Article review

CON STR UCTI NG A FRAM EWOR K FOR CR I M I NAL J USTICE R E S EARCH: LEAR N I NG FROM PACKE R’S M I STAKE S Stuart Macdonald*

This article examines the framework which Herbert Packer devised for analyzing the criminal justice process. Warning against an indiscriminate use of the word “model”—something which Packer introduced, which critics of his work have perpetuated, and which today is commonplace in criminal justice research—it distinguishes three distinct tools which researchers might employ—strong ideal types, weak ideal types and non-ideal types—and underscores the importance of drawing a sharp distinction between empirical work and evaluative work. Exposing other fundamental flaws in Packer’s framework, the article also abstracts two other general lessons for criminal jus- tice research: (1) arguing that a one-dimensional framework like Packer’s is insufficient, it advances the normative claim that a multidimensional frame- work is needed; and (2) it shows that Packer’s simplistic approach to the analy- sis of values is ultimately inadequate. So as well as constructing a number of tools which may be used by criminal justice researchers, and offering examples of how they might be employed, the article establishes the general contours of a framework for criminal justice research.

New Criminal Law Review, Vol. 11, Number 2, pps 257–311. ISSN 1933-4192, electronic ISSN 1933-4206. © 2008 by the Regents of the University of California. All rights reserved. Please direct all requests for permission to photocopy or reproduce article content through the University of California Press’s Rights and Permissions website, http://www.ucpressjournals.com/ reprintInfo.asp. DOI: 10.1525/nclr.2008.11.2.257.

*Lecturer in Law, Swansea University, U.K. I would like to thank all of those who have commented upon drafts of this article, including Andrew Ashworth, Oren Ben-Dor, Julia Fionda, Andrew Halpin, and Andrew Rutherford.

NCLR1102_03.qxd 5/29/08 5:01 PM Page 257

I . I NTROD UCTION

In August 1967 Professor Herbert Packer of Stanford University put aside his administrative duties as vice provost, and went away to Santa Cruz to concentrate his efforts on the book he was writing. The result was the acclaimed The Limits of the Criminal Sanction,1 winner in 1970 of the prestigious triennial Order of the Coif book award, the highest honor that could be bestowed on an American legal scholar. While the book’s primary concern was to question the “far too indiscriminate”2 way in which the criminal sanction was being resorted to, it is best known for Packer’s two models of the criminal justice process, the “crime control” and “due process” models.3

This article begins, in part II, with a description of Packer’s models and the “common ground” which Packer said exists between them. The final section of part II provides an overview of the existing critiques of Packer’s framework, and explains what this article seeks to add to this body of lit- erature. Having shown that Packer’s intention was to construct two ideal- types, and explained why he failed in this task, the article then attempts, in part III, to succeed where Packer did not. This part of the article accordingly constructs four ideal-types, based on the values of investiga- tive efficiency, operational efficiency, and reliability. Part IV then turns to the other values discussed by Packer—prevention of abuse of state power, equality, and crime prevention. Emphasising the importance of distin- guishing between empirical work and evaluation, the article explores the implications of this distinction for those constructs founded upon these three values. The article concludes, in part V, by abstracting from the dis- cussion some general lessons for criminal justice research. As well as con- structing a number of tools which may be used by criminal justice researchers, and offering examples of how they might be employed, the article thus aims to establish the general contours of a framework for crim- inal justice research.

| N E W C R I M I NAL LAW R E VI E W | VOL . 11 | NO. 2 | SPR I NG 2008258

1. Herbert Packer, The Limits of the Criminal Sanction (1968). 2. Id. at 364. For an outline of the central argument of the book, see infra note 129. 3. The part of the book expounding these models is almost identical to an earlier

article. See Herbert Packer, Two Models of the Criminal Process, 113 U. Pa. L. Rev. 1 (1964).

NCLR1102_03.qxd 5/29/08 5:01 PM Page 258

I I . AN I NTROD UCTION TO PACKE R’S MOD E LS

A. The Crime Control and Due Process Models

According to the crime control model, the most important function of the criminal process is to repress criminal conduct and thereby safeguard social freedom. It insists that “primary attention be paid to the efficiency with which the criminal process operates to screen suspects, determine guilt, and secure appropriate dispositions of persons convicted of crime.”4

The model is founded upon the “presumption of guilt.”5 According to this presumption “the screening processes operated by police and prosecutors are reliable indicators of probable guilt,” and so suspects found to be prob- ably guilty by these processes should be convicted “as expeditiously as pos- sible . . . with a minimum of occasions for challenge, let alone post-audit.”6

The model thus finds in the presumption of guilt “a factual predicate for the position that the dominant goal of repressing crime can be achieved through highly summary processes without any great loss of efficiency.”7

Packer stated that if the crime control model resembles an assembly line, the due process model resembles an obstacle course. The due process model insists that reliability is of at least as much importance as efficiency, and so “if efficiency demands short-cuts around reliability, then absolute efficiency must be rejected.”8 The model regards investigative and

CONSTR UCTI NG A FRAM EWOR K F OR CR I M I NAL J USTICE R ESEARCH | 259

4. Packer, supra note 1, at 158. 5. Packer explained that the presumption of guilt is not the opposite of the presumption

of innocence. The presumption of guilt “is purely and simply a prediction of outcome,” whereas the presumption of innocence “is a direction to officials about how they are to pro- ceed.” Packer, supra note 1, at 161. The presumption of innocence requires “that until there has been an adjudication of guilt by an authority legally competent to make such an adjudication, the suspect is to be treated, for reasons that have nothing whatever to do with the probable outcome of the case, as if his guilt is an open question.” Id. Note that Packer’s presumption of guilt only applies where it would be “plainly absurd” to maintain that the suspect did not commit the offense, such as where a person commits a murder in front of many witnesses and confesses to the crime. It must therefore be distinguished from situations where an investiga- tor suspects an individual is guilty of an offense and then seeks to gather admissible evidence to substantiate his hunch. For a stark example of this ethos, see Williamson v. Reynolds, 904 F. Supp. 1529 (E.D. Okla., 1995). Ron Williamson was prosecuted in Oklahoma for first degree murder and spent eleven years on death row before finally being released in April 1999.

6. Packer, supra note 1, at 160. 7. Id. at 162. 8. Id. at 165.

NCLR1102_03.qxd 5/29/08 5:01 PM Page 259

prosecutorial fact finding as prone to error, and so insists upon formal fact-finding processes. But “this is only the beginning of the ideological difference between the two models.”9 First, the due process model stresses that “power is always subject to abuse,” and so “would accept with con- siderable equanimity a substantial diminution in the efficiency with which the criminal process operates in the interest of preventing official oppres- sion of the individual.”10 The doctrine of legal guilt “implements these anti-authoritarian values” by holding that a person may only be held liable for a crime if the finding of guilt was “made in procedurally regu- lar fashion and by authorities acting within competences duly allocated to them.”11 Second, the model asserts that the principle of equality “imposes some kind of public obligation to ensure that financial inability does not destroy the capacity of an accused to assert what may be meri- torious challenges to the processes being invoked against him.”12 And third, the model is sceptical “about the morality and utility of the crimi- nal sanction,” which creates “pressure to limit the discretion with which that power is exercised.”13

B. The Common Ground between the Crime Control and Due Process Models

Packer’s analytical framework was constructed upon four assumptions, which he described as “some common ground” between the two models.14

The first of these assumptions was that “the function of defining conduct that may be treated as criminal is separate from and prior to the process of identifying and dealing with persons as criminals.”15 He described this as one side of the U.S. Constitution’s ex post facto clause coin. The second assumption was the other side of this coin—“the criminal process ordi- narily ought to be invoked by those charged with the responsibility for doing so when it appears that a crime has been committed and there is a

| N E W C R I M I NAL LAW R E VI E W | VOL . 11 | NO. 2 | SPR I NG 2008260

9. Id. at 165. 10. Id. at 166. 11. Id. at 166. 12. Id. at 169. 13. Id. at 170, 171. 14. Id. at 154–158. 15. Id. at 155.

NCLR1102_03.qxd 5/29/08 5:01 PM Page 260

reasonable prospect of apprehending and convicting its perpetrator.”16 The third assumption was that “a degree of scrutiny and control must be exer- cised with respect to the activities of law enforcement officers, that the security and privacy of the individual may not be invaded at will.”17 There should be some limits to the powers of government to investigate and apprehend persons suspected of committing crimes. The final assumption was that “the alleged criminal is not merely an object to be acted upon but an independent entity in the process who may, if he so desires, force the operators of the process to demonstrate to an independent authority (judge and jury) that he is guilty of the charges against him.”18 This assumption encompasses such terms as “the adversary system,” “procedural due process,” “notice and an opportunity to be heard” and “day in court.” Packer explained that these assumptions could be “roughly equated with minimal agreed limits expressed in the Constitution of the United States and, more importantly, with unarticulated assumptions that can be per- ceived to underlie those limits.”19 These assumptions reflect the fact that Packer’s models were intended by him as tools for analysing a criminal process which “operates within the framework of contemporary American society.”20 They are not apt to be used to analyze criminal processes which do not adhere to these premises.

It follows from this “common ground” that a sharp distinction must be drawn between, on the one hand, prioritizing the values associated with one of Packer’s models at the expense of the values associated with the other model, and, on the other hand, a challenge to the assumptions that the framework is constructed upon. Take the void for vagueness doctrine, rooted in the due process clauses of the Fifth Amendment (federal statutes) and the Fourteenth Amendment (state statutes), as an example.21

It is not uncommon for researchers today to use Packer’s models loosely; such a researcher might describe a law which fell foul of the void for vague- ness doctrine as embodying crime control values. In fact, such a law would challenge the first of the assumptions underlying Packer’s framework. In

CONSTR UCTI NG A FRAM EWOR K F OR CR I M I NAL J USTICE R ESEARCH | 261

16. Id. 17. Id. at 156. 18. Id. at 157. 19. Id. at 155. 20. Id. at 154. 21. See, e.g., Lanzetta v New Jersey 306 U.S. 451 (1939).

NCLR1102_03.qxd 5/29/08 5:01 PM Page 261

the eyes of Packer’s framework, then, the law should be described not as embodying crime control values, but as challenging the very framework within which the American criminal process operates.

Unfortunately, however, Packer himself obscured the clarity of this dis- tinction. He explained that in a criminal process which adheres to the four assumptions outlined above, like the American one, two separate value systems compete for priority. Each of his models represents the values underlying one of these value systems “to the exclusion of all of the values underlying the other.”22 He thus stated that “the models are polarities,”23

the crime control model insisting solely on the value of efficiency and the due process model insisting only on the values of reliability, prevention of abuse of state power and equality. These extreme positions exist at either end of a spectrum:

As we examine the way the models operate in each successive stage [of the criminal process], we will raise two further inquiries: first, where on a spec- trum between the extremes represented by the two models do our pres- ent practices seem approximately to fall; second, what appears to be the direction and thrust of current and foreseeable trends along each such spectrum?24

The framework Packer purported to create is thus clear. A policy deci- sion which adheres to the four assumptions constituting the common ground may be located on the spectrum between the crime control and due process models. In this way trends may be analyzed and explicated. But there is a problem with this framework. The assumptions underlying it reflect what are, broadly speaking, rule of law concerns. They are thus rooted in the liberal values which Packer attributed to the due process model. Unsurprisingly, then, there is a tension between the so-called com- mon ground and Packer’s statement that the crime control model insists on the value of efficiency to the “exclusion” of the values underlying the due process model. For example, according to the third of the assumptions which Packer said underlies his framework there must be at least some scrutiny and control of the activities of law enforcement officers in order

| N E W C R I M I NAL LAW R E VI E W | VOL . 11 | NO. 2 | SPR I NG 2008262

22. Packer, supra note 1, at 154. 23. Id. 24. Id. at 153.

NCLR1102_03.qxd 5/29/08 5:01 PM Page 262

to protect individuals’ security and privacy. This is contradicted by the crime control model’s assertion that “[i]t is enough of a check on police discretion to let the dictates of police efficiency determine under what cir- cumstances and for how long a person may be stopped and held for inves- tigation.”25 It seems that Packer sensed a problem here, simply stating, oxymoronically, that “the polarity of the two models is not absolute.”26

This does not deal with the problem adequately. If the crime control model insists on the value of efficiency to the exclusion of the values asso- ciated with the due process model, Packer’s analytical framework implodes; the model at one end of his spectrum challenges the very assumptions upon which the spectrum is constructed. If, on the other hand, the crime con- trol model’s insistence on efficiency is not to the outright exclusion of the values associated with the due process model, the content of the crime con- trol model becomes unclear. To what extent may the “common ground” impinge on the crime control model’s insistence on efficiency? To what extent may the dictates of efficiency be sacrificed so that the activities of law enforcement officers can be scrutinized and controlled? Moreover, if encroachments on the dictates of efficiency are countenanced, the crime control and due process models can no longer be polar opposites at either end of a spectrum of possibilities. So while the intended structure of Packer’s analytical framework is clear enough, his failure to clearly distin- guish the framework’s underlying assumptions from the values which oper- ate within the framework means that ultimately the structure collapses.

C. Existing Critiques of Packer’s Framework

Packer’s framework has been subjected to considerable criticism. Broadly speaking, these criticisms concern the purportedly dichotomous nature of Packer’s framework, the framework’s undue selectivity, and Packer’s claim that he had constructed “models.”

CONSTR UCTI NG A FRAM EWOR K F OR CR I M I NAL J USTICE R ESEARCH | 263

25. Id. at 177. 26. Id. at 154. Packer also stated that the ideology of the due process model “is not the

converse of that underlying the Crime Control Model.” Id. at 163. He immediately quali- fied this statement by explaining that the due process model “does not rest on the idea that it is not socially desirable to repress crime.” Id. This particular statement was thus intended to convey nothing more than the fact that both models agree that the raison d’être of the criminal justice process is to apprehend, convict, and sentence those who engage in con- duct which has been defined as criminal. See infra part II.C.

NCLR1102_03.qxd 5/29/08 5:01 PM Page 263

First, commentators have claimed that Packer’s framework presents a false dichotomy. David Smith has argued that “the Crime Control Model is concerned with the fundamental goal of the criminal justice system, whereas the Due Process Model is concerned with setting limits to the pursuit of that goal.”27 He accordingly claimed that it would be mistaken “to evaluate a system of criminal justice purely by how closely it approxi- mates to the Due Process Model.”28 Smith’s criticism usefully highlights the obfuscatory effect of the title “crime control” model. It obscures the distinction between the raison d’être of the criminal justice process (to apprehend, convict, and sentence those who engage in conduct which has been defined as criminal) and the values of efficiency, speed, and finality— which concern one possible way in which the apprehension and convic- tion of offenders might be pursued.29 But Smith’s assumption that the crime control model is concerned with the first of these—the raison d’être of the criminal justice process—was mistaken. That the crime control model is in fact concerned with the way in which the apprehension and conviction of offenders is pursued is confirmed by Packer’s analogies of an assembly line and an obstacle course. The point about the assembly line is not what product emerges at the end (a criminal conviction), but the speed and efficiency with which that product is produced. To remedy this confusion, Peter Duff has proposed renaming the crime control model the “efficiency model.”30 As will be argued below, however, this new title would itself be deeply problematic.

