1 See Robertson, John A., Rough Justice: Perspectives on Lower Criminal Courts (Boston: Little, Brown & Co. 1974).Google Scholar

2 Dash, Cracks in the Foundation of Criminal Justice, 46 Ill. L. Rev. 385 (1951).Google Scholar

3 Id. at 387.Google Scholar

4 Id. at 388–89.Google Scholar

5 Id. at 389. The rapid pace was apparently not a consequence of caseload pressures. Most of the judges observed by Dash arrived between 9:45 and 10:00) A.M. and left by 10:45. One judge arrived at 9:45 and left by 10:30, after having disposed of 20 cases. Id. Google Scholar

6 Nutter, , The Quality of Justice in Misdemeanor Arraignment Courts, 53 J. Crim. L., Criminology & Police Sci. 215, 216 (1962).Google Scholar

7 President's Comm'n on Law Enforcement and Administration of Justice, Task Force Report: The Courts 30–36 (1967). Other studies at about the same period indicate some significant variations. A 1967 study reports a guilty plea rate of only 25% in Cleveland. Katz, Municipal Courts: Another Urban Ill, 20 Case W. Res. L. Rev. 87 (1968). In a 1969–70 Detroit study, 22% of the defendants were represented by counsel and only 54% pleaded guilty. Donald I. Warren, Justice in Recorder's Court: An Analysis of Misdemeanor Cases in Detroit (1971), reprinted in Robertson, supra note 1, at 326, 330–32. A Boston study found 67% of defendants represented by counsel (among those charged with “real crime”) and a “high” (but unspecified) rate of contested trials. Stephen R. Bing & S. Stephen Rosenfeld, The Quality of Justice in the Lower Criminal Courts of Metropolitan Boston (1970), reprinted in Robertson, supra note 1, at 259, 264, 272.Google Scholar

8 Mileski, , Courtroom Encounters: An Observation of a Lower Criminal Court, 5 Law & Soc'y Rev. 473, 475 (1971).Google Scholar

9 Id. at 486–87.Google Scholar

10 Id. at 493.Google Scholar

11 Id. at 480–85.Google Scholar

12 Id. at 514.Google Scholar

13 407 U.S. 25 (1972).Google Scholar

14 S. Krantz et al., Right to Counsel in Criminal Cases: The Mandate of Argersinger v. Hamlin 4–5, 694–701 (1976); Portman, , Gideon's Trumpet Blows for Misdemeanants–-Argersinger, the Decision and Its Impact, 14 Santa Clara Law. 1, 25 n.25 (1973); Comment, Effect of Argersinger on the Municipal Court of Toledo, Ohio, 4 U. Tol. L. Rev. 577 (1973).Google Scholar

15 See Alfini, & Doan, , A New Perspective on Misdemeanor Justice, 60 Judicature 425, 430 (1977); Ingranam, The Impact of Argersinger–-One Year Later, 8 Law & Soc'y Rev. 615 (1974).Google Scholar

16 Feeley, Malcolm, The Process Is the Punishment 9–10 (New York: Russell Sage Foundation, 1979).Google Scholar

17 Id. at 127.Google Scholar

18 Ryan, John P., Adjudication and Sentencing in a Misdemeanor Court: The Outcome Is the Punishment, in Nat'l Institute of Justice, Misdemeanor Courts: Policy Concerns and Research Perspectives 93, 108, 134 n. 14 (US. Dep't of Justice, 1981).Google Scholar

19 Id. at 100.Google Scholar

20 Id. at 106. These figures may slightly understate the trial rate in Columbus, because some cases were scheduled directly for a bench trial without passing through the pretrial stage from which Ryan drew his sample. See id. at 133 n.2.Google Scholar

21 For discussion of the literature, see Schulhofer, , Is Plea Bargaining Inevitable 97 Harv. L. Rev. 1037, 1041 & n.7 (1984).Google Scholar

22 Id. at 1041–44.Google Scholar

23 Feeley, , supra note 16, at 280, 292–94. For a similar caveat concerning the dangers of introducing formal due process conceptions without basic changes in underlying organization and resources, see Handler, , The Juvenile Court and the Adversary System: Problems of Function and Form, 1965 Wis. L. Rev. 7.Google Scholar

24 Feeley, supra note 16, at 291.Google Scholar

25 Id. at 280.Google Scholar

26 Alschuler, , Implementing the Criminal Defendant's Right to Trial: Alternatives to the Plea Bargaining System, 50 U. Chi. L. Rev. 931, 952, 955–56 (1983).Google Scholar

27 Ryan, supra note 18, at 129–30. Ryan attributed this difference, in part, to the contrast between the court culture of the older eastern cities (e.g., New Haven) and the progressive cities of the Midwest (e.g., Columbus). He noted that the court cultures of Pittsburgh and Minneapolis described by Martin Levin suggested a similar contrast. Id. at 131–32. From this perspective, one would expect the court culture of Philadelphia to approach the New Haven–Pittsburgh pattern, but this turns out not to be the case.Google Scholar

28 Id. at 106.Google Scholar

29 Alschuler, supra note 26, at 956.Google Scholar

30 The comparable figures for New Haven and for Columbus are 46% and 52% respectively. See supra text accompanying notes 17 & 20. The Philadelphia statistics here, and throughout the article are, except as otherwise noted, calculated from Statistical Report of the Common Pleas and Municipal Courts of Philadelphia–-July Term 1984, Bull. No. 84–170; Statistical Report of the Common Pleas and Municipal Courts of Philadelphia–-December Term 1983, Bull. No. 84–28; and from the appropriate annual editions of the Annual Report of the Philadelphia Municipal Court, published by the Office of the Court Administrator.Google Scholar

