on Miranda(68) is that it protects the guilty from apprehension and punishment. We know of no other person who has ever argued that Miranda affirmatively harms innocent individuals who may later become custodial suspects or that it affirmatively prevents the exoneration and release of innocent individuals who have been wrongfully convicted and incarcerated. If Cassell's argument is correct, then he has turned conventional legal wisdom on its head by showing that Miranda not only frustrates the imperatives of crime control, but also that it frustrates the imperatives of due process.(69) But is Cassell's argument correct? According to Cassell, there are two scenarios that describe how Miranda harms the innocent.(70) Under the first, Cassell's frustrated detective scenario, a guilty criminal suspect, who can be arrested only if he confesses, fails to do so, but instead invokes one of his Miranda rights and escapes arrest.(71) Cassell supposes that in the absence of a rule requiring police to issue Miranda warnings to custodial suspects, the guilty party would not have refused to confess and would have been arrested and convicted. In the frustrated detective scenario, having failed to obtain a confession from the truly guilty party, the detective goes on and interrogates some conveniently available innocent suspect. The innocent not only waives Miranda, but thereafter gives a false confession. The innocent is then wrongly convicted.(72) Cassell's vision of police conduct here seems highly unlikely, and more bizarre than anything likely to be voiced seriously even at a conference of the criminal defense bar. Why would police, having correctly identified a perpetrator (presumably because solid evidence points to him), but having failed to successfully interrogate him, go on to extract a confession from an innocent, much less even interrogate him? This makes little sense. While our research indicates that false confessions are caused by police incompetence, negligence and disregard of prohibitions against the use of coercion, we have never seen a case in which police choose to give up their pursuit of a strong suspect and interrogate an innocent suspect merely because they have been foiled by a Miranda invocation.(73) In our experience, a Miranda invocation tends to increase, rather than decrease, police interest in a suspect. In Cassell's second scenario, the ever-diligent detective scenario, an innocent person has already been wrongfully convicted and imprisoned.(74) Despite the fact that the police, as the investigative arm of the state, gathered the evidence supporting the conviction, they strangely keep their investigation open and eventually go on to interrogate the guilty suspect, who can be found out only if he confesses. Instead, however, the guilty suspect decides to invoke one of his Miranda rights and refuses to cooperate with his interrogators. In the absence of a rule requiring police to issue Miranda warnings to custodial suspects, Cassell supposes that the guilty suspect would have confessed and eventually been convicted.(75) But, because Miranda prevented police from obtaining a confession from the guilty party, the imprisoned innocent is neither exonerated nor released.(76) The ever-diligent detective scenario also is unsupported by logic or empirical evidence and thus remains highly unlikely, if not completely implausible. The fatal flaw in Cassell's everdiligent detective scenario is the assumption that police generally keep open cases in which their investigations have already produced a confession and a conviction. This assumption is not only contrary to police practice, but is also completely unsupported by any of the data we have gathered in our study of police-induced false confessions(77) or, for that matter, by any of the research literature on false confessions and miscarriages of justice. While reliable confessions from the true perpetrator are among the leading sources of exoneration of the wrongfully convicted, police virtually always acquire such confessions accidentally or unintentionally.(78) Sometimes--as in the cases of the Phoenix Temple Four,(79) Johnny Lee Wilson,(80) and Melvin Lee Reynolds,(81) for example--the very police who coerced a false confession from an innocent suspect do everything in their power to discredit the confession of the true perpetrator(s). In sum, both of Cassell's scenarios are so highly implausible as to seem fanciful. Instead of marshalling any empirical evidence that might bear on the phenomenon of false confessions, Cassell relies on a series of questionable assumptions and tenuous extrapolations to further his argument that Miranda--contrary to its intent, logic and appearance--somehow endangers the innocent. IV. CONCLUSION: SCAPEGOATING MIRANDA What, then, are we to make of Cassell's conclusions about the competing