Malcolm Feeley has also doubted the dichotomous nature of Packer’s framework—asserting the empirical irrelevance of due process in many minor cases where “the cost of invoking one’s rights is frequently greater

| N E W C R I M I NAL LAW R E VI E W | VOL . 11 | NO. 2 | SPR I NG 2008264

27. David J. Smith, Case Construction and the Goals of Criminal Process, 37 B.J. Criminology 319, 335 (1997). See also Andrew Ashworth & Mike Redmayne, The Criminal Process 39 (3d ed. 2005) (making a similar point).

28. Smith, supra note 27, at 336. 29. This distinction is recognized by other writers. Ashworth distinguished the “general

justifying aim” of criminal justice from principles and policies which qualify the pursuit of that aim. A.J. Ashworth, Concepts of Criminal Justice [1979] Crim LR 412. See also Neil Walker & Mark Telford, Designing Criminal Justice: The Northern Ireland System in Comparative Perspective (Criminal Justice Review Group Research Report 18) (2000) (proposing a two-tiered “meta-model” of the criminal justice process).

30. Peter Duff, 38 B.J. Criminololgy 611 (1998). Note that Smith subsequently refined his analysis of Packer’s models. Smith, 38 B.J. Criminolology 616 (1998).

NCLR1102_03.qxd 5/29/08 5:01 PM Page 264

than the loss of the rights themselves”31—as has Doreen McBarnet, who argued that due process is in fact consistent with crime control: “the law in action is only too close a parallel to the law in the books; due process is for crime control.”32 McBarnet’s contention has, however, been described as overstated by Andrew Rutherford, who, in his study of criminal justice practitioners, found three “working credos.” The first included the belief that “as few fetters as possible be placed upon the authorities in the pur- suit of criminals who, when caught, should be dealt with in ways that are punitive and degrading”; the prevailing concern of the second was “to dis- pose of the tasks at hand as smoothly and efficiently as possible”; and the third encapsulated “empathy with suspects, offenders, and the victims of crime, optimism that constructive work can be done with offenders, adherence to the rule of law so as to restrict state powers, and an insistence on open and accountable procedures.”33

Rutherford’s account ties in with the second set of criticisms commen- tators have leveled at Packer’s framework—that it is unduly selective. This led Michael King to outline a further four models—the medical, bureaucratic, status passage and power models.34 Packer has been particularly criticized for neglecting the role played by resource management35 and victims’ rights.36

Although some have argued that these considerations only assumed greater significance in the years since Packer wrote, so that Packer’s frame- work in fact captured the prevailing concerns of the late 1960s, this seems overly charitable given that Bottoms and McClean highlighted the role of resource management only eight years after the publication of The Limits of the Criminal Sanction.37 In his well-known recent critique of Packer’s framework, meanwhile, Kent Roach has stated that it “cannot explain why women, children, minorities, and crime victims claim rights to the criminal sanction.”38 As a result Packer’s models cannot make sense of

CONSTR UCTI NG A FRAM EWOR K F OR CR I M I NAL J USTICE R ESEARCH | 265

31. Malcolm Feeley, The Process is the Punishment 277 (1979). 32. Doreen McBarnet, False Dichotomies in Criminal Justice Research, in Criminal

Justice: Selected Readings 31 (Baldwin & Bottomley eds., 1978). 33. Andrew Rutherford, Criminal Justice and the Pursuit of Decency 11, 13, 18 (1993). 34. Michael King, The Framework of Criminal Justice ch. 2 (1981). 35. A.E. Bottoms & J.D. McClean, Defendants in the Criminal Process (1976). 36. Ashworth & Redmayne, supra note 27, at 38–40. 37. Id. at 39. 38. Kent Roach, Four Models of the Criminal Process, 89 J. Crim. L. & Criminology

671, 674 (1999).

NCLR1102_03.qxd 5/29/08 5:01 PM Page 265

contemporary debates about pornography or hate speech, or of contem- porary concerns about sexual and domestic violence against women and children and hate crime against minorities, and so have become “as out of date as other hits of the 1960s.”39 Seeking to utilize knowledge gained from victimization studies completed since Packer’s work was published, which demonstrate a widespread underreporting of crime, Roach expounded two further models, the punitive (roller-coaster) and non- punitive (circle) models of victims’ rights.

Third, Packer has been criticized for failing to explain the internal logic of his models.40 John Griffiths has claimed that this failure leaves us with no way of determining whether a particular value belongs more with one model than the other.41 Griffiths attributes this failure to the fact that Packer did not (contrary to his claim) succeed in constructing “models” of the criminal process. The crime control and due process models, he explained, could not have been intended to be “alternative ideals toward which one might strive”; they are also not “entities which have an analogical or metaphoric relation- ship to an actual system of criminal procedure.”42 Developing this, Griffiths argued that the crime control and due process models share an underlying assumption—that the criminal process is a struggle “between two contend- ing forces whose interests are implacably hostile: the Individual (particularly, the accused individual) and the State”43—and so therefore Packer only gave us one model of the criminal process, the battle model. Starting from the opposite assumption, one “of reconcilable—even mutually supportive— interests, a state of love,”44 he then outlined a family model of the criminal

| N E W C R I M I NAL LAW R E VI E W | VOL . 11 | NO. 2 | SPR I NG 2008266

39. Id. at 673. 40. Ashworth & Redmayne, supra note 27, at 40. 41. Griffiths says that Packer gives us no way to determine whether the value of effi-

ciency belongs more with the crime control model than the due process model, except that he happens to assign it to the former. John Griffiths, Ideology in Criminal Procedure or a Third “Model” of the Criminal Process, 79 Yale L.J. 359 (1970). Similarly Ashworth & Redmayne observe that although Packer ascribes the value of speed to the crime control model, since delays are a source of anxiety, inconvenience, and potentially prolonged loss of liberty, an emphasis on speed also belongs to the due process model. Ashworth & Redmayne, supra note 27 at 40.

42. Griffiths, supra note 41, at 362 n.14. 43. Id. at 367. 44. Id. at 371.

NCLR1102_03.qxd 5/29/08 5:01 PM Page 266

process in which the emphasis is on reconciling offenders, victims, and their communities. Commenting that Griffiths “tells us very little about the procedural arrangements which would be likely to flow from the adop- tion of [the battle and family models],” Mirjan Damaska has asserted, in his exposition of adversary and non-adversary models of the criminal process, that what Griffiths “[seems] to have done” is to “present two pos- sible ideological justifications” for these models.45 Meanwhile, according to Damaska, Packer did not offer us two models of the criminal process; instead he presented “a stimulating depiction, rich in implications, of two clashing inner tendencies: the tendency toward efficiency and the tenden- cy toward protecting the rights of the defendant.” The tension between these tendencies “is part and parcel of the dialectics of any criminal process.”46

A possibility which Griffiths did not consider, and which Damaska rejected,47 is that Packer intended to construct “ideal-types” in the sense outlined by Max Weber. Weber explained that ideal-types must be distin- guished from ideals. Whereas an ideal is something against which one evaluates reality, an ideal-type has “no connection at all with value- judgments, and it has nothing to do with any type of perfection other than a purely logical one.”48 An ideal-type is formed “by the one-sided accentu- ation of one or more points of view and by the synthesis of a great many diffuse, discrete, more or less present and occasionally absent concrete individual phenomena, which are arranged according to those one-sidedly

CONSTR UCTI NG A FRAM EWOR K F OR CR I M I NAL J USTICE R ESEARCH | 267

45. Mirjan Damaska, Evidentiary Barriers to Conviction and Two Models of Criminal Procedure: A Comparative Study, 121 U. Pa. L. Rev. 506, 572–73 (1973).

46. Id. at 576. 47. Damaska wrote: “Notwithstanding sporadic passages which seem to suggest the

contrary, closer reading reveals that [Packer] does not purport to contrast two rival ideal- types in designing the criminal process.” Id. at 575. This is because, he claims, the due process model is a negative model, and “it is conceptually impossible to imagine a crimi- nal process whose dominant concern is a desire to protect the individual from public offi- cials. In its pure form, it would lead not to an obstacle course, but rather to mere obstacles and no course on which to place the former.” Id. Cf. infra part IV (discussing Jareborg’s defensive model).

48. Max Weber, “Objectivity” in Social Science and Social Policy, in The Methodology of the Social Sciences 49, 97–98 (Edward A. Shils & Henry A. Finch trans. and eds., 1949).

NCLR1102_03.qxd 5/29/08 5:01 PM Page 267

emphasized viewpoints into a unified analytical construct,” and may be used for both research and exposition:

Historical research faces the task of determining in each individual case, the extent to which this ideal-construct approximates to or diverges from real- ity, to what extent for example, the economic structure of a certain city is to be classified as a “city-economy.” When carefully applied, these concepts are particularly useful in research and exposition.49

The crime control and due process models were not meant as ideals; Packer claimed that “a person who subscribed to all of the values underly- ing one model to the exclusion of all of the values underlying the other would be rightly viewed as a fanatic.”50 Rather they were intended to aid analysis by providing “a convenient way to talk about the operation of a process whose day-to-day functioning involves a constant series of minute adjustments between the competing demands of two value systems.”51

Packer constructed the models by “[abstracting] two separate value sys- tems that compete for priority in the operation of the criminal process” and then imagining a value system that insisted wholeheartedly on one of these sets of values, to the exclusion of the other.52 That Packer’s purpose was to construct something like Weberian ideal-types is thus clear. Ironically, Griffiths himself hinted at this possibility:

What [Packer] is really telling us is that among American lawyers there are two main perspectives on the criminal process. He has caricatured them a bit and exaggerated their differences so we can clearly see the terms of the debate between those who hold more to one than to the other.53

This article argues that, while Packer’s aim may have been to create ideal-types, he failed in this objective since he did not accentuate the rel- evant features of the crime control and due process models to their purest

| N E W C R I M I NAL LAW R E VI E W | VOL . 11 | NO. 2 | SPR I NG 2008268

49. Id. at 90. 50. Packer, supra note 1, at 154. 51. Id. at 153. 52. Id. Although the models were not intended to be taken in the sense of “Is and

Ought,” the values which Packer held dear are clear from his son’s book Blood of the Liberals. George Packer, Blood of the Liberals (2000).

53. Griffiths, supra note 41, 362 n.14. See also Ashworth & Redmayne, supra note 27, at 38.

NCLR1102_03.qxd 5/29/08 5:01 PM Page 268

form. By remedying this error, and by correcting other flaws in Packer’s work, part III of the article constructs four ideal-types based on the values of investigative efficiency, operational efficiency, and reliability. In part IV the article stresses the importance of the distinction between empirical work and evaluation, and explores the implications of this distinction for research tools constructed upon the values of prevention of abuse of state power, equality, and crime prevention. The article seeks to demonstrate the importance of drawing a sharp distinction between the different tools which criminal justice researchers can employ. It also emphasizes the importance of distinguishing between the distinct activities of empirical work and evaluation. With these distinctions in mind, the article warns against an indiscriminate use of the word “model.” This point is particu- larly important given that Packer’s indiscriminate use of the word “model” has been perpetuated by many of his critics, as the preceding brief overview of the existing critiques of Packer’s work illustrates. Indeed, the use of “models” is now commonplace in criminal justice research. As well as advancing an alternative framework for criminal justice research—a tabular summary of which can be found in the appendix to this article— the article abstracts other general lessons for criminal justice research. It argues that a one-dimensional framework like Packer’s is insufficient and it illustrates the inadequacy of Packer’s simplistic analysis of values. So while owing to the constraints of space, the article focuses on those values which Packer attributed to one or the other of his two models, by estab- lishing the general contours of a framework for criminal justice research it paves the way for the development of further research tools founded upon such values as resource management and victims’ rights.

I I I . SUCCE E D I NG WH E R E PACKE R FAI LE D:

CON STR UCTI NG I D EAL-TYPE S BAS E D ON TH E

VALU E S OF ( I NVE STIGATIVE AN D OPE RATIONAL)

E FFICI E NCY AN D R E LIAB I LITY

Having established that Packer’s objective was to construct Weberian ideal-types of the criminal process, the aim of part III of this article is, first, to explain that Packer failed in his objective since he did not accen- tuate the relevant features of the crime control and due process models to their purest form, and, second, to seek to achieve what Packer did not,

CONSTR UCTI NG A FRAM EWOR K F OR CR I M I NAL J USTICE R ESEARCH | 269

NCLR1102_03.qxd 5/29/08 5:01 PM Page 269

by constructing ideal-types of the criminal process based on the values of investigative efficiency, operational efficiency, and reliability. Before constructing these ideal-types, however, it is necessary to do two things: to distinguish three different ways in which Packer used the word “efficiency,” and, first, to explain why, of the various values associated with the due process model, this part of the article focuses on the value of reliability.

A. Reliability

One of the strands of thought underlying the due process model is a skepticism about how the criminal sanction is used. Packer stated that this skepticism leads in turn to concern about the criminal justice process. However, since the range of possible concerns about the use of the criminal sanction is so diverse, this strand of thought does not in itself give us any guidance on what form the criminal justice process should take.54 It simply acts as a catalyst for further evaluation. It is there- fore not possible to construct an ideal-type with concern about the use of the criminal sanction as a starting-point. This is not to say that concern about the use of the criminal sanction is irrelevant. Not only might it provoke consideration of the criminal justice process, it might also be rel- evant in making the value judgments that the ideal-types open up.

The other values which Packer attributed to the due process model— reliability, prevention of abuse of state power, and equality—are different. They are not contingent on how the criminal sanction is used. Each has a substantive content and has something to say on what form the criminal justice process should take. It is immediately apparent, however, that the values of reliability, prevention of abuse of state power, and equality do not always pull in the same direction. Take, for example, a confession, obtained through torture of the suspect, but verified as true by evidence

| N E W C R I M I NAL LAW R E VI E W | VOL . 11 | NO. 2 | SPR I NG 2008270

54. Packer was skeptical about the broadening use of the criminal sanction. He argued that it was at odds with the movement of the criminal justice process towards the due process model, and that a more refined approach to the criminal sanction was therefore needed. See infra note 129. However, a person could be skeptical about the use of the crim- inal sanction and not share Packer’s views. They might, for example, argue that there is lit- tle point broadening the use of the criminal sanction unless the criminal justice process is first streamlined, with fewer procedural safeguards.

NCLR1102_03.qxd 5/29/08 5:01 PM Page 270

subsequently discovered as a result of the confession. If our sole concern is reliability, the confession ought to be relevant evidence.55 But if we are concerned purely with preventing abuses of state power, the confession ought not to be considered as evidence. Refusing to consider such evidence sends out a strong message that obtaining evidence in such a manner is unacceptable.56 The due process model is thus constructed upon values that may conflict.57 Although Packer recognized this, he seems to have shied away from the inescapable upshot that a model which is internally inconsistent is flawed.58 For this reason, this article considers the values of reliability, prevention of abuse of state power, and equality separately.

At this stage a further distinction needs to be made. The ideal-types constructed upon the values of prevention of abuse of state power and equality have a normative content. The features of these ideal-types could, in other words, be employed as evaluative tools and advanced for practi- cal implementation. For this reason, they will be discussed in part IV, when we turn to the distinction between empirical work and evaluation.