31 The trial rate has been rising steadily in recent years. For the years 1980 through 1983 the trial rate (as a percentage of all cases resolved by trial or by plea) was 41%, 45%, 58%, and 69% respectively. In contrast, there were no trial dispositions in the New Haven study, and trials represented only 3% of total dispositions in Columbus.Google Scholar

32 The New Haven study found no trials at all in that city, and in Columbus, trials accounted for only 3% of all dispositions. See supra text accompanying notes 17 & 20. In a study of Austin, Texas, the trial rate was found to be 3.5% of all dispositions. Anthony J. Ragona & John Paul Ryan, Beyond the Courtroom: A Comparative Analysis of Misdemeanor Sentencing 26 (Washington, D.C.: National Institute of Justice, 1985).Google Scholar

33 See infra text accompanying notes 137–49.Google Scholar

34 See infra text accompanying note 177.Google Scholar

35 The civil cases consist of claims under $1,000, code-enforcement matters, and landlord-tenant complaints. Summary offenses include curfew violations, teenage drinking, loitering, littering, and similar matters; in theory such offenses are punishable by up to 90 days' imprisonment, but in practice the judges almost invariably dispose of such cases in very informal proceedings held at the police district, and they virtually never impose imprisonment. In 1983, the municipal court disposed of 80,501 civil cases and 24,207 summary offense cases.Google Scholar

36 Almost half the states have cities or counties using similar two-tiered systems for misdemeanor cases. See Justices of Boston Municipal Court v. Lydon, 104 S. Ct. 1805, 1815 n.8 (1984).Google Scholar

37 In 1983, appeals were filed in 689 criminal cases, or about 15% of the total municipal court criminal trials ending in conviction. In practice, appeals are discouraged by the fact that the defendant may, if convicted again, receive a higher sentence after the trial denovo. See Kentucky, Colton V., 407 U.S. 104 (1972). On the other hand, if the municipal court sentence is unexpectedly severe, defendants can obtain relief by the trial de novo route even if they have no hope of winning acquittal; in at least one case in our sample (No. 0590), a defendant who received a jail sentence in municipal court appealed and after conviction received only a probationary sentence. For a detailed description of trial procedure and practice in the court of common pleas, see Schulhofer, supra note 21.Google Scholar

In Pennsylvania a defendant can challenge a conviction without risking a higher sentence, by filing a “writ of certiorari,” but this mode of review is essentially equivalent to a traditional appeal–-it is available only for errors of law and does not afford a trial de novo. The Defender Association (which represents about 70% of the municipal court defendants) estimates that it files only about 60–120 writs of certiorari each year from municipal court convictions.Google Scholar

38 Of these, 10,924 cases were held for trial in the court of common pleas and 3,813 cases were discharged; the discharge rate was thus 26%.Google Scholar

39 See Alfini & Doan, supra note 15, at 427–28. In New Jersey, misdemeanors are punishable by up to seven years in prison. Id. at 428 n.15.Google Scholar

40 Offense definitions under the Pennsylvania Crimes Code are generally identical to those of the Model Penal Code (MPC), but the grading terminology and authorized punishments are different. The four lowest MPC classifications are violation (fine only), petty misdemeanor (30-day maximum), misdemeanor (1-year maximum), and third-degree felony (5-year maximum). The four lowest Pennsylvania classifications are summary offense (90-day maximum), third-degree misdemeanor (1-year maximum), second-degree misdemeanor (2-year maximum), and first-degree misdemeanor (5-year maximum). Generally, the position of each Pennsylvania offense in this four-level classification corresponds to the position of the same offense under the MPC. As a result, the “misdemeanor” range in Pennsylvania tends to include more serious offenses, and for misdemeanors under both codes, Pennsylvania's authorized punishments are higher. For example, involuntary manslaughter, a third-degree felony under the MPC, is a first-degree misdemeanor in Pennsylvania. Aggravated assault, a second- or third-degree felony under the MPC, is a second-degree felony or a first-degree misdemeanor in Pennsylvania. Simple assault, a misdemeanor or petty misdemeanor under the MPC, is a second- or third-degree misdemeanor in Pennsylvania. Similarly, reckless endangerment, a misdemeanor under the MPC, is a second-degree misdemeanor in Pennsylvania; prostitution, a petty misdemeanor under the MPC, is a third-degree misdemeanor in Pennsylvania.Google Scholar

41 For detailed discussion of the mix of misdemeanor offenses in Philadelphia and comparison with that of other cities, see infra appendix 3.Google Scholar

42 The imprisonment rate has jumped sharply in the recent past. For the years 1980–83, the percentage of convicted defendants receiving jail or prison sentences was 12%, 14%, 17%, and 23% respectively. Significantly, the sharp jump in the imprisonment rate has been paralleled by a correspondingly sharp jump in the rate of cases taken to trial. See supra note 31. Even so, the trial rate throughout this period has consistently remained much higher than the imprisonment rate–-41% versus 12% in 1980 and 74% versus 35% in the first half of 1984.Google Scholar

43 See Schulhofer, , supra note 21, at 1047–49.Google Scholar

44 The interviews covered the interviewee's background and experience; perceptions about trials, guilty pleas, and sentencing practice; attitudes toward other participants; and, for attorneys, their training, client contacts, and preparation for trial.Google Scholar

45 The sampling procedure was not strictly random and, for practical purposes, could not be. Appendix 1 infra discusses the construction of the sample and demonstrates that it is not biased with respect to the questions under investigation. The case numbers and initials of participants that will be reported here have been modified to avoid unnecessary identification of individuals. Actual docket numbers and full names corresponding to the initials are on file with the author.Google Scholar