risks to the innocent from false confessions and lost confessions? Does anyone seriously believe that Miranda prevents "tens of thousands"(82) of criminals from being convicted every year? Or that Miranda actually harms the very innocents whom it empowers to resist and/or terminate stationhouse interrogation? Whether or not Cassell fully believes in the validity and reliability of his statistical manipulations, resort to quick and dirty quantification offers him a unique rhetorical strategy in the policy debate about Miranda's future. By casting his argument in quantitative terms--no matter how flawed his assumptions, and no matter how contrived his reasoning--Cassell eschews the doctrinal arguments traditionally necessary to argue before legal scholars that a Supreme Court decision--one of the most famous and influential Supreme Court decisions, no less--should be overturned. Cassell also cloaks his attack on Miranda in a language and style of analysis with which most legal scholars are wholly unprepared to engage, and may even impress readers who have no familiarity with social science research methods. Moreover, Cassell's willingness to engage in crude quantification offers him a rhetorical opportunity to pitch his "estimation methodology" and numerical conclusions as providing the most reasonable empirical basis for public policy regulating interrogation. Positing that the problem of wrongful conviction from false confessions "is an empirical or numbers issue that cannot be resolved by a priori, theoretical reasoning,"(83) Cassell criticizes actual empirical research of others such as ourselves for "reason[ing] solely from anecdotal example,"(84) and then concludes that "[f]or public policy purposes, the anecdotal evidence collected by Leo and Ofshe tells us little."(85) Like the emperor who wears no clothes, however, Cassell's argument here is based solely on illusion. Cassell makes empty claims about a superior "estimation methodology" and advocates only a dark vision about the possible disastrous consequences of Miranda. But Cassell might as well admit that his analysis is a priori reasoning because it is not based on data that reveals anything about the actual workings of the criminal justice system. Cassell's try at quantification eschews original data altogether and instead relies on assumptions and extrapolations that, masquerading as sound estimates of a real world phenomenon, ultimately amount to no more than elaborate speculation. There is a danger, however, in treating speculation as a reasonable estimate of fact, especially when the speculation may have fateful consequences for human lives. There is also a danger in bashing Miranda when it diverts our attention from the real problems of the criminal justice system--such as the wrongful conviction of the innocent--and the possible solutions to those problems. Unless and until Professor Cassell can marshal some compelling data demonstrating Miranda's allegedly tragic consequences for the innocent, we must continue to presume that Miranda is a procedural safeguard for, not a procedural detriment to, the innocent. If Cassell's primary goal is to bash Miranda, then it is easy to understand why he has relied on quantitative guesswork to argue now that Miranda harms the innocent more than it helps them. This argument appears to be rooted in Cassell's longstanding contempt for Miranda and offers him yet another rhetorical weapon in his highly charged anti-Miranda crusade.(86) Not only does Miranda prevent tens of thousands of lost confessions, lost clearances and lost convictions every year--as Cassell has repeatedly suggested in the past--but now it can also be said to hurt the very innocent whom it otherwise seems to help resist coercive police interrogation. For everyone involved then, Miranda is, according to Cassell, a social fraud: it produces no social benefits, it creates only victims. It must be overturned or else more innocents will be harmed. This newfound discovery--surprisingly unnoticed by anyone else in more than three decades of voluminous popular and academic writing on Miranda--would, if true, seem to offer anti-Miranda critics like Cassell yet another point of attack in any policy debate about Miranda's legitimacy. However, if Cassell's primary goal is indeed to protect the innocent, then his emphasis on overturning Miranda is entirely misplaced. Instead, a proper reading of the miscarriage of justice literature would suggest that Cassell should redirect his energies to the advocacy of tougher safeguards to protect custodial suspects and criminal defendants--not only against police-induced false confession, but also against other prominent sources of wrongful conviction such as eyewitness misidentification,(87) the prosecutorial withholding of exculpatory evidence,(88) the use of perjured testimony by so-called jailhouse informants,(89) and ineffective assistance of counsel.