CONSTR UCTI NG A FRAM EWOR K F OR CR I M I NAL J USTICE R ESEARCH | 271

55. Packer defined reliability as “a high degree of probability in each case that factual guilt has been accurately determined.” Packer, supra note 1, at 164 (emphasis added). The value of reliability thus requires that we look at every piece of evidence which helps deter- mine factual guilt, including improperly obtained confessions that have been verified as true. It is implicit in the distinction between factual and legal guilt that sometimes we should sacrifice reliability and accept the possibility that a verdict is factually inaccurate in order to, e.g., deter abuses of state power.

56. A confession obtained by torture is involuntary, and so may be excluded under the Constitution’s due process clauses. It would also be inadmissible under the Fifth Amendment privilege against compelled self-incrimination.

57. Although the example in the main text involves only two of the three values associ- ated with the due process model, the same point applies if we take an example involving equality. A defendant is convicted on the basis of incontrovertible evidence of his guilt, but he did not have legal representation because he could not afford it. According to the dictates of reliability there is no point in spending public money on providing legal representation for a defendant whose guilt is not in doubt. But according to the dictates of equality, all defen- dants should have access to legal representation regardless of their financial ability and the cogency of the case against them.

58. Packer wrote: “the Due Process Model, although it may in the first instance be addressed to the maintenance of reliable fact-finding techniques, comes eventually to incor- porate prophylactic and deterrent rules that result in the release of the factually guilty even in cases in which blotting out the illegality would still leave an adjudicative fact finder con- vinced of the accused person’s guilt.” Packer, supra note 1, at 168. See also supra note 55.

NCLR1102_03.qxd 5/29/08 5:01 PM Page 271

By contrast, of the four ideal-types we will construct upon the values of investigative efficiency, operational efficiency, and reliability, three are founded upon non-implementable premises which mean they could not sensibly be advanced for practical implementation. Since Packer argued that only a “fanatic” would insist wholeheartedly on the values of one of his models to the exclusion of those of the other,59 this part of the article will focus on these ideal-types.

B. Efficiency

As mentioned above, it has been suggested that the crime control model should be renamed the “efficiency model.” However, this new title would also be deeply problematic, because, as closer examination of Packer’s work reveals, he used the word “efficiency” inconsistently. After attribut- ing the value of efficiency to the crime control model he defined it as the “expeditious handling of the large numbers of cases that the process ingests.”60 However, as the following extract on the police’s power of arrest demonstrates, this is not the only way in which Packer used it:

The innocent have nothing to fear. It is enough of a check on police discretion to let the dictates of police efficiency determine under what circumstances and for how long a person may be stopped and held for investigation.61

The fact that cases are dealt with expeditiously does not, in itself, afford any guarantee to the innocent. Someone who has been wrongly accused could have his case processed swiftly. The logic in this extract is that the innocent have nothing to fear because the police are efficient at discover- ing the truth. We may call this investigative efficiency (or alternatively investigative effectiveness).62 Investigative efficiency is closely related to the presumption of guilt:

| N E W C R I M I NAL LAW R E VI E W | VOL . 11 | NO. 2 | SPR I NG 2008272

59. Packer, supra note 1, at 154. 60. Id. at 164. 61. Id. at 177. 62. The label investigative effectiveness perhaps conveys the essence of the concept

more clearly than the label investigative efficiency. Nonetheless, the term investigative effi- ciency will be used in this article in order to maintain the connection with Packer’s work. For a definition of investigative efficiency, we may employ Packer’s statement that “[b]y

NCLR1102_03.qxd 5/29/08 5:01 PM Page 272

The presumption of guilt is what makes it possible for the system to deal efficiently with large numbers, as the Crime Control Model demands. The supposition is that the screening processes operated by police and prosecu- tors are reliable indicators of probable guilt.63

In this extract Packer uses the word efficiency to refer to the expeditious handling of cases. We may call this second form of efficiency operational efficiency. The extract also makes it clear that the crime control model’s demand for operational efficiency is premised upon the reliability (or investigative efficiency) of the police/prosecutorial screening processes.

In his outline of the due process model and its associated values, Packer’s failure to recognize that he attributes different meanings to the word efficiency becomes particularly glaring. This important section begins with an explanation for the due process model’s rejection of admin- istrative fact finding and its insistence on adjudicative fact finding. Having defined reliability as “a high degree of probability in each case that factual guilt has been accurately determined” and efficiency as the “expeditious handling of the large numbers of cases that the process ingests,” i.e., oper- ational efficiency, Packer then sought to illustrate the divergence between the crime control and due process models by posing the question, “how much reliability is compatible with efficiency?” He continued:

The Due Process Model insists on the prevention and elimination of mis- takes to the extent possible; the Crime Control Model accepts the proba- bility of mistakes up to the level at which they interfere with the goal of repressing crime, either because too many guilty people are escaping or, more subtly, because general awareness of the unreliability of the process leads to a decrease in the deterrent efficacy of the criminal law. In this way, reliability and efficiency are not polar opposites but rather complementary

CONSTR UCTI NG A FRAM EWOR K F OR CR I M I NAL J USTICE R ESEARCH | 273

‘efficiency’ we mean the system’s capacity to apprehend, try, convict and dispose of a high proportion of criminal offenders whose offenses become known.” Id. at 158. This differs from Packer’s definition of reliability, in that he defined reliability as “a high degree of probability in each case that factual guilt has been accurately determined,” whereas this def- inition of investigative efficiency refers to convicting “a high proportion of criminal offend- ers whose offences become known.” Id. at 164, 158. A lack of investigative efficiency will thus result in a “justice gap.” The notion of a “justice gap” was influential in the change that occurred in Dutch criminal justice policy in the 1980s. See Andrew Rutherford, Transforming Criminal Policy ch. 3 (1996).

63. Packer, supra note 1, at 160.

NCLR1102_03.qxd 5/29/08 5:01 PM Page 273

characteristics. The system is reliable because efficient; reliability becomes a matter of independent concern only when it becomes so attenuated as to impair efficiency.64

This quotation begins with a reaffirmation of the due process model’s concern to ensure that the criminal justice process operates reliably. Packer contrasts this with the crime control model, which also insists on reliabil- ity, but to a lesser extent. Only if too many guilty people are escaping, or if awareness of the unreliability of the criminal justice process is affecting the deterrent efficacy of the criminal law, does the probability of mistakes need to be addressed. He then states that, according to the crime control model, reliability and efficiency are “complementary characteristics.” The preceding expression “[i]n this way” reveals that what Packer means is that if the criminal justice process is reliable, it will be efficient at deterring peo- ple from crime. Packer then purports to restate this idea, saying that “the system is reliable because efficient.” But the fact that a system is efficient at deterring people from committing crime does not mean the system is also reliable. Efficiency in this sense provides no guarantee of reliable fact finding.65 Indeed only two sentences previously Packer stated the causal relationship the other way around—“because general awareness of the unreliability of the process leads to a decrease in the deterrent efficacy of the criminal law.” Reliability is, however, the consequence of investigative efficiency. If the criminal justice process is effective at discovering the truth, reliable verdicts will inevitably ensue. So in order to make sense of this statement we must conclude that Packer is switching to the investiga- tive efficiency sense of efficient.

Packer then completes this sentence by stating that “reliability becomes a matter of independent concern only when it becomes so attenuated as to impair efficiency.” Since reliability is always of concern to fact finders who possess investigative efficiency, Packer must here be reverting to the idea expressed two sentences previously. For the crime control model, reliability only becomes a matter of independent concern when it is so attenuated as to impair the deterrent efficacy of the criminal law.

| N E W C R I M I NAL LAW R E VI E W | VOL . 11 | NO. 2 | SPR I NG 2008274

64. Id. at 164–65. 65. For example, individuals might be deterred from committing crimes by an onerous

penalty scale, even where the criminal justice process is unreliable.

NCLR1102_03.qxd 5/29/08 5:01 PM Page 274

So although immediately before this passage Packer defined efficiency in the sense of operational efficiency, at no point in the passage does he use the word efficiency to refer to the expeditious handling of cases. Instead he uses the word primarily to refer to the way in which the unreliability of the criminal justice process will affect the efficiency with which individu- als are deterred from committing crimes, while also lapsing into using the word in the sense of investigative efficiency without offering any indica- tion that he is doing so.

Having described the perspective of the crime control model, Packer immediately goes on to say,

All of this the Due Process Model rejects. If efficiency demands short-cuts around reliability, then absolute efficiency must be rejected. The aim of this process is at least as much to protect the factually innocent as it is to con- vict the factually guilty.66

What exactly does the due process model reject? Efficiently deterring individuals from committing crime does not demand “short-cuts around reliability”; on the contrary, as Packer has already pointed out, this form of efficiency depends upon reliability. Neither does the due process model, with its emphasis on reliability, reject investigative efficiency. Packer must, of course, mean operational efficiency, i.e., according to the due process model, if the expeditious handling of cases is incompatible with having reliable fact-finding processes, then it is the expeditious handling of cases that must be sacrificed. It is therefore misleading to say that “[a]ll of this the Due Process Model rejects,” for it suggests that the due process model is rejecting efficiency in the senses attributed immediately before to the crime control model, when in fact it is only rejecting the primacy of oper- ational efficiency.

Packer finishes this section by likening the due process model to a fac- tory that cuts down on quantitative output in order to improve its quality control. The implication is that the due process model chooses to sacrifice the expeditious handling of cases in order to improve reliability, while the crime control model insists on the expeditious handling of cases at the expense of reliability. What Packer fails to recognize, however, is that the crime control model does not demand “short-cuts around reliability”; on

CONSTR UCTI NG A FRAM EWOR K F OR CR I M I NAL J USTICE R ESEARCH | 275

66. Id. at 165.

NCLR1102_03.qxd 5/29/08 5:01 PM Page 275

the contrary, as noted above, its demand for operational efficiency is premised upon the reliability of the police/prosecutorial screening process. Operational efficiency is only deemed to be a sustainable ideal because the administrative fact-finding processes possess investigative efficiency. Packer’s dialogue between the two models thus reveals two voices speaking at cross- purposes: a crime control voice that fails to articulate its model clearly, and a due process voice that has failed to understand the model of its opponent.

It is clear then that Packer’s analysis is confused as a result of his failure to distinguish these three different forms of efficiency: investigative efficiency, operational efficiency, and deterrent efficacy. So, as we turn to the task of constructing ideal-types, it is imperative that these different forms of effi- ciency are kept distinct.

C. The Investigative Efficiency and Operational Efficiency Ideal-types

According to the crime control model, since police and prosecutors possess investigative efficiency, “an early determination of probable innocence or guilt emerges.” The model maintains that once those determined to be “probably innocent are screened out,” those determined to be probably guilty can be “passed quickly through the remaining stages of the process.”67

If there is confidence in the reliability of informal administrative fact- finding activities that take place in the early stages of the criminal process, the remaining stages of the process can be relatively perfunctory without any loss in operating efficiency.68

This raises the question, when does the screening process end? As the following passage shows, Packer saw the prosecutor’s decision whether or not to prosecute as the final stage of the screening process:

[W]e will assume that the police have satisfied themselves that the original decision to arrest was sound and that the suspect is factually guilty. . . . It

| N E W C R I M I NAL LAW R E VI E W | VOL . 11 | NO. 2 | SPR I NG 2008276

67. Id. at 160. 68. Id. at 160–61. Packer again uses the word efficiency in a confusing manner. It

would not make sense for the word efficiency to be referring simply to the expeditious handling of cases, since Packer’s logic here is based on efficiency also involving some degree of reliability.

NCLR1102_03.qxd 5/29/08 5:01 PM Page 276

is clear that now the initiative must pass from the police to the prosecutor, from the expert in factual guilt to the expert in legal guilt. The decision to be made at this stage is a screening decision: should the suspect be held for further stages of the process?69

This passage also notes that, while the police are concerned with factual guilt, prosecutors are concerned with legal guilt, and so will be swayed by extra considerations, e.g., whether there is sufficient admissible evidence to secure a verdict of guilty. So, according to the crime control model, all those who emerge from the police/prosecutorial screening process and are passed on to the remaining stages of the criminal justice process70 are likely to be held to be legally guilty.

Having distinguished the police/prosecutorial screening process from the remainder of the process, it is now possible to construct an investiga- tive efficiency ideal-type of the police/prosecutorial screening process. The starting point is the crime control model’s perspective on the reliability of administrative fact finding:

[S]ubsequent processes, particularly those of a formal adjudicatory nature, are unlikely to produce as reliable fact-finding as the expert administrative process that precedes them. . . . It becomes important, then, to place as few restrictions as possible on the character of the administrative fact-finding processes and to limit restrictions to such as enhance reliability, excluding those designed for other purposes.71

To construct the investigative efficiency ideal-type, the two key features of this perspective must be accentuated to their purest form. So the first feature of this ideal-type is that administrative fact finders have perfect investigative efficiency. They will be completely reliable fact finders, and so the innocent can be sure that their innocence will emerge.72 Furthermore, since they have perfect investigative efficiency, administrative fact finders

CONSTR UCTI NG A FRAM EWOR K F OR CR I M I NAL J USTICE R ESEARCH | 277

69. Id. at 205–06. 70. It is assumed here that the screening process comprises part of the criminal justice

process. An alternative approach would be to construe the criminal justice process nar- rowly and hold that it begins only once the screening process is complete.

71. Id. at 162. 72. For example, on the issue of electronic surveillance, the crime control model states,

“Law-abiding citizens have nothing to fear.” Id. at 196. For a similar statement with regard to the power of arrest, see supra note 61.

NCLR1102_03.qxd 5/29/08 5:01 PM Page 277

will not waste time and effort employing their powers in a manner that is not constructive. For example, they will not extract confessions by torture, since such evidence is inherently unreliable. Neither will they waste time pursuing personal agendas, for example, arresting and questioning some- one simply because they have a personal vendetta against him.

The perfect investigative efficiency of the police is the basis for the second feature of this ideal-type—that no restrictions be imposed upon the admin- istrative fact-finding process. According to the crime control model, “[c]riminal investigation is a search for truth, and anything that aids the search should be encouraged.”73 Packer’s statement that “as few restrictions as possible” be placed on the work of administrative fact finders failed to reach the pure form of this perspective. Restrictions which conflict with the demands of reliability will merely hinder the work of administrative fact finders, while restrictions which “enhance reliability” are unnecessary since the police (and prosecutors) have perfect investigative efficiency and there- fore will, through self-regulation, impose such restrictions on themselves.