46 In 1983, 4% of all arrestees were discharged at or before preliminary arraignment, 17% were released on recognizance, and another 47% were released on money bail. Only 32% of arrestees were detained. These figures include both felony and misdemeanor defendants.Google Scholar

47 Under Pa. R. Crim. P. 6003 (e)(2), the first listing may not occur less than 20 days after preliminary arraignment. Sometimes first listing can be as much as 2–2 1/2 months after preliminary arraignment.Google Scholar

48 After completion of our study, a revised system provided for skipping the first listing in crash court cases and scheduling them directly for a trial courtroom.Google Scholar

49 After completion of our study, the preliminary arraignment function was assumed by bail commissioners, only one of whom was a lawyer.Google Scholar

50 Roughly 10–20% of the arraignments were conducted by closed-circuit television to facilitate processing of defendants held at police districts around the city. Whether conducted in person or by TV, the arraignment time could vary in light of the personal style of the judge. Judge KJ, for example, took more time with arraignments.Google Scholar

51 Only once in our observations did a district attorney make a request for especially high bail, and the judge denied it. Public defenders made bail objections somewhat more often, but not frequently, and seldom vigorously.Google Scholar

52 Sometimes a defender will also note the defendant's need for a particular social service or drug program so that office staff can get the defendant admitted to the program. Once this is accomplished, a defender can use the program as a basis for obtaining the defendant's release on reduced bail.Google Scholar

53 Stenographers were supposed to be on call, but so far as we could tell, the judges never used them. More recently, see supra note 49, the preliminary arraignment function has been entrusted to bail commissioners, who have no authority to take a felony or misdemeanor guilty plea.Google Scholar

54 The judge can also discharge a case, but if this is done over the D. A.'s objection, the prosecutor can simply order the defendant rearrested. In practice there are relatively few discharges at this stage. See supra note 46.Google Scholar

55 It remains unclear to what extent these practices are legal under Pennsylvania law. See Pa. R. Crim. P. 6002.Google Scholar

56 A total of 40 juveniles had been arrested at a well-known Italian restaurant, but only those with out-of-town addresses had been detained downtown. The judge, ND, remarked that he thought the restaurant owners had failed to contribute to someone's election campaign.Google Scholar

57 See infra text accompanying note 67.Google Scholar

58 This impression was confirmed by the court statisticians and by a number of attorneys. In the courtroom sample only 1.9% of the cases were processed as summary offenses. See table 3 infra. But we did not have a large enough sample of preliminary arraignment cases to permit us to quantify the frequency with which misdemeanor charges might be reduced to the summary-offense level at that stage.Google Scholar

59 By this stage, defense attorneys have usually had an opportunity to meet with their clients, and the attorneys use first listing as another chance to discuss with them the evidence and possible sentences. Defense attorneys also use first listing as an opportunity to request discovery if it has not already been provided.Google Scholar

60 For discussion of D.A.-defender interactions and prevailing practices with respect to plea bargaining, see infra text accompanying notes 63–74. The low guilty plea rate reported in court statistics tends to confirm these reports that plea negotiation seldom leads to plea agreements.Google Scholar

61 The trial commissioners are relatively high-ranking court personnel, appointed by the judges. Most are lawyers.Google Scholar

62 In one case that the D.A. had approved for ARD, the defendant expressed the desire for a trial. The trial commissioner made no attempt to change the defendant's mind and replied simply, “Fine, list it for trial.”Google Scholar

63 At the time of our observations, assistant D.A.s were sent to crash court as their first assignment after a vacation, so that they would not have to do any preparation the night before. Beginning around March 1985, this system was changed; an assistant D.A. was assigned regularly to crash court, the D.A.'s office began giving higher priority to achieving reduced detention rates, and larger numbers of defendants began to be released either on negotiated pleas or on reduced bail.Google Scholar

64 Jail officials dislike having to bring defendants in from the cells, and they tend to hassle any defender who asks to see a client, at least when the D.A.'s proposal is one the defendant is clearly not likely to accept. Thus, defenders did not insist on seeing a client when they had decided to make a strong negative recommendation.Google Scholar

65 Defenders did, however, often press the prosecutor to agree to reduced bail.Google Scholar

66 Because crash court does not convene every day, the observation period produced only five full days of court sessions. Three different prosecutors, two defenders, and 62 cases were observed during this period. In 14 of the cases, no plea offer was made because of an incomplete file or some similar problem.Google Scholar

67 Supervising prosecutor BE denied that attorneys on her staff are pressured not to bargain; she said that assistant prosecutors were told that what to offer is “up to you. Can you live with it?” Assistant prosecutors consistently painted a different picture, however. They noted that office memos convey a variety of caveats against bargaining and “lenient” deals. For example, one memo instructs D.A.s to base their offers only on the gravity of the crime and never on the evidence of guilt–-the latter can be built up if necessary. Similarly, BE conceded, office policy stressed that “it is better to try a case than to be a party to a farce.” Prosecutor TU explained why prosecutors, in practice, have much less discretion than BE had indicated: at the close of each day their work is reviewed systematically and “they are well aware that while they will not be criticized for conservative offers, lenient offers don't sit well with the office.” In addition, since supervisors do not systematically observe the trial performance of assistant D.A.s, the latter run no risk in trying and losing a case, but they may be criticized severely for negotiating a plea. Finally, prosecutor RN explained that D.A.s don't offer good bargains because, simply, “they like to try cases.” As prosecutor PT expressed it, “I'm not in a hurry, and I don't care about judicial economy. I'm paid to try cases.”Google Scholar