(90) For it is these kinds of abuses--not the constitutionally-based Miranda warnings--that lead to miscarriages of justice in the first place and that prevent the exoneration of the wrongfully convicted and incarcerated.(91) (1) Paul Cassell, Protecting the Innocent from False Confessions and Lost Confessions -- And From Miranda, 88J. CRIM. L. & CRIMINOLOGY 497 (1998). (2) Others have likewise criticized Cassell for the ideological nature of his scholarship. See, e.g., Hugo Adam Bedau 8,: Michael L. Radelet, The Myth of Infallibility: A Reply to Markman and Cassell, 41 STAN. L. REV. 161, 169 (1988) (observing that Markm material."); George C. Thomas, III, Telling Half Truths, LEGAL TIMES, Aug. 12, 1996, at 20 ("While Miranda is not immune from questioning, advocacy cannot replace careful scholarship"); id. at 24 ("[S]cholars have a duty to describe all the evidence and to acknowledge contrary interpretations if they are (3) See generally Paul G. Cassell & Richard Fowles, Falling Clearance Rates After Miranda: Coincidence or Consequences, 50 STAN. L. REV. 1181 (1998); 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 (1998); Paul G. Cassell, Miranda's "Negligible" Effect on Law Enforcement: Some Skeptical Observations, 20 HARV. J.L. & PUB. POL'Y 327 (1997); Paul G. Cassell, All Benefits, No Costs: The Grand Illusion of Miranda's Defenders, 90 NW. U. L. REV. 1084 (1996); Paul G. Cassell, Miranda's Social Costs: An Empirical Reassessment, 90 NW. U. L. KEV. 387 (1996); 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 (1996). (4) 384 U.S. 436 (1966). (5) See supra note 3 and accompanying text. (6) Cassell, supra note 1, at 552; see also id. at 538-51. (7) Id. at 552. (8) Id. at 513-24. (9) Id. at 502. (10) Id. at 503. (11) For the argument that Miranda neither dispels nor affects the conditions of modern interrogation that lead to false confession, and therefore Miranda does not offer any significant protection against the admission of false confessions, see Richard A. Leo, Miranda and the Problem of False Confessions, in THE MIRANDA DEBATE: LAW, JUSTICE AND POLICING 271, 271-82 (Richard A. Leo & George C. Thomas, III eds., 1998). (12) Professor Stephen Schulhofer has already replied to the merits of Cassell's argument that the Miranda warning and waiver requirements be replaced with videotaping. See Stephen J. Schulhofer, Miranda's Practical Effect: Substantial Benefits and Vanishingly Small Social Costs, 90 NW. U.L. REV. 500, 556-57 (1996). (13) See also Richard J. Ofshe & Richard A. Leo, The Social Psychology of Police Interrogation: The Theory and Classification of True and False Confessions, 16 STUD. L., POL. & SOC'Y 189, 191-94 (1997) [hereinafter Ofshe & Leo, Social Psychology]; Richard A. Leo & Richard J. Ofshe, Missing the Forest for the Trees: A Response to Paul Cassell's Balanced Approach" to the False Confession Problem, 74 DENV. U.L. REV. 1135, 1136-39 (1997) [hereinafter Leo & Ofshe, Missing the Forest]. (14) Professor Schulhofer has advanced this same criticism against Cassell in another context. See Stephen J. Schulhofer, Bashing Miranda Is Unjustified--And Harmful, 20 HARV. J.L. PUB. POL'Y 347, 372 (1997) ("All the talk about `terrible Miranda' is therefore worse than just wrong. That kind of talk amounts to picking on a convenient scapegoat and distracting our attention from the real problem."). (15) This discussion necessarily draws on and repeats an earlier debate we had with Cassell on this very issue. See Leo & Ofshe, Missing the Forest, supra note 13, at 1135. (16) Cassell, supra note 1, at 513-24. (17) See Leo & Ofshe, Missing the Forest, supra note 13, at 1137-39; Ofshe & Leo, Social Psychology, supra note 13, at 191. (18) See Leo & Ofshe, Missing the Forest, supra note 13, at 1135, 1143. (19) Cassell, supra note 1, at 500. (20) Id. at 500-03, 505-07, 524, 537-38. (21) See 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) [hereinafter Leo & Ofshe, The Consequences of False Confessions]. (22) Cassell tries to minimize the frequency and importance of police-induced false confession by arguing that "common sense suggests that suspects will more often `confess' for understandable reasons (such as protecting a loved one) than because police have somehow convinced them that they actually committed the crime." Cassell, supra note 1, at 519. Curiously, unlike the problem of wrongful conviction from false confession, which Cassell asserts is an empirical issue that "cannot be resolved by a priori reasoning--Cassell's a priori reasoning. See id. at 500. However, Cassell's sense here is not that common at all, and, in fact, is clearly contradicted by the empirical research literature on the social psychology of false confessions. See Leo & Ofshe, The Consequences of False