Since the investigative efficiency ideal-type only operates in the police/ prosecutorial screening process, a further ideal-type—which operates in the post-screening part of the process—must also be constructed. This is formed by accentuating the crime control model’s confidence in admin- istrative fact finding and the model’s insistence on operational efficiency in the post-screening part of the process to their purest forms. The result- ant ideal-type, which we can call the operational efficiency ideal-type,74 is thus premised on police and prosecutors having perfect investigative effi- ciency. In other words, it presupposes that the police/prosecutorial screening process is a perfectly reliable indicator of legal guilt, so that we can say that everyone who emerges from the screening process is legally guilty.75 As a result of this premise, the operational efficiency ideal-type

| N E W C R I M I NAL LAW R E VI E W | VOL . 11 | NO. 2 | SPR I NG 2008278

73. Packer, supra note 1, at 189. 74. See Julia Fionda, Public Prosecutors and Discretion: A Comparative Study (1995)

(advancing an operational efficiency model of prosecutorial sentencing). While this bears some resemblance to our operational efficiency ideal-type, Fionda’s model bases its demand for administrative efficiency on the pragmatic concern to “control and manage an increasing workload within the constraints of a limited workforce and budget.” Id. at 176.

75. What this does not mean, of course, is that everyone who is factually and legally guilty of a crime will emerge from the screening process, since not everyone who is factu- ally guilty will be entered into the screening process in the first place.

NCLR1102_03.qxd 5/29/08 5:01 PM Page 278

insists on operational efficiency in the remaining stages of the criminal justice process. In its purest form, this requires that there be no delays in resolving the case.

Examination of the operational efficiency ideal-type again reveals that Packer failed to accentuate the key features of the crime control model to their purest form. According to his presumption of guilt, those who emerge from the screening process are only “probably guilty,”76 while the remaining stages of the process are “relatively unimportant and should be truncated as much as possible.”77 By contrast, the operational efficiency ideal-type states that, since those who emerge from the screening process are legally guilty, the remaining stages of the process need have no bearing on the defendant’s guilt. All cases should therefore be dealt with as expe- ditiously as possible, since all that remains to be determined is the sen- tence to be served. The image of the conveyor belt thus resonates with the operational efficiency ideal-type. Once a defendant has been placed on the conveyor belt (i.e., been determined by the screening process to be legally guilty), his case should be processed and disposed of as quickly as possi- ble. What use would any obstacles on the conveyor belt be? His guilt is not in doubt.

D. The Administrative Reliability Ideal-type

Turning our attention to the value of reliability, the next task is to con- struct an administrative reliability ideal-type of the police/prosecutorial screening process. Taking as its starting point the importance the due process model purportedly attaches to the value of reliability, this ideal- type maintains that the police/prosecutorial screening process should be entirely devoted to achieving reliable determinations of legal guilt. Where the value of reliability conflicts with other values, the conflict must be resolved by meeting the demands of reliability. Prohibitions should be imposed to prevent the police from engaging in activities that will not result in pro- bative evidence being uncovered (e.g., extracting confessions by torture). Conversely, no restrictions should be imposed upon activities that will result in probative evidence being uncovered. If, for example, bugging a

CONSTR UCTI NG A FRAM EWOR K F OR CR I M I NAL J USTICE R ESEARCH | 279

76. Packer, supra note 1, at 160. 77. Id. at 163.

NCLR1102_03.qxd 5/29/08 5:01 PM Page 279

suspect’s telephone line is likely to result in cogent evidence being uncov- ered, there should be no restrictions placed on the police’s power to do so. The primary goal is for the screening process to reach the correct outcome, and fetters should not be imposed on the pursuit of this goal.

Comparing the investigative efficiency and administrative reliability ideal-types reveals certain differences between the two. The investigative efficiency ideal-type is premised upon administrative fact finders having perfect investigative efficiency, while the administrative reliability ideal- type is not—it simply attaches primacy to the value of reliability.78 The administrative reliability ideal-type consequently insists that the work of administrative fact finders be regulated, so that they are prohibited from engaging in activities that will not enhance reliability, and are directed into activities that do. According to the investigative efficiency ideal-type, on the other hand, it is unnecessary to regulate the work of administrative fact finders, since the dictates of perfect investigative efficiency mean that they are self-regulating. In spite of these differences, however, the two ideal-types share the same underlying assumption. Both regard reliability as of primary importance, and subjugate competing concerns to the demands of reliability. For example, any electronic surveillance that pro- duces probative evidence, and helps secure reliable determinations of guilt, would be endorsed by them, regardless of issues of personal privacy. As a result of this shared underlying assumption, the scope of the powers of administrative fact finders is identical in each ideal-type. Although the scope of these powers is determined by external regulation in one, and by self-regulation in the other, the boundaries are the same.

It might seem surprising that there is such a high degree of similarity between these two ideal-types, given that one was formed by accentuating the crime control model’s perspective on the police/prosecutorial screen- ing process, and the other was formed using one of the values at the heart of the due process model. The explanation lies in Packer’s failure to explain clearly the role of the value of reliability in his two models. Using

| N E W C R I M I NAL LAW R E VI E W | VOL . 11 | NO. 2 | SPR I NG 2008280

78. The investigative efficiency ideal-type is thus an ideal-type in the strong sense since its premise means it could not sensibly be advanced for practical implementation. By con- trast, since the administrative reliability ideal-type is not constructed upon such a premise, its features could theoretically form part of an agenda for practical implementation and it is therefore an ideal-type in the weak sense. See also infra part IV.

NCLR1102_03.qxd 5/29/08 5:01 PM Page 280

the factory illustration he implied that the due process model sacrifices the expeditious handling of cases for the sake of reliability, while the crime control model insists on the expeditious handling of cases at the expense of reliability. But while reliability is of concern to the due process model in the post-screening part of the criminal justice process, in the screening process itself the model is pessimistic about the possibility of reliable determinations of guilt—it rejects “informal fact-finding processes as definitive of factual guilt” and adopts “a view of informal, nonadjudicative fact finding that stresses the possibility of error”79—and so as a result attaches greater significance to other concerns. Instead, it is the crime control model, with its insistence that administrative fact finders should be allowed to employ their expertise unhindered, that attaches great significance to the value of reliability in the screening process. This is confused by Packer’s sim- ple ascription of the value of reliability to the due process model.

E. The Adversarial Reliability Ideal-type

Having constructed an ideal-type of the police/prosecutorial screening process based on the value of reliability, the next task is to construct an ideal-type based on the value of reliability which operates in the post- screening part of the process. The starting point for this is the due process model’s skepticism about the reliability of the police/prosecutorial screen- ing process. But, as we accentuate this view to its purest form, we find a problem with the meaning of the word reliable. If we take reliability in the sense of accuracy (consistently with Packer’s definition)80 and accentuate this to its purest form, we have a screening process that is perfectly inac- curate. Everyone who is innocent will be found by the screening process to be guilty, and vice versa. This leaves two possible scenarios. Either everyone who is passed on to the remainder of the criminal justice process, having been found by the screening process to be guilty, would have to be acquitted, or only those found by the screening process to be innocent should be passed on to the remainder of the process. Neither of these pos- sibilities are worthy of further consideration; the first would result in no crime ever resulting in a conviction, while the second would result in a far- cical criminal justice process.

CONSTR UCTI NG A FRAM EWOR K F OR CR I M I NAL J USTICE R ESEARCH | 281

79. Packer, supra note 1, at 163. 80. See supra note 55.

NCLR1102_03.qxd 5/29/08 5:01 PM Page 281

However, the word reliable may also be taken in the sense of trustwor- thy. In fact, when Packer described the due process model’s perspective on the police/prosecutorial screening process, he used reliability primarily in the sense of trustworthiness.81 The proponent of the due process model stresses the possibility of error and so all determinations made by adminis- trative fact finders must be examined and reconsidered. If we accentuate this to its purest form, we have a screening process that is entirely untrust- worthy. The people who emerge from the screening process are regarded as nothing more than a randomly selected group of people. The fact they have been selected by the screening process offers no indication of guilt. So in order to describe the due process model’s rejection of administrative fact finding, it is important to take reliability in the sense of trustworthi- ness. This is consistent with Packer’s depiction of the due process model. However, Packer obscured this because he consistently defined reliability in the sense of accuracy.

It is now possible to construct an adversarial reliability ideal-type of the post-screening part of the criminal process. Packer wrote that the due process model “resembles a factory that has to devote a substantial part of its input to quality control.”82 Again, he failed to accentuate this to its purest form. According to the adversarial reliability ideal-type, the police/ prosecutorial screening process is completely untrustworthy, so emphasizing the expeditious handling of cases is no longer appropriate.83 The post- screening part of the process must be entirely devoted to scrutinizing and testing the case against the defendant in order to achieve reliable determinations of legal guilt.84

Comparing the administrative and adversarial reliability ideal-types further illustrates Packer’s failure to explain clearly the role of the value of reliability in his two models. One would have expected these ideal-types

| N E W C R I M I NAL LAW R E VI E W | VOL . 11 | NO. 2 | SPR I NG 2008282

81. For example, he wrote of “the distrust of fact-finding processes that animates the Due Process Model,” and he explained that the model’s “rationale [for excluding improp- erly obtained confession evidence] is not that the confession is untrustworthy. . . .” Packer, supra note 1, at 164, 191.

82. Id. at 165. 83. Of course, this is not to give a license for delay. But this is very different from the

ideal-type operational efficiency model’s emphasis on passing the case through the remain- der of the criminal justice process as quickly as possible. Cf. supra note 41.

84. Here the accuracy and trustworthy senses of reliability coincide.

NCLR1102_03.qxd 5/29/08 5:01 PM Page 282

to be complementary, since both were constructed upon the same value— reliability. But in fact the two make uneasy bedfellows. While the adminis- trative reliability ideal-type endorses the wholehearted pursuit by adminis- trative fact finders of reliable outcomes, the adversarial reliability ideal-type premises its insistence on securing reliable outcomes in the post-screening part of the criminal justice process on the untrustworthiness of the police/prosecutorial screening process. The adversarial reliability ideal- type thus rejects the possibility of reliable administrative fact finding, while the administrative reliability ideal-type embraces it.

Closer examination of Packer’s application of the due process model to the police/prosecutorial screening process confirms that, in spite of his statement that the value of reliability is central to the due process model, reliability has little role to play in the due process model’s perspective on the police/prosecutorial screening process. Packer’s outline of the due process model’s perspective on arrests for investigation,85 detention, and interrogation after a “lawful” arrest,86 electronic surveillance,87 illegally secured evidence,88 access to counsel,89 and the decision to charge90 is instead dominated by a desire to protect privacy and the dignity and invi- olability of the individual, and a concern to prevent the state from abusing its power. For example,91 when describing the due process model’s perspec- tive on arrests for investigation, Packer stated that a stringent test must be satisfied before any arrest is made. This is based on a concern not to “[open] the door to the possibility of grave abuse,” and to protect “per- sonal privacy” and “the dignity and inviolability of the individual.”92

Similarly, on the issue of electronic surveillance, Packer stated that “the right of privacy . . . cannot be forced to give way to the asserted exigencies of law enforcement.”93 And third, when considering access to counsel, he

CONSTR UCTI NG A FRAM EWOR K F OR CR I M I NAL J USTICE R ESEARCH | 283

85. Packer, supra note 1, at 179–81. 86. Id. at 190–92. 87. Id. at 196–97. 88. Id. at 200. 89. Id. at 203. 90. Id. at 207–09. 91. The same is true of Packer’s description of the due process model’s perspective on

detention and interrogation after a lawful arrest, id. at 190–92, illegally secured evidence, id. at 200, and on the decision to charge, id. at 207–09.

92. Id. at 179. 93. Id. at 196.

NCLR1102_03.qxd 5/29/08 5:01 PM Page 283

said that the suspect must be immediately apprised of his right to remain silent and to have a lawyer, he must promptly be given access to a lawyer, and failing the presence of a lawyer he must not be subjected to police interrogation, for “there is no moment in the criminal process when the disparity in resources between the state and the accused is greater than at the moment of arrest. There is every opportunity for overreaching and abuse on the part of the police.”94 Indeed, where these concerns conflict with the value of reliability, they are to be given priority, so, for example, the due process model insists that improperly obtained confessions and illegally secured evidence should be inadmissible regardless of their proba- tiveness.95 So although Packer attributed the value of reliability to the due process model, it is clear that, in the police/prosecutorial screening process, the model attaches greater significance to competing concerns. Packer thus obscured the role that reliability has to play by simultaneously attributing it to the due process model and subordinating it to other con- cerns in his application of this model, without offering any explanation of his reasons for doing so. What is more, the due process model actually rejects the possibility of administrative fact finders arriving at reliable determinations of guilt:

People are notoriously poor observers of disturbing events—the more emotion-arousing the context, the greater the possibility that recollection will be incorrect; confessions and admissions by persons in police custody may be induced by physical or psychological coercion so that the police end up hearing what the suspect thinks they want to hear rather than the truth; witnesses may be animated by a bias or interest that no one would trouble to discover except one specially charged with protecting the interests of the accused (as the police are not).96

It is incoherent to attribute the value of reliability to the due process model and then to apply the model to administrative fact finding, if the model’s insistence on reliability is premised upon the impossibility of administrative fact finders arriving at reliable determinations of guilt. Applying the model, with its insistence on reliability, to administrative fact finding must either be futile, for the model itself asserts that reliable

| N E W C R I M I NAL LAW R E VI E W | VOL . 11 | NO. 2 | SPR I NG 2008284

94. Id. at 203. 95. See supra note 55. 96. Packer, supra note 1, at 163.

NCLR1102_03.qxd 5/29/08 5:01 PM Page 284

administrative fact finding is impossible, or, if it is not futile, the very premise on which the model bases its demand for reliability in the post- screening part of the criminal justice process is challenged.

F. Using These Ideal-types: An Example

Having constructed these ideal-types of the criminal process, it is now possible to offer an example of how they might be used. Packer ascribed a uniform threshold of probability to the word reliability.97 In reality, how- ever, parties at different stages of the criminal justice process are asked to make quite different assessments of a case. A prosecutor, for example, may only decide to charge a suspect if the charge is supported by probable cause.98 Finders of fact in criminal trials, by contrast, employ the consti- tutionally required standard of “beyond a reasonable doubt.”99 The “prob- able cause” and “beyond a reasonable doubt” thresholds are obviously not synonymous.100 Given the lower threshold applied by prosecutors, it fol- lows that suspects may be required to stand trial when in fact the case against them does not fulfill the criminal standard of proof. Suppose that, against this background, the government directed the courts to speed up criminal trials in order to ensure a swifter turnover of cases (with the resourcing of the courts remaining the same).

Packer used the image of a factory to imply that the crime control model sacrifices reliability in order to process cases expeditiously, while the due process model insists on ensuring reliable determinations of guilt at the expense of operational efficiency. As was explained previously, this

CONSTR UCTI NG A FRAM EWOR K F OR CR I M I NAL J USTICE R ESEARCH | 285

97. He defined reliability as “a high degree of probability in each case that factual guilt has been accurately determined.” Id. at 164 (emphasis added).

98. American Bar Association Standards for Criminal Justice: Prosecution Function and Defense Function, standard 3–3.9 (3d ed. 1993).

99. In re Winship, 397 U.S. 358 (1970). 1o0. Although Packer ascribed a uniform threshold of probability to the value of relia-

bility, he did implicitly recognize that prosecutors apply a lower threshold than the crimi- nal standard of proof, but without considering the implications this has for his models. Packer, supra note 1, at 160. Note also that it is not only prosecutors and finders of fact in criminal trials that have to make assessments of a case. The police must make an assess- ment of a case when deciding whether or not to exercise their power of arrest. For exam- ple, an officer may arrest without a warrant a suspect whom he has reasonable cause to believe has committed a felony.