68 In No. 2467, a drug case, defender NT refused an offer for a $200 fine, and in No. 0293, a theft case, defender NT refused an offer of nine months' reporting probation. In both cases the defender persuaded the trial commissioner to reduce bail so that the defendants won immediate release pending trial. In No. 1116, an assault case, defender WT refused an offer of a $200 fine when the trial commissioner warned that the defendant faced a detainer on other charges. And in No. 2845, a drug case, defender NT refused an offer of a $300 fine; a detainer had already been lodged, and the plea would have created two violations of probation and, thus, a substantial risk of imprisonment.Google Scholar

For defendants already on probation, defenders strongly stressed the dangers of “pleading into a V.O.P. [violation of probation].” Their recommendations depended heavily on the identity of the sentencing judge in the previous case. Judge GP usually responded to a V.O.P. by adding additional time on probation; Judge JM was reputed to do no more than give defendants a stern lecture. Judge GE, on the other hand, almost invariably revoked probation and gave a jail or prison Sentence; for defendants on probation before Judge GE, a guilty plea to the subsequent charge was almost always out of the question.Google Scholar

69 Schuman, Stuart H., Answers to Common Questions About Crash Court and How to Effectively Represent Defendants in Crash Court (Room 888) 25 (Sept. 1983) (unpublished memorandum on file with the author).Google Scholar

70 See supra note 68.Google Scholar

71 For example, in No. 2678, a prostitution case, defender NT rejected an offer of a $350 fine; in No. 0599, an assault case, defender WT rejected an offer of 18 months' probation. Note that defenders made these decisions without direct consultation with their clients. See supra note 64. In borderline cases, they would consult with the defendant and explain why the plea should be rejected. Usually their efforts to counteract the “overwhelming lure” were successful. But No. 2481 was an exception. Defender NT urged a shoplifting defendant to reject an offer of one year's reporting probation, because he felt almost certain that the charges would be dismissed at trial. Nonetheless, the defendant chose to plead guilty.Google Scholar

72 In No. 2746, a defendant facing drug charges, held in lieu of $2,500 bail, was offered a guilty plea and fine of $250. In conference with his attorney, the defendant initially resisted the plea offer and swore that he was innocent. Defender NT explained that although the arresting officer might have lied, a judge would likely believe the officer's testimony; with some difficulty, NT finally persuaded the defendant that it would be better to take the deal and to state (falsely) that he was guilty.Google Scholar

73 The trial commissioner speculated that in this case the defendant, a male, was making a good deal of money practicing his profession while in custody. Defender PP reported that male prostitution defendants frequently avoid release for the same reasons. For such defendants the process becomes a bonanza rather than a punishment.Google Scholar

74 Defender perceptions about the severity of prosecutorial offers were largely borne out by subsequent developments in the cases we observed. For example, in No. 0931, a prostitution case in which the defendant had 13 prior prostitution convictions, the defender rejected a plea offer involving six months' incarceration. After conviction at trial the judge imposed only a $300 fine and court costs. Altogether, we were able to trace outcomes in 24 of the crash court cases in which we had observed the rejection of prosecutors' offers. Of these, 9 cases were eventually dismissed; in 1 the record was expunged (presumably because the defendant had qualified for ARD); and in 7 the defendant received a sentence more lenient than the prosecutor's offer. In 2 cases the ultimate sentence was equivalent to that of the offer, and only in 5 cases (21% of the total) was the sentence after conviction at trial more severe than the prosecutor's guilty plea offer.Google Scholar

75 See infra text accompanying notes 117–21.Google Scholar

76 Beddall, Jane, University of Pennsylvania Department of Criminology, Private Criminal Complaint Processing in Philadelphia 38 (Spring 1984) (unpublished paper on file with the author). In cases not approved for filing, the complainant may appeal the district attorney's decision, but such appeals are rarely granted. Id. at 8.Google Scholar

77 See id. at 11, for discussion of the content of consent agreements.Google Scholar

78 Annual Report of the Philadelphia Municipal Court, 1983, at 52. The agreement to seek arbitration is not absolutely binding. If the arbitrator does not expect the parties to honor any proposed solution, the case is referred back to the private complaint courtroom and eventually to a trial courtroom. Beddall, supra note 76, at 11–12.Google Scholar

79 Annual Report, supra note 78, at 28, 43–45; Beddall, supra note 76, at 38. Note that only the 1,024 cases referred for trial are given a criminal docket number and counted in the municipal court statistics on criminal dispositions.Google Scholar

80 In such cases waiver of the disqualification is noted in the record. Although the Department of Public Welfare could, in theory, ignore the waiver, it does not do so, apparently because of its interest in cooperating with the D.A.'s office. In addition, a failure to honor the waiver would presumably entitle the defendant to withdraw the plea.Google Scholar

81 See Pa. R. Crim. P. 175–85. In 1983, pretrial diversion accounted for 8,114 cases, or 33% of total misdemeanor dispositions in the municipal court. For the first half of 1984 the proportion was 36%.Google Scholar

82 The program is intended primarily for first offenders charged with nonviolent misdemeanors–-e.g., prostitution, gambling, theft, and welfare fraud. Also eligible are defendants charged with weapons possession, drug possession (small amounts), drunk driving, minor assaults, and even such felonies as auto theft and burglary (if the building was uninhabited).Google Scholar

83 Defender BI and Judge LH agreed that the D.A.'s office does not make ARD decisions this way. And BI added that the ARD judge will refuse diversion and simply discharge a case when the hearing reveals a problem of insufficient evidence.Google Scholar