Confessions, supra note 21, at 429-91; Richard J. Ofshe & Richard A. Leo, The Decision to Confess Falsely: Rational Choice and Irrational Action, 74 DENV. U.L. REV. 979, 1001-106 (1997) [hereinafter, Ofshe & Leo, The Decision to Confess Falsely]; Ofshe & Leo, Social Psychology, supra note 13, at 191. Cassell further suggests that his "common sense" argument is supported by a single study of police-induced versus suspect-induced false confessions among Icelandic prisoners, and that the Icelandic figure probably overstates the proportion of suspect induced confessions in America because of Iceland's inquisitorial legal system. See Cassell, supra note 1, at 519. However, Cassell not only ignores the relevant American research literature on this issue, but also fails to mention that, unlike in America, Icelandic interrogators are not legally permitted to employ trickery or deception during interrogation. Personal Correspondence from Gisli Gudjonsson to Richard Leo (Jan. 19, 1998) (on file with authors). In addition, in Iceland all confessions have to be repeated in front of a judge. Id. Thus, there is no reason to believe that the figures Cassell cites from a single Icelandic study translate at all into a reasonable estimate of what in fact occurs in America. There are other problems--such as the differences in the nature and rates of crime, the racial composition of the universe of suspects, and the cultural attitudes toward authority--that limit the generalizability of the research in Iceland to the research in America. See infra note 46 and accompanying text. Cassell also suggests that, "other things being equal, it should be easier for the police to persuade someone to falsely confess to a larceny than to confess to a murder. The Gudjonsson and Sigurdsson study provides strong support for this hypothesis." See Cassell, supra note 1, at 523 (citing Gisli H. Gudjonsson & Jon F. Sigurdsson, How Frequently Do False Confessions Occur?: An Empirical Study Among Prison Inmates, 1 PSYCHOL, CRIME & L. 21, 23 (1994)). The problem with Cassell's hypothesis here, of course, is that all other things are not equal. While in theory it may be easier for police to elicit false confessions to less serious crimes, in practice they put far more resources, time and effort into eliciting confessions in more serious cases. For this reason, as well as the methodological problems of comparing American and Icelandic interrogation practices, Gudjonsson and Sigurdsson's study hardly provides "strong" support for Cassell's hypothesis here. Moreover, Cassell fails to mention our field research in America, which indicates that the problem of police-induced false confession occurs disproportionately in more serious cases. See Leo & Ofshe, Missing the Forest, supra note 13, at 113940. (23) See Stephen J. Schulhofer, Pointing In The Wrong Direction, LEGAL TIMES, Aug. 12, 1996, at 21 ("Readers should understand that these are simply advocacy numbers, derived from indefensibly selective accounts of the available data."). (24) It is curious that Cassell characterizes the problem of wrongful conviction from false confessions as "rare" and "exotic" in light of his own multi-step speculation that 10 to 394 wrongful convictions from false confessions occur yearly. See Cassell, supra note 1, at 502, 538. If one takes Cassell's quantitative estimates seriously--we do not, but certainly Cassell does--then in the 32 years since the Supreme Court decided Miranda v. Arizona, somewhere between 320 and 12,608 individuals have been wrongly convicted from false confessions alone. If true, this would surely be evidence that police-induced false confession represents a serious problem in America. (25) See Cassell, supra note 1, at 497-98. (26) Id. (27) Id. at 499. (28) Id. at 500. (29) Id. (30) Id. at 513-32. (31) Id. at 538-56. (32) Id. at 501 (quoting Schulhofer, supra note 12, at 505). (33) Schulhofer, supra note 12, at 546. (34) See Cassell, supra note 1, at 504-05. (35) For research on the ultimate miscarriage of justice, see Hugo Adam Bedau & Michael L. Radelet, Miscarriages of Justice in Potentially Capital Cases, 40 STAN. L. Rev. 21 (1987); Samuel R. Gross, The Risks of Death: Why Erroneous Convictions are Common in Capital Cases, 44 BUFF. L. REV. 469 (1996). (*) We thank Charles Weisselberg and Welsh White for their helpful comments. We also thank Paul Cassell for graciously providing pre-publication copies of various drafts of his article. RICHARD A. LEO(**) AND RICHARD J. OFSHE(***) (**) Assistant Professor of Criminology, Law and Society, University of California, Irvine; Assistant Professor of Sociology and Adjoint Professor of Law, University of Colorado, Boulder; Ph.D., University of California, Berkeley; J.D., Univers Berkeley. (***) Professor of Sociology, University of California, Berkeley; Ph.D., Stanford University.
|