NCLR1102_03.qxd 5/29/08 5:01 PM Page 285

contrast is flawed. Both models only recognize the expeditious handling of cases to the extent that this is compatible with the dictates of reliability. However, in reality there will often be instances where the demands of operational efficiency and reliability conflict. Comparing such situations to each of our ideal-types aids analysis of these situations and helps clarify discussion. In our example, for instance, the government has shown a concern to promote the expeditious handling of cases. This is akin to the emphasis placed on operational efficiency by the operational efficiency ideal-type. Unlike this construct, however, the background to the govern- ment’s decision is a screening process that is not an effective indicator of legal guilt. Furthermore, if the resources available to the courts remain the same, then in order to ensure a swifter turnover of cases, the scrutiny with which cases are examined must be reduced. The government has thus decided to reduce the thoroughness of the adversarial fact-finding process, even though the background to our example is similar to the premise upon which the adversarial reliability ideal-type is built. Of course, one would have to take into account the factors behind the government’s deci- sion. Perhaps it was motivated by a concern to reduce the length of time individuals are remanded in custody pending trial, or perhaps it wanted to send out a message that they are cracking down on crime. In order to eval- uate the decision these background factors would have to be considered. The key point, however, is to note how the operational efficiency and adversarial reliability ideal-types open up this further evaluative discourse.

If we compare this methodological approach to the one advocated by Packer, the shortcomings of his proposed analytical framework are further exposed. Packer encouraged the use of a spectrum, with his crime control and due process models (purportedly) at either end. On this approach, our example is simply an instance of the crime control model’s concern for the expeditious handling of cases being given preeminence over the values under- lying the due process model. There are two problems with this reasoning. First, it fails to question whether the reliability of the criminal justice process has in fact been diminished. Instead of examining the reliability of the police/prosecutorial screening process, it simply assumes that the post- screening part of the process ought to be devoted to reliability. Second, by reducing all decisions to a simple conflict between the values of the crime control and due process models, Packer’s framework offers only one possible explanation for reducing the reliability of the post- screening part of the process—to increase the expeditiousness with

| N E W C R I M I NAL LAW R E VI E W | VOL . 11 | NO. 2 | SPR I NG 2008286

NCLR1102_03.qxd 5/29/08 5:01 PM Page 286

which cases are handled—when in fact other values might have formed the basis for the decision. For example, the reliability of the post-screening part of the process might have been reduced in order to try and deter abuses of state power101—a possibility which Packer implicitly rejected by attributing the value of prevention of abuse of state power to the same model as the value of reliability. Packer’s proposed framework is thus inad- equate for proper consideration of the issues involved.

IV. TH E N E E D TO D I STI NG U I S H E M PI R ICAL WOR K

FROM EVALUATION

Of the ideal-types constructed in part III of this article, three were founded upon premises which mean that they could not sensibly form part of an agenda for practical implementation.102 They may thus be described as ideal- types in a strong sense.103 They are theoretical constructs which are non- implementable and which could not be regarded as prescriptions of what ought to exist.104 As we turn our attention in this part of the article to the val- ues of prevention of abuse of state power, equality, and crime prevention, we will find an important difference with the tools which we will construct using these values. For, as well as being useful in empirical work, we will find that the features of these tools might also be advanced as evaluative standards.

At first glance this possibility seems to blur the distinction, outlined in part II above, which Weber drew between ideal-types and ideals:

[W]e should emphasise that the idea of an ethical imperative, of a “model” of what “ought” to exist is to be carefully distinguished from the analytical construct, which is “ideal” in the strictly logical sense of the term.105

CONSTR UCTI NG A FRAM EWOR K F OR CR I M I NAL J USTICE R ESEARCH | 287

101. See supra note 55. 1o2. The exception is the administrative reliability ideal-type. See supra note 78. 1o3. The word “ideal” is here being used in the abstract theoretical sense, not the ethi-

cal imperative sense. See infra note 105. 1o4. Note that while an ideal-type in the strong sense could not be regarded as a pre-

scription of what ought to exist, it might be plausible to use it to describe an ideal. Weber gave the concept of a free market as an example of an ideal-type. Weber, supra note 48, at 90. A perfectly free market cannot be achieved, but moving towards a free market might nonetheless be described as an ideal.

1o5. Id. at 91–92. The distinction Weber draws here between the abstract theoretical and ethical imperative senses of the word “ideal” is also found in the Oxford English

NCLR1102_03.qxd 5/29/08 5:01 PM Page 287

Weber himself commented that writers often do blur the distinction between the two.106 Using the example of Christianity, he claimed that ostensibly ideal-types often “contain what, from the point of view of the expositor, should be and what to him is ‘essential’ in Christianity because it is enduringly valuable. . . . In this sense, however, the ‘ideas’ are naturally no longer purely logical auxiliary devices, no longer concepts with which reality is compared, but ideals by which it is evaluatively judged.”107 He concluded:

[T]he elementary duty of scientific self-control and the only way to avoid seri- ous and foolish blunders requires a sharp, precise distinction between the logically comparative analysis of reality by ideal-types in the logical sense and the value-judgment of reality on the basis of ideals.108

The crucial point that emerges from this is that a researcher should not confuse the two distinct activities of empirical work and evaluation. On this reasoning, however, it remains possible that a theoretical construct may, if appropriate, be used both as an ideal-type when engaged in an empirical study, and as an ideal when engaged in evaluative work—a point which is confirmed by Weber’s recognition that such constructs may sometimes, perhaps unwittingly, be “ideal-types not only in the logical sense but also in the practical sense.”109 Before using such a construct as an evaluative tool, however, it is important to heed Weber’s warning that a failure to distinguish between “the ‘idea’ in the sense of the ideal [and] the ‘idea’ in the sense of the ‘ideal-type’ . . . on the one hand hampers the

| N E W C R I M I NAL LAW R E VI E W | VOL . 11 | NO. 2 | SPR I NG 2008288

Dictionary’s entry for “ideal.” Strand B.2 of the definition states that an ideal is “Something existing only as a mental conception,” whereas strand B.1.a. states that an ideal is “an object to be realized or aimed at.” Oxford English Dictionary 615–16 (2d ed. 1989). The terms “ideal-type” and “non-ideal-type” use the word in the abstract theoretical sense. See supra note 103, infra note 118.

1o6. Weber wrote: “As fundamental as this distinction is in principle, the confusion of these two basically different meanings of the term ‘idea’ appears with extraordinary fre- quency in historical writings.” Weber, supra note 48, at 98.

1o7. Id. at 97–98. 1o8. Id. at 98. 1o9. Id. at 97. Weber stated that an ideal-type is a “mental construct [that] cannot be

found empirically anywhere in reality.” Id. at 90. It follows that if a weak ideal-type is used in evaluative work as a prescription of what ought to exist, and is realized, then it must cease to be an ideal-type (in the abstract theoretical sense).

NCLR1102_03.qxd 5/29/08 5:01 PM Page 288

value-judgment and on the other, strives to free itself from the responsi- bility for its own judgment.”110 So, if the constructs founded upon the val- ues of prevention of abuse of state power, equality, and crime prevention are to be used as evaluative tools, it is essential that such use first of all be justified. To fail to do this would be to replace reasoned argument with mere assertion and so result in an impoverished evaluation.

With this in mind, this part of the article discusses the use of research tools founded upon the values of prevention of abuse of state power, equality, and crime prevention in empirical and evaluative work. In par- ticular, it explores the implications of Weber’s warning for the use of these constructs in evaluative work. First, though, it introduces the work of Swedish scholar Nils Jareborg, which at first glance bears some superficial resemblance to Packer’s analytical framework.

A. Jareborg’s Defensive Model of, and Offensive Approach to, Criminal Law Policy

Swedish scholar Nils Jareborg has outlined a defensive model of, and an offensive approach to, criminal law policy.111 Unlike Packer’s due process model, the defensive model is based purely on the concepts of prevention of abuse of state power and the primacy of the individual. Jareborg described the model as “an ‘ideal type’ model in a Weberian sense.”112 It con- sists of principles for criminalization (respect for which means the criminal code lists a set of “socially sanctioned basic moral demands” and so acquires a “value-expressive function”), procedural safeguards, and principles for sen- tencing (which recognize that “the courts cannot have an independent func- tion in ‘combatting’ crime”).113 The offensive approach, meanwhile, is not a conceptual construct, and so is not an ideal-type.114 Rather, it is an attempt

CONSTR UCTI NG A FRAM EWOR K F OR CR I M I NAL J USTICE R ESEARCH | 289

110. Id. at 98. 111. Nils Jareborg, What Kind of Criminal Law Do We Want?, in Beware of

Punishment: On the Utility and Futility of Criminal Law 17 (Anikka Snare ed. 1995). 112. Id. at 20. 113. Id. at 22–23. 114. Jareborg also insisted that the offensive approach ought not to be regarded as a

model, because “important parts of the defensive model are kept or only slightly modified— there is no general rejection of the anchorage in a Rechtstaat ideology. . . . [I]t is not (yet) possible to formulate an offensive model, i.e., to describe the offensive approach in isolation from the defensive model.” Jareborg, supra note 111, at 24.

NCLR1102_03.qxd 5/29/08 5:01 PM Page 289

to describe a “strong ideological counter-current” which “is undermining [the] dominance” of the ideology of the defensive model.115 This emerging attitude to criminal law policy “regards the criminal justice system as an at least potential repertoire of methods for solution of social or societal prob- lems”116 and so is “best described in terms of its methods and conse- quences.”117 The offensive approach is thus an attempt to characterize a particular strategy towards criminal justice, which adopts a mixture of pre- vailing attitudes and some ideological leaning, without achieving a level of coherent articulation found in a theorist’s model. It is a type of approach to criminal justice, but a non-ideal-type.118

The offensive approach’s benevolent view of state power contrasts with the perspective of the defensive model:

All criminal law aims at protecting the interests of individuals, collective or public interests, or state interests, by using threats of punishment and by using execution of punishment to make the threat credible. But the defen- sive model . . . also aims at protecting individuals against power abuse, against abuse of state power, excessive repression in legal or illegal forms, as well as against abuse of private, informal power, of which “lynch justice” is the most obvious form. . . . [T]he defensive model does not deny that the crim- inal law has a social task or function, but its criminal law policy implies that criminal law is meant to be an obstacle, not only for offenders, but also for authorities and politicians.119

| N E W C R I M I NAL LAW R E VI E W | VOL . 11 | NO. 2 | SPR I NG 2008290

115. Id. at 23. 116. Id. at 24–25. 117. Id. at 26. 118. Weber distinguished two senses of the word “ideal,” one a strictly logical sense, the

other an ethical imperative sense. See supra note 105. The expression “non-ideal-type” uses the first sense of the word. Since Jareborg’s offensive approach to criminal law policy was an attempt to describe a “strong ideological counter-current,” it may be regarded by some as an ethical imperative. But it is a description of an emerging attitude, not a conceptual construct, and so is not “ideal” in the abstract theoretical sense.

119. Jareborg, supra note 111, at 21, 24. Cameron argues that “the emphasis in the Swedish system is placed on preventive, legislative safeguards on abuse of rights,” later attributing this in part to the fact that the “courts in Sweden . . . unreservedly accept the primacy of the prin- ciple of parliamentary democracy.” Iain Cameron, Protection of Constitutional Rights in Sweden [1997] P.L. 488, 502, 504. It might be argued that this emphasis on preventive, leg- islative safeguards informs Jareborg’s defensive model. The model asserts, for example, that “the point of having a criminal justice system as a response to unwanted behaviour is . . . to protect the offending individual from power abuse.” Jareborg, supra note 111, at 24.

NCLR1102_03.qxd 5/29/08 5:01 PM Page 290

As this extract indicates, the defensive model “does not regard state power as necessarily benevolent.”120 It sees the state itself as a “potential enemy,”121 hence its concern to impose obstacles “for authorities and politicians.” The defensive model and offensive approach’s contrasting views of state power mean that Jareborg’s framework has a wider ambit than Packer’s. It encompasses the work of legislators as well as the work of the executive. For example, one of the principles for criminalization insisted upon by the defensive model is that crimes must be defined by statutory law and that the definitions must be understandable and determinate, whereas one of the methods of the offensive approach is the use of lin- guistically indeterminate definitions of crimes. This contrasts with Packer’s framework, which focuses on the stages of the criminal process from initial arrest through to appeals and collateral attack, and the poten- tial for individuals involved in the process to abuse the powers vested in them.122 As explained previously, his framework would not regard a lin- guistically indeterminate crime definition as a contest between the values of the crime control and due process models. It would be considered a chal- lenge to the assumptions upon which the framework is constructed. In other words, elements of Jareborg’s defensive model comprise the rule of law concerns at the heart of the so-called common ground between Packer’s models. The offensive approach to criminal law policy challenges these.

B. Prevention of Abuse of State Power

Packer’s attribution of the value of prevention of abuse of state power to the due process model was based on the fact that the criminal justice process is “coercive, restricting, and demeaning” and that criminal

CONSTR UCTI NG A FRAM EWOR K F OR CR I M I NAL J USTICE R ESEARCH | 291

120. Jareborg, supra note 111, at 22. 121. Id. at 25. 122. For example, the due process model stresses the scope for police officers to abuse

their power of arrest and employ it in a discriminatory manner. Packer, supra note 1, at 179–81. It stresses the possibility of suspects being detained and interrogated improperly. Id. at 190–92. It worries that “an unscrupulous policeman or prosecutor” could use elec- tronic surveillance “to pry into the private lives of people almost at will.” Id. at 196–97. It stresses that police officers might resort to illegal searches in order to obtain evidence. Id at 200. And it states that a prosecutor “with nobody looking over his shoulder” might charge a suspect even when there is insufficient evidence. Id. at 207–09.

NCLR1102_03.qxd 5/29/08 5:01 PM Page 291

penalties are the “heaviest deprivation that government can inflict on the individual.”123 Closely related to the value of prevention of abuse of state power is the concept of the primacy of the individual.124 According to this concept, the state should act in a manner that respects the auton- omy, liberty, and rights of every individual. When it abuses its power it violates these demands.