84 See Pa. R. Crim. P. 178–79.Google Scholar

85 At times the total runs as high as 100 cases per day. The court “day,” moreover, runs from 10:00 A.M. until roughly 12:30 or 1:00 P.M.Google Scholar

86 Similarly, when private attorneys were present, they typically offered comments on the defendant's strong community ties and good character.Google Scholar

87 Of a total of 69 cases not continued, 32 (46%) were resolved in about one minute, 17 (25%) in two minutes, 15 (21%) in three or four minutes, 4 (6%) in five to eight minutes, and 1 (1%) in ten minutes.Google Scholar

88 In Nos. 2405 and 2429 two defendants were charged with siphoning $5 worth of gasoline from a parked car. The first defendant commented that the usual disposition, three months' nonreporting probation, would prevent him from enlisting in the Navy the following week, as he had planned. The judge ordered one week's nonreporting probation for the first defendant. After initial hesitation, he decided it was “only fair” to give the co-defendant the same disposition. In No. 2348, the defendant was charged with stealing $41 from a cash register. The judge originally ordered six months' reporting probation, but the defendant complained that he lived in New Mexico and had just been passing through Philadelphia. He also said, “It wasn't my fault anyway. The box was open, and there was no one in the store.” Judge LH replied, “If you were in New Mexico, they'd put you against the wall and shoot you!” Then, he modified the disposition to nonreporting probation.Google Scholar

89 No. 2433, a drug case, had been diverted into ARD after a previous prosecutor had discovered a clear problem of mistaken identity at a prior hearing. Judge LH, visibly irritated that this “bad news” prosecutor had attempted to subvert ARD in that way, ordered a dismissal. A contrasting approach was taken in No. 2454, a theft case in which the complainant charged that someone had taken his wallet at a gas station where the defendant worked. The defendant claimed innocence, and the police report itself alleged that the wallet was taken by the defendant's friend. Judge LH was incredulous: “How can you hold him on this?… There is no case.” The defense attorney insisted on a complete discharge, but the prosecutor responded, “I don't know why he was arrested… I want it to go to trial.” As a result, the file was marked “ARD rejected by defendant,” and the case was listed for trial.Google Scholar

90 No. 2421 involved a dispute triggered by a dogcatcher's attempt to impound the defendant's dog as it slept on his front steps. The defendant and a friend called the police; the officer then arrested them, and the defendant was charged with resisting arrest. A private defense attorney vigorously protested the episode, and the defendant commented, “The dog got out of the slammer sooner than we did.” Judge LH indignantly observed that this was a “shameful incident… [it] should never have happened.” Nonetheless, he was unable to discharge the case. Rather than list it for trial, he ordered one months' nonreporting probation.Google Scholar

No. 2432 produced a similar outcome. In this case, the defendant called the police to break up a fight involving his niece and another woman. After the fight was over, the defendant, an older man, told one officer that his niece should be taken to the hospital. At that point, he said, another officer grabbed him from behind and threw him to the ground. (The police report, alleging that the defendant struck one of the officers, charged him with aggravated assault.) The private defense attorney said, “My client is not interested in ARD. He wants to prosecute the officer.” Again, Judge LH was indignant: “This happens again and again. You've done nothing wrong. But the officer will lie in court. He must justify his arrest…. If the judge believes the officer, he'll convict you of aggravated assault… I think you should sue the officer… Being in ARD won't preclude you.” The defendant then accepted a disposition of one months' nonreporting probation. (Note that in plea negotiation in other cities, a common condition in such a case is that the defendant must waive any claim against the police.).Google Scholar

91 The treatment program typically involves a program of 15 weekly counseling sessions of several hours each and regular attendance at meetings of Alcoholics Anonymous. If the defendant has a particularly acute problem, a more extensive treatment program may be ordered. In theory, inpatient care may be required, but this is said to be extremely rare. (At the Defender Association, supervisor BI reported that in a year's work he had never heard of such a requirement's being imposed.).Google Scholar

92 Pa. Vehicles Code, 75 Pa. Cons. Stat. Ann. § 3731 (e) (6) (ii) (Purdon 1985).Google Scholar

93 Retail theft school consists of a single five- to six-hour session. In the first half, an attorney explains the retail theft statute, emphasizing the toughness of the law and discussing the impact of retail theft on the merchant. Then, there are small group discussions led by clinical psychologists. The group leaders also look for drug and alcohol problems and urge those with such problems to seek counseling.Google Scholar

In the 32 drunk driving cases we observed, 12 defendants were considered nonproblem drinkers, and all received six months' nonreporting probation, safe driving school, and a one month's license suspension. The 20 problem drinkers received a year's reporting probation with a required treatment and rehabilitation program. We also observed 69 dispositions not involving drunk driving; 28 of the defendants received nonreporting probation, nearly all for a three months' period; 9 welfare fraud defendants received probation for purposes of restitution; and 32 defendants received probation with some reporting requirement, evaluation, or other program.Google Scholar

Some defendants referred for evaluation are found not to need counseling or other programs, and in the end, they too receive only nonreporting probation. In June 1984 a total of 6,757 active cases were in Philadelphia ARD programs. Of these, 21% involved only nonreporting probation; 25% were in the various drunk driving programs (evaluations, safe driving school, and rehabilitation); 10% were in other drug and alcohol treatment programs: and 43% were in some other kind of reporting probation, including restitution programs, counseling programs, and retail theft school. Interview with Frank Eaverly, chief of Diversion Services, Philadelphia Probation Department, Philadelphia (Aug. 8, 1984).Google Scholar

94 The D.A.'s ARD unit maintains a record of ARD dispositions in order to determine whether a defendant should be refused ARD eligibility in the event that a subsequent offense is charged. Similarly, the state police and state Department of Transportation maintain records of ARD dispositions in drunk driving cases in order to assure that repeat offenders in such cases are granted ARD no more than once every seven years. In each case these internal records are not available for any other purpose. The criminal records kept by the FBI and by the Philadelphia and state police are expunged. AU participants agreed that the expungement procedure is scrupulously respected. ARD dispositions do not, for example, play any part in sentencing in the event of a subsequent offense, and insurance companies have been refused access to the ARD drunk driving records maintained by the state Department of Transportation.Google Scholar

95 These include not only the impact on job prospects but, of course, the prospect of more severe sentencing in the event of a subsequent offense. Under Pennsylvania's sentencing guidelines system, prior convictions lead to direct enhancement of the presumptively applicable sentence.Google Scholar

96 See Franklin Zimring & Richard Frase, The Criminal Justice System 386–87 (Boston: Little, Brown & co., 1980).Google Scholar

97 For the first half of 1984, there were 4,368 adjudications by plea or trial; of these, 1,149 (26%) were guilty pleas. If the 3,972 ARD cases are treated as pleas of guilty, then there would be a total of 8,340 adjudications by plea or trial; of these, 5,121 (61%) would be guilty pleas. For the full year 1983, guilty pleas accounted for 31% of the adjudications by plea or trial, and if ARD cases are treated as pleas of guilty, then guilty pleas would represent 65% of the adjudications by plea or trial.Google Scholar

98 For the first half of 1984, there were a total of 10,951 misdemeanor dispositions. If the ARD cases are treated as pleas of guilty, there would be 8,340 adjudications by trial or plea (see supra note 97), and nonadjudications would represent only 24% of the total caseload. On similar assumptions, nonadjudications would represent only 33% of the total caseload for the year 1983.Google Scholar

99 For New Haven and Columbus, see supra text accompanying notes 17–20. In the Boston municipal court, 69% of the misdemeanor cases (including motor vehicle violations carrying a penalty of imprisonment and/or a fine of more than $100) were dismissed, withdrawn, or diverted in the 1983 fiscal year. Annual Report of the Massachusetts Trial Court–-1983, at 23, 26–27. In Chicago the reported statistics group together misdemeanor, ordinance, and conservation violations, and report the disposition of charges rather than cases; in 1983, 81% of such charges were dismissed or withdrawn. Statistical Report on the Circuit Court of Cook County, Illinois, for 1983, at 51 (on file with the author). In the District of Columbia, the dismissal/diversion rate was 56% of all dispositions of U.S. misdemeanors and 92% of all dispositions of D.C. misdemeanors. Annual Report of the District of Columbia Courts–-1983, at 53. Such comparisons can only be suggestive, however, as they do not (and probably cannot) control for the possibility of differential screening at the point of arrest or at other early stages before a charge is filed.Google Scholar

100 See infra appendix 2.Google Scholar

101 A trial commissioner normally makes these room assignments at first listing, on the basis of the complexity of the charges and the expected number of witnesses and defendants. The method for making specific room assignments is in constant flux and is no longer identical to the one in effect at the time of our study.Google Scholar

102 One judge, BL, even made a practice of talking on the telephone himself during court sessions.Google Scholar

103 For example, in No. 2663, a drunk driving case, and in No. 0467, an assault case heard immediately afterward, Judge CR apologized to witnesses for the din of jazz music, construction, and traffic noises coming in through the open window.Google Scholar

104 See Feeley, , supra note 16, at 8–9, quoting letter from John Hersey to the Forum (Aug. 11, 1972).Google Scholar

105 The court officer in room 295 (a “mass production” courtroom) was particularly insistent that members of the audience sit up straight. In other courtrooms, a more casual atmosphere frequently prevailed. But court officers uniformly objected, for example, to any attempts to read a newspaper while court was in session.Google Scholar

106 Similarly, in No. 0148, a disorderly conduct case, Judge CR interrupted the guilty plea colloquy to tell two police officers to stop talking.Google Scholar

107 There are 22 active municipal court judges. In addition, cases involving juvenile witnesses (usually indecent assault, corrupting a minor, and the like) are heard in juvenile court by a common pleas judge who, for this purpose, sits as a municipal court judge and handles cases on the misdemeanor docket. For discussion of the sampling procedure, see infra appendix 1. In order to be able to follow the proceedings, the courtroom observers positioned themselves in the jury box, which is close to the bench and the witness stand, and they arranged to borrow a copy of the printed daily docket sheets, which include information on the docket number, charges, attorneys, and bail status for each case. Observers were not present at confidential conferences in chambers, but they could tell when such conferences occurred because any absence of the prosecutor and public defender from the courtroom was immediately apparent. Often we obtained an account of such a conference from one or more of the participants, but in any event, such conferences were quite rare. We are also reasonably confident that prosecutors and defenders did not discuss cases with judges after courtroom hours. We did not know, however, whether private attoryneys did so.Google Scholar

108 In the courtroom sample 20% of defendants were in custody, 21% had been released on recognizance, 40% had been released on bail, and 9% were released on undetermined status; 10% of defendants appeared on private complaints and, therefore, had not been subject to arrest at all.Google Scholar

109 Sixty percent of the defendants were represented by the public defender, and 39% were represented by private attorneys, either retained or court appointed. It was not possible to determine from observations or from the court records available to us whether a private attorney had been appointed or retained.Google Scholar