In part III we saw that much of Packer’s application of the due process model to the police/prosecutorial screening process centers on preventing abuses of state power. The same is true of the post-screening part of the process. For example, the model insists that since “a person accused of crime is not a criminal,” it is an abuse of state power to detain an accused person pending trial unless the orderly processes of criminal justice are under threat.125 On the issue of guilty pleas the model states that “no kind of pressure, either by the prosecutor or by the judge, should be brought to bear on a defendant to induce him to plead guilty. . . . [I]t can only defeat the ends of the system to penalize a defendant for insisting on a trial.”126

And on the issue of appeals, the model urges that “the discretion to allow bail pending appeal not be manipulated coercively to discourage the pur- suit of any appeal that has a semblance of merit.”127 Moreover, an appeal should result in the reversal of a conviction whenever an organ of the state has abused its powers—“When an appellate court finds it necessary to cas- tigate the conduct of the police, the prosecutor, or the trial court, but fails to reverse a conviction, it simply breeds disrespect for the very standards it is trying to affirm.”128

The features of the due process model which are based on the value of prevention of abuse of state power may be used as an ideal-type to analyze and expound trends in the criminal justice process, in the same way as Jareborg’s defensive model. In fact, with each issue he considered (pre-trial detention, guilty pleas, appeals, etc.), Packer compared the contempora- neous situation in the United States with the perspectives of the crime control and due process models, and concluded that the criminal justice

| N E W C R I M I NAL LAW R E VI E W | VOL . 11 | NO. 2 | SPR I NG 2008292

123. Id. at 165–66. 124. Id. at 165. 125. Id. at 214–18. 126. Id. at 223–25. 127. Id. at 231. 128. Id. at 230–32.

NCLR1102_03.qxd 5/29/08 5:01 PM Page 292

process was moving increasingly towards the due process model—an important stepping stone in the argument of his book.129 However, in con- trast to the investigative efficiency, operational efficiency, and adversarial reliability ideal-types constructed in part III, Jareborg’s defensive model and the features of the due process model which are based on the value of prevention of abuse of state power could also be regarded as prescriptions of what ought to exist, and so could be used as evaluative tools.130 The due process model, for example, states that a prosecutor should never take the lead in proposing or suggesting a compromise plea, for this could place coercive pressure on an accused person, while Jareborg’s defensive model insists that all persons accused of crime should be provided with access to independent legal counsel. These are normative standards. The defensive model and, to the extent it is based on the value of prevention of abuse of state power, the due process model are thus ideal-types in a weak sense. They are ideal-types which may also be used as ideals.

As explained above, if the defensive model or the features of the due process model based on the value of prevention of abuse of state power were to be used in evaluative work, such use would first have to be justi- fied. This is especially important given that the defensive and due process models’ concerns about abuses of state power are grounded in a liberal outlook—the ideological base of the defensive model is classical criminal law and the philosophies of the Enlightenment and its views on human

CONSTR UCTI NG A FRAM EWOR K F OR CR I M I NAL J USTICE R ESEARCH | 293

129. Id. at 239. This conclusion, at the end of part two, paved the way for the ulti- mate argument of the book. In part three Packer went on to argue that the criminal sanction was being resorted to too indiscriminately, which created the situation where “its processes are being forced to conform to values that reduce its efficiency [while] we place heavier and heavier demands on those processes.” Id. at 365. Packer thus con- cluded, “The process cannot function effectively unless the subject matter with which it deals is appropriately shaped to take advantage of its strengths and to minimize its weaknesses. The prospect of spending billions of dollars . . . on improving the capacity of the nation’s system of criminal justice to deal with gamblers, narcotics addicts, pros- titutes, homosexuals, abortionists, and other producers and consumers of illegal goods and services would be seen for the absurdity that it is if we were not so inured to simi- lar spectacles. Our national talent runs much more to how-to-do-it than to what-to-do. We sorely need to redress the balance, to ask ‘what’ and ‘why’ before we ask ‘how.’” Id. at 366.

130. Indeed, immediately after stating that the defensive model is an ideal-type, Jareborg added that “[i]n many respects, it is also meant to be ‘ideal.’” Jareborg, supra note 111, at 20.

NCLR1102_03.qxd 5/29/08 5:01 PM Page 293

nature,131 while the due process model’s concern about abuse of state power stemmed from the U.S. Constitution,132 the civil rights movement,133 and the Warren Court of the day134—which many might not share. Jareborg’s offensive approach usefully outlines another possible approach to criminal justice:

From its point of view, the most serious criticism is not that the system is, in some respects, unjust or lacking in legal certainty but that it is too inef- ficient, not “rational” enough (in the sense of “goal rational”). Prevention of harm or wrong-doing is the dominating perspective. . . . [F]or the defen- sive model the state is a potential enemy. For the offensive approach, the state is an ally. The possibility of power abuse is not completely forgotten: a Rechtsstaat ideology is still the background. But the important thing is to show results. . . . From its own point of view, the offensive approach is legit- imate only if it is efficient in preventing crime.135

It would be misleading to suggest that goal fulfillment is the sole domain of the offensive approach. For while a proponent of the offensive approach may prioritize crime prevention, and so might criticize the crim- inal justice process if it is not “efficient in preventing crime,” proponents of the defensive model believe that the criminal justice process should strive to prevent the state from abusing its power, and so might criticize the criminal justice process if it is ineffective in achieving this goal. It is equally incorrect to say that it is only the offensive approach that is con- cerned with “methods,” for while the defensive model may be described as consisting of principles and procedural safeguards, the virtue of these prin- ciples and procedural safeguards is deemed to lie in the fact that their imposition is the method by which abuses of state power are prevented from occurring. The difference between, on the one hand, the defensive model and the due process model (insofar as it is based on the prevention of abuse of state power) and, on the other hand, the offensive approach is

| N E W C R I M I NAL LAW R E VI E W | VOL . 11 | NO. 2 | SPR I NG 2008294

131. Id. at 20. By classical criminal law, Jareborg meant the “kind of criminal law that began to dominate in the beginning of the 19th century, especially in what could roughly be described as German- and French-dominated parts of Europe.” Id.

132. Packer, supra note 1, at 173. 133. Id. at 243–44. 134. Id. at 239–40. 135. Jareborg, supra note 111, at 25, 26, 27–28.

NCLR1102_03.qxd 5/29/08 5:01 PM Page 294

that they disagree on three fundamental questions. The first question asks what constitutes an abuse of state power. While the defensive and due process models emphasize the autonomy, liberty, and rights of every indi- vidual, the offensive approach emphasizes crime prevention, and so is will- ing to condone more intrusive uses of state power in pursuit of this aim. This leads it to view the question of what constitutes an abuse of state power differently. The second question asks what safeguards are necessary to prevent such abuses from occurring. Both the defensive and due process models are mindful of the potential for abuse and so insist on strict safe- guards. The offensive approach, by contrast, regards the state as an ally. Its benevolent view of state power means that, while the possibility of abuse is not forgotten, it is willing to entrust the state with wider powers in order to give it greater leeway in its efforts to tackle crime. The third question asks whether such safeguards should be enacted if doing so restricts our pursuit of the apprehension and conviction of offenders. According to the defensive model and (insofar as it is based on prevention of abuse of state power) the due process model, all possible steps should be taken to prevent abuses of state power, even if this places restrictions on the apprehension and conviction of offenders. The evaluative priorities associated with these models are constructed on this basis. By contrast, the offensive approach’s emphasis on crime prevention leads it to reject safeguards which restrict our pursuit of the apprehension and conviction of offenders.

For example, the due process model maintains that to arrest someone in order to attempt to compile a case against him/her would amount to an abuse of state power: “the police should not arrest unless information in their hands at that time seems likely, subject to the vicissitudes of the liti- gation process, to provide a case that will result in a conviction.”136 For many people, however, this is too stringent. They may share the view of the crime control model, that it will often be necessary to arrest suspects in order to investigate offenses.137 But this does not mean that they are willing to tolerate abuses of state power. Rather, their different priorities lead them to adopt a different view of what constitutes an abuse, which in

CONSTR UCTI NG A FRAM EWOR K F OR CR I M I NAL J USTICE R ESEARCH | 295

136. Packer, supra note 1, at 190. 137. The crime control model argues that since “the best source of information is usually

the suspect himself,” the police should not “be expected to solve crimes by independent investigation alone.” Id. at 187. They should therefore be able to “interrogate the suspect in private before he has a chance to fabricate a story or to decide that he will not cooperate.”

NCLR1102_03.qxd 5/29/08 5:01 PM Page 295

turn leads to a different view of what safeguards are necessary to prevent (what they regard as) abuses from occurring.

Similarly, one of the defensive model’s procedural safeguards is the placing of the burden of proof on the prosecutor. In In re Winship, the Supreme Court held that the Due Process Clause of the Constitution “protects the accused against conviction except upon proof beyond a reasonable doubt of every fact necessary to constitute the crime with which he is charged.”138 Many would agree with this description, urging that it should always be for the state to prove the guilt of a person suspected of having committed an offense. The power and resources of the state are immense in comparison to individuals, and to presume that those prosecuted for an offense are guilty would place an oppressive burden on defendants and considerable power in the hands of those who decide on prosecution. However, others would say that to always prohibit the state from placing the burden of proof on defendants is too inflexible. Some matters may be far easier for one party to prove than the other (e.g., possession of a license), or it may simply be more expedient to require the defendant to disprove an element of the offense than to require the prosecution to prove it. The Supreme Court’s accept- ance that a defendant may constitutionally bear the burden of proving an affirmative defense139 illustrates that not all agree that the prosecutor should invariably bear the burden of proof.140

| N E W C R I M I NAL LAW R E VI E W | VOL . 11 | NO. 2 | SPR I NG 2008296

Id. The length of time for which the suspect may be held is “the length of time, given all the circumstances, during which it is reasonable to suppose that legitimate techniques of interrogation may be expected to produce useful information or that extrinsic investiga- tion may be expected to produce convincing proof either of the suspect’s innocence or of his guilt.” Id. at 188.

138. In re Winship, 397 U.S. 358, 364 (1970). Per Justice Brennan, on behalf of a majority of the Court.

139. See Patterson v. New York, 432 U.S. 197 (1977). 140. In a similar vein, the European Court of Human Rights has accepted that article 6(2),

which enshrines the presumption of innocence, does not prohibit the reversal of the burden of proof, provided that reverse onus provisions are “[confined] within reasonable limits which take into account the importance of what is at stake and maintain the rights of the defence.” Salabiaku v. France, 13 E.H.R.R. 379, ¶ 28 (1988). See also Andrew Ashworth & Meredith Blake, The Presumption of Innocence in English Criminal Law, [1996] Crim. L.R. 306 (find- ing that 40 percent of offenses triable in the Crown Court in England and Wales place a bur- den of proof on the defendant). This is in spite of Lord Sankey LC’s celebrated description of the presumption of innocence as the “golden thread” of English criminal law. Woolmington v. DPP, [1935] A.C. 462 (H.L.).

NCLR1102_03.qxd 5/29/08 5:01 PM Page 296

To take one final example from the jurisprudence of the House of Lords, another of the defensive model’s procedural safeguards is the pro- hibition on retroactive application of law to the detriment of the accused. While many would agree that such a prohibition is an important restraint, there are others who would argue that the state should have the power to apply law retroactively if and when the necessity arises. In the controver- sial case R. v. R.,141 their Lordships held that the pronouncement of Sir Matthew Hale in 1736142 that a husband could not be guilty of raping his wife had ceased to apply, and so upheld the conviction of the defendant for the rape of his wife. A number of different interpretations of this deci- sion are possible: that the House of Lords did not apply the law retrospec- tively at all; that the House of Lords’ decision was a reasonably foreseeable step in the evolution of the law; or that the House of Lords usurped Parliament’s role by abolishing the marital rape exception.143 The impor- tant point for present purposes is to notice how contentious the question whether the House of Lords’ verdict represented an abuse of state power is. How one answers this question will depend not only on how one interprets the decision itself, but also, for example, on whether one regards it as an abuse to afford a reasonably foreseeable development in the law retrospective effect.

CONSTR UCTI NG A FRAM EWOR K F OR CR I M I NAL J USTICE R ESEARCH | 297

141. R. v. R., [1992] 1 A.C. 599. 142. 1 Matthew Hale, History of the Pleas of the Crown 629 (1st ed. 1736). 143. Their Lordships did not consider themselves to be applying the law retrospectively,

with Lord Keith (with the unanimous support of the House) declaring that “in modern times the supposed marital exemption in rape forms no part of the law of England.” R. v. R., [1992] 1 A.C. 599, 623. The difficulty with this reasoning is that “the [House of Lords] did not hold that Hale had misstated the law . . . but that (this must have been on some unspecified day before R had forcible intercourse with his wife in October 1989) the law had changed as no longer compatible with modern conditions.” David Ormerod, Smith & Hogan Criminal Law: Cases & Materials 728 (9th ed. 2006). Although the European Court of Human Rights subsequently declined to find that the decision violated article 7 E.C.H.R., it is submitted that the decision went beyond what the Strasbourg Court described as “evolution, which was consistent with the very essence of the offence.” SW and CR v. United Kingdom, 21 E.H.R.R. 363 ¶ 41 (1996). However distasteful one finds it, before the House of Lords’ decision the exception for marital rape existed (albeit sub- ject to some erosion in previous cases). After the decision it ceased to exist. The definition of rape had thus been rewritten. See also R. v. C. (Barry), [2004] E.W.C.A. Crim. 292 (convicting C in 2002 of raping his wife in 1970, at which time they were married).

NCLR1102_03.qxd 5/29/08 5:01 PM Page 297

C. Equality

Racial discrimination formed no part of Packer’s outline of the principle of equality. But he did not overlook this issue—to have done so would have been quite a stark omission given Packer’s strong personal views144

and the fact that The Limits of the Criminal Sanction was written against the backdrop of the struggle over civil rights. Instead he dealt with it as one example of how the state can abuse its power.145 As this demonstrates, there is some overlap between the principles of prevention of abuse of state power and equality, in that it is an abuse for the state to unjustifiably dis- criminate against particular (groups of ) people. But the state may also abuse its power without being guilty of unjustifiable discrimination. And abuse of state power is not the only reason that people may not be treated equally. Packer, for example, applied the principle of equality to an issue which is not caused by abuse of state power—financial inability. The two principles should, therefore, be dealt with separately.

George Fletcher commented, “[e]quality is at once the simplest and the most complex idea that shapes the evolution of the law.”146 As this indi- cates, how the principle of equality should be applied is contentious. What is clear, however, is that equality of treatment in the criminal justice process cannot mean uniformity of treatment; suspects cannot all be investigated by the same officers, tried by the same judge and jury, and represented by the same counsel. Packer’s decision to concentrate the due process model’s application of the principle of equality on financial (in)ability was based on the decision of the U.S. Supreme Court in Griffin v. Illinois that “there can be no equal justice where the kind of trial a man gets depends on the amount of money he has.”147 So, for example, the due process model argues that, since there is a large class of persons who cannot afford any bail payment, “a system that makes pre-trial freedom conditional on

| N E W C R I M I NAL LAW R E VI E W | VOL . 11 | NO. 2 | SPR I NG 2008298

144. George Packer recalls that his father felt strongly about racial discrimination: “I was fairly alert to the plight of Negroes; the subject was much discussed in our house, in strong moral tones. I gathered that Negroes had been treated unfairly and we owed them something.” Packer, supra note 52, at 238.

145. See Packer, supra note 1, at 180 (outlining the due process model’s perspective on the police’s power of arrest); id. at 239–46 (commenting on the apparent trend towards the due process model).