110 Prosecutor PT said that she never left before 8 P.M. and sometimes had to stay until 11 P.M. in order to complete her preparation for the next day.Google Scholar

111 Conclusions about the adequacy of preparation can be drawn from what the attorneys themselves told us, supplemented by observations in court. Inadequate preparation becomes obvious when an attorney frequently stumbles or hesitates in posing questions, or when he or she has to refer frequently to the the in order to recall what the case is about. It is more difficult to conclude with complete confidence that preparation was adequate, but when an attorney asks questions smoothly, in logical sequence, and develops an argument with a clear sense of direction (as was usually true, for example, of public defenders), the observer has a strong impression of adequate preparation. More rigorous conclusions about the adequacy of preparation (and about the related matter of investigation) would require a methodology different from the one employed in the present study.Google Scholar

112 Defender GL estimated that 99% of his clients do report for their client interview. We did, however, occasionally observe cases in which the defender's first conference with the client occurred in court—e.g., No. 2322, a prostitution case handled by defender LZ.Google Scholar

113 At the Defender Association, the supervisor of the municipal court unit reviews each file after preliminary arraignment and after each subsequent listing to determine whether investigation or other preparatory steps are required. In addition, the supervisor stresses the need for the attorney handling each listing to develop trial memoranda and to note impressions of witnesses and the results of any witness interviews conducted. The adequacy of this work product is also examined when the supervisor reviews the files. His slogan, which becomes a watchword for the attorneys in the unit, is, “Advance the file! Advance the file!”.Google Scholar

114 For the first six months of 1984, there were 4,368 adjudications by trial and by plea; 73.7% of these were by trial and 26.3% by guilty plea. During the same period, 60% of the overall caseload was disposed without an adjudication by trial or plea. If one excludes ARD cases, 37% of the remaining cases were disposed without adjudication. The much lower figure in our own sample (12%) reflects the fact that most dismissals occur not in the trial courtrooms (from which our sample was drawn), but rather in the preliminary arraignment or listing rooms (which we observed without attempting to draw a sufficiently large sample).Google Scholar

115 See supra text accompanying note 55.Google Scholar

116 Several of the 107 cases involved more than one type of motion. Six cases (6%) involved speedy trial motions, 92 (86%) involved challenges to search or arrest, 17 (16%) sought suppression of statements, 4 (4%) sought suppression of identifications, and 10 (9%) involved other issues.Google Scholar

117 The mean motions time was 17.7 minutes, and the median was 13 minutes, excluding recesses. In No. 1799, a weapons possession case in which the defendant had been identified by the D.A.'s office as a “career criminal” (one against whom a severe sentence will be sought), private defense attorney RF filed motions seeking suppression of physical evidence, identifications, and a statement by the defendant. The motions hearing before Judge ND consumed 2 hours and 21 minutes. Not all the long motions occurred in unusually serious cases. In No. 0914, a drunk driving case before Judge VA, a motion to suppress a breathalizer test consumed 58 minutes. In Nos. 0175 and 0229, defender BN and private attorney AS, representing two co-defendants charged with theft of a TV set, filed motions challenging physical evidence, identifications, and statements; after a 92-minute hearing, Judge JM granted the motions in full.Google Scholar

118 In No. 1217, a drug possession case before Judge JM, an inexperienced intern from the D.A.'s office said diffidently, “I have some case law if it would help.” JM replied, “It always does.” Close attention to legal standards was common among a variety of judges, attorneys, and cases. E.g., No. 0175, a theft case before Judge JM (prosecutor AB; defender BN); No. 0177, a drug possession case before Judge JM (prosecutor MP; defender CA); No. 1343, a weapons possession case before Judge KA (prosecutor HK; defender BN); and No. 1981, a gambling case before Judge HL (prosecutor EE; defender PE). On the other hand, a few judges, notably GE, GP, and CR, seemed relatively uninterested in technical legal issues, and VA seemed to have a poor command of legal subtleties.Google Scholar

119 Examples include No. 1343, supra note 118, and No. 1466, a theft case before Judge JM. In No. 0089, a weapons possession case, Judge HL called the attorneys' attention to a decisive case on point and then, after giving them time to argue it, granted the motion to suppress. In No. 1343, supra, after testimony and cross-examination, Judge KA raised a series of questions concerning the proper scope of a frisk, the actual grabbing distance, probable cause, and the possible applicability of special rule, governing automobiles. After initially denying the motion, the judge decided to take it under advisement, explaining that “this is a very sticky area.”Google Scholar

120 E.g., No. 1026, a shoplifting case (Judge MN); No. 0032, an assault case (Judge GO); No. 1343, supra note 118. In No. 1466, supra note 119, the prosecutor lost a speedy trial motion because his office had failed to file a timely brief. Afterward, defender LZ remarked that “in all fairness [to the D.A.'s office], many judges don't read briefs.” Judge JM retorted, “1 do.”Google Scholar

121 In No. 2147, a drug possession case before Judge MN, prosecutor BE and defender GL relied on a variety of legal theories and case law in a thorough argument that lasted a total of only 7 minutes; the suppression motion was granted. In No. 2441, a theft case before Judge JM, defender TS did relatively little questioning and pursued a strategy of “the less said the better”; the judge found prosecutor EE's argument unconvincing and granted the motion to suppress after a total of 12 minutes. Private attorneys, in contrast, sometimes pursued cross-examination and argument well beyond the point of diminishing returns; in No. 0914, supra note 117, defense attorney RN dragged out a losing suppression motion for 58 minutes.Google Scholar