146. George Fletcher, Basic Concepts of Legal Thought 121 (1996). 147. Griffin v. Illinois, 351 U.S. 12, 19 (1956).

NCLR1102_03.qxd 5/29/08 5:01 PM Page 298

financial ability is discriminatory.”148 And it asserts that the right of appeal must not be restricted by financial inability: “[i]f the appellant cannot afford to pay a filing fee, it must be waived; if he cannot afford to buy a transcript, it must be given to him; if he cannot afford to hire a lawyer, he must be assigned one.”149

These propositions are clearly contentious. The proponent of the crime control model, for example, views financial ability as a fundamental aspect of the bail system. Being able to set bail at a level that the defendant can- not afford enables the magistrate to “select those people who—for what- ever reason—ought not to be at liberty pending trial, and to see to it that they are not.”150 On the right of appeal, the crime control model states that the costs of filing an appeal, buying a transcript, and having legal repre- sentation should only be waived/defrayed if “the appeal is screened and determined to be probably meritorious.”151 While this places a limit on the exercise of the right of appeal by the financially unable, it is deemed to be sufficient to satisfy the demands of equality. It might even be argued that to waive/defray these costs for the financially unable, but to impose them on the financially able, is in itself discriminatory, for it attaches a cost to the exercise of the right for some but not for others.

The features of the due process model that are derived from the prin- ciple of equality comprise an ideal-type in the weak sense. They may be used in empirical work. By comparing these features to reality, they can be used to analyze the extent to which the criminal justice process reflects the liberal approach to the concern that justice is not dependent upon finan- cial ability. And they may also be used in evaluative work. But, for the rea- sons explained previously, before the features of the due process model that are based on the principle of equality can be employed as evaluative standards, their usage in this way must be justified. The researcher must

CONSTR UCTI NG A FRAM EWOR K F OR CR I M I NAL J USTICE R ESEARCH | 299

148. Packer, supra note 1, at 217. 149. Id. at 231. 150. Id. at 213. The difference of opinion fundamentally flows from the fact that the due

process model holds that there should be a right to pre-trial liberty pending a formal adju- dication of guilt, id. at 215, whereas the crime control model argues that, since a formal charge “has behind it a double assurance of reliability” (the judgment of the police officer and the prosecutor), then “[f ]or all practical purposes, the defendant is a criminal,” and so there “is no reason for him to go free.” Id. at 211.

151. Id. at 229.

NCLR1102_03.qxd 5/29/08 5:01 PM Page 299

not only justify invoking the principle of equality, but also explain why the features of the due process model which are based on the principle of equality are suitable norms for evaluating whether suspects are treated equally, particularly since there will almost always be debate over what the principle of equality requires in a given situation. Furthermore, the scope of the inquiry must be delineated. For example, Packer’s due process model only applies the principle of equality to financial (in)ability, and so may only be used to evaluate the extent to which equal justice depends on money. Equality may also be a concern in other areas, such as race or gen- der, and an inquiry along these lines would require another set of evalua- tive standards.

When expounding his own theoretical framework for evaluating the criminal justice process, Andrew Ashworth commented that Packer’s two models are “unsatisfactory in their failure to propose any normative or evaluative criteria.”152 As we have seen, however, Packer’s purpose in this part of his book was to demonstrate that the criminal justice process in the United States was moving towards the due process model.153 In other words, he was engaged in empirical work, not evaluative work. Furthermore, as has been explained, the features of the due process model based on the values of prevention of abuse of state power and equality could be employed as eval- uative criteria, if their use in this way was first of all justified. Packer did not embark on this task because he was engaged in an empirical study, but if it were to be undertaken the result would be a set of evaluative standards. The distance between Packer’s models and Ashworth’s proposed normative framework is thus not as great as it at first appears.

D. Crime Prevention

Packer stated that the proposition “the repression of criminal conduct is by far the most important function to be performed by the criminal

| N E W C R I M I NAL LAW R E VI E W | VOL . 11 | NO. 2 | SPR I NG 2008300

152. Andrew Ashworth, The Criminal Process: An Evaluative Study 29 (2d ed. 1998). Cf. Ashworth & Redmayne, supra note 27.

153. See supra note 129. When Packer arrived at this conclusion, he qualified it by say- ing, “[i]n theory at least.” Packer, supra note 1, at 239. This presumably alludes to the pos- sible disparity between the “law in the books” and the “law in action.” Like the due process/crime control dichotomy, the dichotomy between the “law in the books” and the “law in action” has been challenged by Doreen McBarnet. McBarnet, supra note 32.

NCLR1102_03.qxd 5/29/08 5:01 PM Page 300

process”154 underlies the crime control model. However, closer examination reveals that the features of the crime control model are not contingent upon the belief that repressing criminal conduct is the most important function of the criminal process. Although the crime control model insists that the work of administrative fact finders be given “special weight,” with “as few restrictions as possible” being placed on it,155 this need not stem from a concern to repress crime. Someone who shares the crime control model’s confidence in administrative fact finding will not only regard the potential for abuse of power as slight,156 but will also view the police/pros- ecutorial screening process as the best available opportunity to obtain pro- bative evidence which will help secure a reliable determination of legal guilt. Nor need the crime control model’s insistence on operational effi- ciency in the post-screening part of the process157 be motivated by a belief that primacy should be attached to the repression of crime. Managerialist concerns would also suggest that if the police/prosecutorial screening process is a reliable indicator of legal guilt, the operational efficiency of the post-screening process should be increased so as to avoid an unnecessary duplication of resources. This reasoning is consistent with the analysis presented in part III of this article. The investigative and operational efficiency ideal-types were constructed by accentuating the features of

CONSTR UCTI NG A FRAM EWOR K F OR CR I M I NAL J USTICE R ESEARCH | 301

154. Packer, supra note 1, at 158. 155. Id. at 162. For example, the crime control model insists that the police should have

extremely broad powers of arrest since “the dictates of police efficiency” provide sufficient regulation, and that access to a lawyer should be refused because “it is absolutely necessary for the police to question the suspect at this point without undue interference.” Id. at 177, 202.

156. The crime control model asserts, for example, that “the innocent have nothing to fear” from broad police powers of arrest, since the dictates of efficiency are sufficient reg- ulation, and that “[l]aw-abiding citizens have nothing to fear” from electronic surveillance since “law enforcement has neither the time nor the inclination to build up files of infor- mation about activity that is not criminal.” Id. at 196.

157. For example, it takes a restrictive view of pre-trial liberty because “[t]he vast majority of persons charged with crime are factually guilty. . . . Just because the assem- bly line cannot move fast enough for him to be immediately disposed of is no reason for him to go free.” Id. at 211. It also encourages guilty pleas because “[i]f the earlier stages of the process have functioned as they should” there should be only a very small num- ber of cases in which there is “genuine doubt about the factual guilt of the defendant.” Id. at 222.

NCLR1102_03.qxd 5/29/08 5:01 PM Page 301

the crime control model to their purest form. Yet neither ideal-type pur- ports to attach primacy to the repression of criminal conduct. Instead, both are premised on administrative fact finders having perfect investiga- tive efficiency. Since there is no necessary connection between the features of the crime control model and the belief that the criminal process should attach primacy to the repression of criminal conduct, the crime control model is not apt to be used to analyze the extent to which the criminal process reflects this belief.

By contrast, crime prevention is at the heart of the offensive approach. However, it is important to recognize that the offensive approach focuses on one particular strategy for achieving crime prevention. Six of the nine methods listed by Jareborg concern criminalization, concentrating on an increased willingness to use the criminal sanction in ways that are broader and more severe.158 Its focus is thus on the work of legislators, with the threat of penal sanctions forming their chief weapon.159 So while a researcher might wish to analyze the extent to which contemporaneous criminal law policy resembles the offensive approach by comparing the criminal justice process to the methods and consequences Jareborg identi- fied, he or she should remember that the fact that current criminal law policy does not resemble the offensive approach does not mean that crime prevention is not a priority. It might be that a different crime prevention strategy is being employed.

This also has implications for the use of the offensive approach in eval- uative work.160 A researcher might seek to evaluate legislation by assessing

| N E W C R I M I NAL LAW R E VI E W | VOL . 11 | NO. 2 | SPR I NG 2008302

158. These six methods are: a threat against or a violation of a legitimate interest of value is regarded as a sufficient reason for criminalization; so too is culpability; emphasis shifts from offenses against individuals to offenses against the state machinery or an anonymous public; criminalizations increasingly concern potentially dangerous deeds or deeds which are peripheral to caused harm; crime definitions are linguistically indeterminate; and decriminalizations are rare. Jareborg, supra note 111, at 26.

159. The other three methods could also be employed by legislators: severer legislation can increase repression within the criminal justice process, criminal procedure can be rationalized through new legislation, and legislators can encourage the demonization of offenders.

160. Jareborg himself makes it clear that he opposes the offensive approach. Jareborg, supra note 111, at 32–33. As a result, much of his description of the offensive approach is at least impliedly critical. A proponent of the offensive approach would describe its methods and consequences quite differently.

NCLR1102_03.qxd 5/29/08 5:01 PM Page 302

the extent to which it employs the methods of the offensive approach.161

However, such evaluation will be impoverished if the researcher does not first of all do three things. First, grounds must be given for believing that crime will be prevented if the methods of the offensive approach are employed. Second, attempting to tackle crime in the manner envisaged by the offensive approach, rather than in some other way, must be justified. And third, where attaching primacy to repressing criminal conduct raises issues relating to other values, this primacy must be justified. This might involve the researcher justifying the view that seeking to tackle crime in this particular way does not constitute an abuse of state power, justifying the view that safeguards which others regard as necessary to prevent abuses of state power from occurring are in fact unnecessary, and/or the researcher justifying the view that safeguards that are conceded would have some value in helping to prevent abuses of state power would nonetheless be an unjustified restriction on the pursuit of the apprehen- sion and conviction of offenders. Alternatively (or additionally), it might involve the researcher justifying the view that seeking to tackle crime in this particular way is not at odds with the demands of the value of equal- ity, or the researcher justifying the view that the demands of the value of equality are insufficient to warrant qualifying the pursuit of the appre- hension and conviction of offenders.

V. CONCLUS ION

Any analysis or evaluation of criminal justice policy will be flawed without a detailed appreciation of the way that competing policies are constructed and assessed. This article has accordingly warned against an indiscriminate use of the word “model” to describe the different tools which may be

CONSTR UCTI NG A FRAM EWOR K F OR CR I M I NAL J USTICE R ESEARCH | 303

161. A proponent of the offensive approach might also carry out evaluation by study- ing the effect a piece of legislation has on the crime level. However, the fact that crime lev- els are unaffected does not necessarily mean that the legislation in some way fails to employ the methods of the offensive approach. For example, it might be because inadequate resourcing has hampered the enforcement of the legislation, or because there has been a lack of enthusiasm amongst enforcement agencies for the legislation. Note too that, even if crime levels were to fall, it does not follow that the new legislation is responsible. It would have to be shown that the reduction was caused by the new legislation as opposed to some other factor(s).

NCLR1102_03.qxd 5/29/08 5:01 PM Page 303

employed in criminal justice research. It has distinguished three separate tools: strong ideal-types, weak ideal-types, and non-ideal-types. A strong ideal-type is a theoretical construct. It may be used in empirical work for analysis and exposition, but, since it could not sensibly be regarded as a prescription of what to exist, is not apt to be used in evaluative work. A weak ideal-type is also a conceptual construct, but, as well as being used in empirical work, it may also be employed in evaluative work as an ideal. A non-ideal-type (such as the offensive approach to criminal law policy) is not a conceptual construct; it is a description of an actual strategy or approach. Like a weak ideal-type, it may be used in both empirical and evaluative work. This article has also shown that the distinct activities of empirical work and evaluation must be distinguished. In particular, it has shown that, if a weak ideal-type or a non-ideal-type is to be used in eval- uative work, its use in this way must first be justified. This involves the researcher (1) explicating and defending his interpretation of contentious concepts, (2) demonstrating that the provisions that he supports are not only capable of achieving their goal but are also to be preferred to other possible methods of doing so, and, (3) where enacting the provision involves imposing/removing restrictions on our efforts to apprehend and convict offenders, justifying the imposition/removal of these restric- tions. If the use of a weak ideal-type or a non-ideal-type as an ideal is not justified in this way, the result will be an impoverished approach to the consideration of policy which fails to take responsibility for its own judgment.

In addition, it is also possible to abstract from the foregoing discussion two other general lessons for criminal justice research. First, this article has demonstrated that criminal justice policy cannot be satisfactorily under- stood if the analytical framework used is a spectrum. A spectrum such as the one Packer purported to create is a one-dimensional device. It states that there are two sets of values which are polar opposites, and that as adherence to one set of values increases so adherence to the other set necessarily dimin- ishes. It has been shown, however, that the challenge in criminal justice pol- icy is not to balance the competing demands of two value systems; it is to balance the competing demands of many different values. The analytical framework employed must therefore be multidimensional. Second, the ana- lytical framework employed must have regard for the different ways in which values are held. A simple “yes/no” approach—either a value is a pri- ority or it isn’t—is inadequate. Such an approach fails to recognize that

| N E W C R I M I NAL LAW R E VI E W | VOL . 11 | NO. 2 | SPR I NG 2008304

NCLR1102_03.qxd 5/29/08 5:01 PM Page 304

many values are contentious. One person’s view of what a particular value requires might not be the same as that of other people. Moreover, there might be differences over how best to meet the demands of a value. The “yes/no” approach also obscures the fact that a person might identify as pri- orities a number of values which, on occasion, conflict. In particular, a value may still be a priority even if its demands are not always met.

Packer’s exposition of the crime control and due process models is infested by a failure to appreciate the way that competing policies are con- structed and assessed. He failed to distinguish between strong and weak ideal-types and between empirical work and evaluative work, he advocat- ed the use of a spectrum an as analytical framework, and he employed a simple “yes/no” approach to describe the packages of values held by dif- ferent parties within the criminal justice process. Explaining how Packer committed each of these errors provides a useful way of summarizing the various arguments that have been advanced. We will work through them in reverse order.

First, Packer (purportedly) constructed his two models by dichotomiz- ing the various values competing for priority in the American criminal process into two separate value systems. He thus adopted a yes/no approach to these values—certain values could be attributed to the due process model and not the crime control model, and vice versa. Ironically, Packer himself appeared to concede the inadequacy of this approach. As part II explained, he outlined certain rule of law concerns which constitute “com- mon ground” between his models, thus undermining his assertion that the crime control model insists on the value of efficiency to the exclusion of the values associated with the due process model. This article has shown how this simple yes/no approach to values blights Packer’s work.

Packer ascribed the value of reliability to the due process model. Yet it has little part to play in his application of the due process model to the police/prosecutorial screening process; other values are not only given greater prominence than, but are also given priority over, the demands of reliability. He also suggested that it is of only secondary importance to the crime control model, even though this model attaches great significance to the value of reliability in the police/prosecutorial screening process, insist- ing that administrative fact finders should be allowed to work unhindered. And although Packer states that the crime control model demands short- cuts around reliability in order to increase operational efficiency, we have seen that it is in fact the case that the crime control model premises its

CONSTR UCTI NG A FRAM EWOR K F OR CR I M I NAL J USTICE R ESEARCH | 305

NCLR1102_03.qxd 5/29/08 5:01 PM Page 305

demand for operational efficiency upon the police/prosecutorial screening process operating reliably.

The second value which Packer ascribed to the due process model is prevention of abuse of state power. According to this model, “the criminal process must . . . be subjected to controls that prevent it from operating with maximal [operational] efficiency”; the model would happily accept “a substantial diminution in the [operational] efficiency with which the criminal process operates in the interest of preventing official oppression of the individual.”162 The implication is that the crime control model, with its insistence that “primary attention be paid to the efficiency with which the criminal process operates,”163 refuses to impose any such controls upon the criminal justice process. It is, however, too simplistic to say that the crime control model subjugates the concern to prevent abuses of state power to the dictates of efficiency in this way. For a start, it is at odds with the so-called “common ground” between the models. Moreover, propo- nents of the crime control model view state power with greater benevo- lence than the due process model, and so do not share the latter’s liberal concerns. This means that they are happy to afford greater leeway to the state and that they have different views on the prior question of what con- stitutes an abuse of state power. For Packer to assert that the crime control model subordinates the need to prevent abuses of state power to the dic- tates of efficiency is to mistakenly assume that there is uniform under- standing on both sides of what safeguards are necessary to prevent abuses of state power from occurring and what constitutes an abuse of state power.

The value of equality, the final value Packer attributed to the due process model, may also be understood in diverse ways. The due process model applies the principle to the issue of financial (in)ability. What is required to ensure that financial ability does not affect the “kind of trial a man gets”164 is also contentious; the level of intervention that the due process model considers necessary is greater than that deemed necessary by the crime control model. This is not to say, however, that the due process model attaches greater significance to the value of equality than the crime control model. Rather the two models have different views on what is suf- ficient to prevent discrimination on the basis of financial ability.

| N E W C R I M I NAL LAW R E VI E W | VOL . 11 | NO. 2 | SPR I NG 2008306

162. Packer, supra note 1, at 166. 163. Id. at 158. 164. Griffin v. Illinois, 351 U.S. 12, 19 (1956).

NCLR1102_03.qxd 5/29/08 5:01 PM Page 306

Packer attributed the value of efficiency to the crime control model, and said that the due process model rejects “absolute efficiency” if it “demands short-cuts around reliability.”165 This statement on the role of efficiency is misleading, however, due to Packer’s failure to distinguish, and clarify the roles of, three different forms of efficiency. The first of these, investigative efficiency, is welcomed not only by the crime control model, but also—with its emphasis on reliability—by the due process model. The second form of efficiency is the deterrent efficacy of the crim- inal law. This form of efficiency is not in competition with the demands of reliability—on the contrary, it is dependent upon it. General awareness that the criminal justice process is unreliable will lead to a decrease in the deterrent efficacy of the criminal law. It is the third form of efficiency, operational efficiency, that the due process model rejects when it demands shortcuts around reliability. The crime control model, however, does not insist on the expeditious handling of cases at the expense of reliability. Its demand for operational efficiency in the post-screening part of the process is premised upon its faith in the police/prosecutorial screening process; according to the crime control model operational efficiency is only a sus- tainable ideal if the administrative fact finding processes have investigative efficiency.

Packer stated that the goal of the crime control model is the repression of crime. We have seen, however, that the features of the model are in fact shaped by its great confidence in administrative fact finding, and that someone who shares this confidence might agree with the features of the model even if his or her primary concern is not the repression of crime.

The simple yes/no approach proves to be a defective foundation for the analytical framework which Packer built upon it. It obscures the fact that Packer’s due process model identifies as priorities values which, on occa- sion, conflict. It also means that there is no scope within Packer’s frame- work for examining whether different opinions over a policy decision emanate from different views of what a particular value requires. Suppose, for example, a decision is made to repeal a provision which the due process model regards as essential to prevent the abuse of state power. According to Packer’s framework the reason for the decision against the values of the due process model must be a concern for efficiency and crime prevention.

CONSTR UCTI NG A FRAM EWOR K F OR CR I M I NAL J USTICE R ESEARCH | 307

165. Packer, supra note 1, at 165.

NCLR1102_03.qxd 5/29/08 5:01 PM Page 307

But this is too simplistic.166 It could be the case that those in favor of the decision have different views either on what constitutes an abuse of state power or on what is necessary to prevent such an abuse from occurring. In their eyes the provision might be an unnecessary restriction upon the power of the state, one which could potentially hinder efforts to achieve other goals. Any analysis that ignores this, and explains the decision sim- ply as an example of the values of efficiency and crime prevention being given priority over preventing abuses of state power, will be distorted.

Packer’s second error was to advocate the use of a spectrum to analyze the criminal process. We have seen that a spectrum is inadequate as an analytical tool. The challenge in criminal justice policy is to balance the competing demands of many different values, and so the analytical framework employed must be multidimensional. The spectrum Packer purported to construct reduces all policy decisions to a conflict between two value systems—those of the crime control and due process models. This one-dimensional approach leads to the mistaken assumption that any policy decision which is based upon the values of one model must necessarily have a detrimental effect upon the values of the other model. But this is not the case. Suppose, for example, that a decision is made which increases the operational efficiency of the post- screening part of the criminal justice process. Packer’s spectrum tells us that the reliability of the process as a whole must have been diminished. But to say this is to simply assume that the post-screening part of the process ought to be devoted to reliability without examining first the police/prosecutorial screening process. It might be the case that the reliability of the whole process has not been adversely affected at all. The one-dimensional approach also leads to the mistaken assumption that any decision which detrimentally affects one of the values Packer associated with the due process model must be motivated by a concern for the values he attributed to the crime control model. But such a decision might not be based on the values he associated with the crime control model. For example, a decision which reduces the

| N E W C R I M I NAL LAW R E VI E W | VOL . 11 | NO. 2 | SPR I NG 2008308

166. One might seek to defend Packer’s spectrum by arguing that those making the deci- sion to repeal the provision might be primarily concerned with preventing crime. There are two flaws in such reasoning. First, since they do not regard the provision as essential to pre- vent abuses of state power, those making the decision would not, in their eyes, be prioritiz- ing crime prevention over preventing abuses of state power as Packer’s spectrum suggests. Second, while the provision might be repealed to help crime prevention, the provision might equally be repealed in order to help efforts to pursue other values, e.g., efforts to obtain probative evidence which leads to reliable determinations of guilt.

NCLR1102_03.qxd 5/29/08 5:01 PM Page 308

reliability of the criminal justice process in order to prevent abuses of state power, or to achieve greater equality, would not be concerned with the values associated with the crime control model at all.

Finally, Packer’s work fails to clearly distinguish between strong and weak ideal-types and between empirical work and evaluative work. Although Packer’s aim was to construct something like ideal-types, he failed in this task by not accentuating the viewpoints of either of his models to their purest form. Part III of this article succeeded where Packer failed, by con- structing four ideal-types—the operational efficiency, investigative efficiency, adversarial reliability and administrative reliability ideal-types—based on the values of operational efficiency, investigative efficiency and reliability. The factual premises upon which the first three of these ideal-types are based mean that they are ideal-types in the strong sense. Packer’s objective appears to have been to construct strong ideal-types, since he stated that his models could not plausibly be regarded as goals to strive towards, describing anyone who subscribed to the values of one model to the exclusion of the values of the other as a “fanatic.”167 However, this statement obscures the fact that, to the extent that it is based upon the principles of prevention of abuse of state power and equality, the due process model is an ideal-type in the weak sense. The features of the model derived from these two principles may not only be used in empirical work, but could also be used as evaluative standards. Packer thus obscured the necessity of justifying the use of these features as evaluative standards before engaging in evaluative work.

Although, after decades of critical scrutiny, Herbert Packer’s attempt to devise a theoretical framework for analyzing criminal justice policies may not exert the influence it once held, there remain lessons to be learned through exploring the fundamental approach to modeling criminal justice policy pioneered by Packer. First, a distinction must be drawn between the different tools which may be used to analyze criminal justice policies (strong ideal-types, weak ideal-types, non-ideal-types) and between the activities of empirical work and evaluation. Second, since the challenge in criminal justice policy is to balance the demands of many competing val- ues, a multidimensional analytical framework must be used. Third, the simple yes/no approach to describing the ways in which different values are held must be abandoned. By paying attention to these lessons, a fuller understanding of criminal justice policy can be achieved.

CONSTR UCTI NG A FRAM EWOR K F OR CR I M I NAL J USTICE R ESEARCH | 309

167. Packer, supra note 1, at 154.

NCLR1102_03.qxd 5/29/08 5:01 PM Page 309

| N E W C R I M I NAL LAW R E VI E W | VOL . 11 | NO. 2 | SPR I NG 2008310

A P

P E

N D

IX :

TH E

D IF

F E

R E

N T

TO O

LS W

H IC

H C

R IM

IN A

L JU

S TI

C E

R E

S E

A R

C H

E R

S M

IG H

T E

M P

LO Y

A N

D T

H E

IR U

S E

S

IN E

M P

IR IC

A L

A N

D E

V A

LU A

TI V

E W

O R

K

H ow

is it

H

ow m

ig ht

it b

e us

ed fo

r th

e pu

rp os

es o

f

W ha

t i s

it? co

ns tru

ct ed

? em

pi ric

al w

or k?

ev al

ua tio

n?

S tr

on g

D es

cr ip

tio n

A th

eo re

tic al

c on

st ru

ct Th

e on

e- si

de d

ac ce

nt ua

tio n

C om

pa rin

g th

e th

eo re

tic al

It

co ul

d no

t s en

si bl

y id

ea l-

w hi

ch in

p ra

ct ic

al te

rm s

of th

e ke

y fe

at ur

e( s)

o f a

co ns

tr uc

t t o

cu rr

en t

be a

dv an

ce d

fo r

ty pe

is n

on -im

pl em

en ta

bl e

pa rt

ic ul

ar p

er sp

ec tiv

e po

lic ie

s an

d tr

en ds

m ay

pr

ac tic

al

ai d

an al

ys is

a nd

e xp

os iti

on im

pl em

en ta

tio n

E xa

m pl

e Th

e op

er at

io na

l Th

e po

lic e/

pr os

ec ut

or ia

l In

c on

ju nc

tio n

w ith

th e

W hi

le it

m ay

b e

us ed

ef fic

ie nc

y sc

re en

in g

pr oc

es s

is d

ee m

ed ad

ve rs

ar ia

l r el

ia bi

lit y

to o

pe n

up a

n id

ea l-t

yp e

to p

ro vi

de a

p er

fe ct

ly r

el ia

bl e

id ea

l-t yp

e, it

m ay

b e

ev al

ua tiv

e di

sc ou

rs e,

in

di ca

tio n

of le

ga l g

ui lt.

T he

us

ed to

a na

ly ze

it is

n ot

a pt

fo r

us e

as

id ea

l-t yp

e th

er ef

or e

in si

st s

on si

tu at

io ns

w he

re th

e an

id ea

l b ec

au se

op

er at

io na

l e ffi

ci en

cy in

th e

de m

an ds

o f o

pe ra

tio na

l it

is fo

un de

d on

a

po st

-s cr

ee ni

ng p

ar t o

f t he

ef

fic ie

nc y

an d

re lia

bi lit

y no

n- im

pl em

en ta

bl e

pr oc

es s

co nf

lic t

pr em

is e

W ea

k D

es cr

ip tio

n A

th eo

re tic

al c

on st

ru ct

Th e

on e-

si de

d ac

ce nt

ua tio

n C

om pa

rin g

th e

th eo

re tic

al Th

e co

ns tr

uc t m

ig ht

id

ea l-

w hi

ch c

ou ld

a ls

o be

of th

e ke

y fe

at ur

e( s)

o f a

co

ns tr

uc t t

o cu

rr en

t be

a dv

an ce

d as

a n

ty pe

ad va

nc ed

a s

a pa

rt ic

ul ar

p er

sp ec

tiv e

po lic

ie s

an d

tre nd

s m

ay a

id id

ea l,

an d

its fe

at ur

es pr

es cr

ip tio

n of

w ha

t an

al ys

is a

nd e

xp os

iti on

us

ed a

s ev

al ua

tiv e

ou gh

t t o

ex is

t st

an da

rd s

E xa

m pl

e Ja

re bo

rg ’s

B as

ed o

n th

e lib

er al

no tio

n It

m ay

b e

us ed

fo r

It m

ay b

e pr

om ot

ed de

fe ns

iv e

th at

s ta

te p

ow er

m us

t b e

an al

yz in

g an

d ex

po un

di ng

as a

n id

ea l,

an d

m od

el vi

ew ed

w ith

s us

pi ci

on , i

t th

e ex

te nt

to w

hi ch

its

in di

vi du

al fe

at ur

es ou

tli ne

s w

ha t i

s co

ns id

er ed

cr im

in al

la w

p ol

ic y

m ay

b e

us ed

a s

ne ce

ss ar

y to

p ro

te ct

re fle

ct s

th e

lib er

al c

on ce

rn ev

al ua

tiv e

to ol

s in

di vi

du al

s fro

m a

bu se

s of

to

p ro

te ct

in di

vi du

al s

fro m

NCLR1102_03.qxd 5/29/08 5:01 PM Page 310

st at

e po

w er

ab us

es o

f s ta

te p

ow er

N on

– D

es cr

ip tio

n It

is n

ot a

th eo

re tic

al B

y id

en tif

yi ng

a nd

e xp

ou nd

in g

B y

co m

pa rin

g it

to c

ur re

nt A

p ar

tic ul

ar s

tr at

eg y

id ea

l- co

ns tr

uc t,

bu t r

at he

r th

e ke

y fe

at ur

es o

f a p

ar tic

ul ar

po lic

ie s

an d

tre nd

s, a

or

a pp

ro ac

h m

ig ht

ty pe

a de

sc rip

tio n

of

st ra

te gy

o r

ap pr

oa ch

no n-

id ea

l-t yp

e m

ay b

e be

e nd

or se

d, a

nd

a pa

rt ic

ul ar

s tr

at eg

y us

ed fo

r an

al ys

is a

nd its

fe at

ur es

u se

d or

a pp

ro ac

h ex

po si

tio n

to e

va lu

at e

cu rr

en t

(h is

to ric

al o

r pr

op os

ed )

po lic

ie s

an d

tr en

ds

E xa

m pl

e Ja

re bo

rg ’s

A

d es

cr ip

tio n

of a

n em

er gi

ng It

m ay

b e

us ed

fo r

A r

es ea

rc he

r m

ig ht

of fe

ns iv

e ap

pr oa

ch to

c rim

in al

la w

an al

yz in

g an

d ex

po un

di ng

en do

rs e

th e

st ra

te gy

ap pr

oa ch

po lic

y w

hi ch

v ie

w s

th e

th e

ex te

nt to

w hi

ch of

th e

of fe

ns iv

e st

at e

as a

n al

ly a

nd is

cr im

in al

la w

p ol

ic y

in a

ap pr

oa ch

a nd

u se

ge ar

ed to

w ar

ds e

ffi ci

en t

pa rt

ic ul

ar s

et tin

g ad

op ts

its fe

at ur

es to

cr im

e pr

ev en

tio n

th e

st ra

te gy

o f t

he ev

al ua

te c

rim in

al of

fe ns

iv e

ap pr

oa ch

la w

p ol

ic y

CONSTR UCTI NG A FRAM EWOR K F OR CR I M I NAL J USTICE R ESEARCH | 311

NCLR1102_03.qxd 5/29/08 5:01 PM Page 311

Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.