122 In No. 0847, a private complaint charging harassment and assault, when prosecutor SW tried to explain the background of an unusual situation, Judge CR cut her off, saying that she was a D.A. and therefore couldn't testify. In No. 2226, a private complaint charging harassment and criminal mischief, prosecutor TU, unfamiliar with CR's style in such cases, kept trying to ask questions. A contrasting approach was taken by Judge GA. In No. 1542, a private complaint growing out of a custody dispute in which the defendant was charged with harassment, assault, and terroristic threats, GA played a very passive role and allowed the attorneys to talk to one another and to the parties until they all agreed on dismissal with a protective order—the standard outcome for such cases.Google Scholar

123 Of the 26 defendants in cases handled by mediation, 5 appeared without counsel. In contrast, among the 554 cases resolved by plea or trial and the 51 cases involving only pretrial motions, only 1 defendant appeared without counsel.Google Scholar

124 E.g., No. 0373, an assault case before Judge CR.Google Scholar

125 In 2 of the 26 cases, the judge achieved the same result by holding the cases indefinitely under advisement rather than formally dismissing them.Google Scholar

126 The protective order reads: “The Court hereby directs that pursuant to P.C.C. § 4954 a protective order be entered against defendant ordering that person not to retaliate against or intimidate any witness or victim in this matter; ordering that person to maintain [his] distance from [name of complainant]; ordering that person not to communicate with [name of complainant]. Violation will result in revocation of bail or other penalty as prescribed in P.C.C. § 4955.”Google Scholar

127 Mediation also required less court time than a full trial. Mediation cases lasted anywhere from 2 to 15 minutes, with an average of about 7 minutes each. Dispositions of this kind represented only about 4% of the cases in the courtroom sample, but 50% of the private complaints (excluding welfare fraud complaints) were processed by the mediation approach.Google Scholar

128 In Nos. 1698 and 1699, in which a mother and daughter both complained of harassment and assault, the defendant (representing himself) laughed throughout the proceedings and bragged to the judge about having made love with the daughter. Although CR told the defendant, “You're a wise guy,” and repeatedly warned him to stop laughing, the defendant seemed to regard the whole process as a joke.Google Scholar

129 Of 143 guilty or nolo pleas in the courtroom sample, 70(49%) were taken individually and 4 (3%) were taken in pairs of related defendants. Fifteen (10%) were taken in groups of 2, 3, or 4 unrelated defendants, and 54 (38%) were taken in groups that consisted of 7, 10, 11, or 12 unrelated defendants. Cases were more likely to be heard individually when defendants were in custody or when, for example, an interpreter was required. Surprisingly, the 8 summary offense guilty pleas were all taken individually.Google Scholar

130 In No. 1426, a theft case before Judge LC, the defendant disagreed with the facts summarized by the D.A. and said, “I can't go on with this.” The case was then listed for trial.Google Scholar

131 There were only two exceptions. In No. 3378, an attempted theft case, Judge GP reduced the agreed disposition to a summary offense and imposed only a $50 fine. And in No. 0388, a weapons possession case, Judge GP, after hearing the statement of facts, found the defendant not guilty.Google Scholar

132 The mean was 9.9 minutes and the median was 8 minutes, excluding recesses. In No. 1812, Judge HL allowed a prostitution defendant to waive the colloquy, and the entire proceeding was completed in 1 minute.Google Scholar

133 See supra text accompanying notes 69–74, and infra note 139. In No. 2426, a drug defendant held in lieu of $2,500 bail was offered a guilty plea with two years' reporting probation. Defender NT recommended that he accept, but the defendant refused. NT then urged him to talk it over with the other defendants in the lockup. When the defendant was brought down to the courtroom for trial, he said he would take the offer but urged the defender to try for 18 months' probation; prosecutor EW then agreed to this proposal. In No. 0305, private attorney PR persuaded his client to accept prosecutor AB's offer of two years' probation for sale of marijuana in a case before Judge GP. PR was evidently unfamiliar with the judge's sentencing practices because GP never gave such a long probation in that kind of case, and a public defender would have negotiated a better deal; courtroom observers felt that PR “got taken” by a more experienced prosecutor.Google Scholar

134 Summary-offense guilty pleas typically involved such charges as disorderly conduct, possession of a mall quantity of marijuana, or retail theft under $150. In some of these cases the decision not to charge at the misdemeanor level conceivably could reflect a tacit bargain. And in No. 2020 the prosecutor reduced a misdemeanor charge (obstructing the highway) to a summary offense. But as table 3, supra, shows, there were altogether very few summary-offense pleas in the sample.Google Scholar

135 Roughly 69% of the guilty pleas in that court were either open pleas or pleas tendered in exchange for dropping inconsequential lesser counts, and in neither case did judicial sentencing practice provide tacit sentencing concessions for the plea. Only 15% of the pleas were given in exchange for a specific sentence recommendation. Schulhofer, supra note 21, at 1059.Google Scholar

136 Id. at 1061.Google Scholar

137 E.g., No. 1908, a charge of obstructing the highway (Judge JM); No. 0754, a case of assault, drug possession, and resisting arrest (Judge JM).Google Scholar

138 In No. 1324, a defendant caught in the electric company's yard was charged with theft and criminal trespass and was detained pending trial. On the day of trial his court-appointed private attorney failed to appear, so Judge CR asked what the case was about and then asked defender LZ to talk to the defendant: “See if we can take care of this. I can give him time in, or 30 days, if it won't mess him up.” When the defendant was brought in, LZ met with him, and the defendant agreed to a stipulated trial on the basis proposed. After prosecutor EW read the police report, CR cut off defender LZ's argument and said, “I see trespass in there. I acquit on everything else.” A time-served sentence was then imposed. Similarily, in No. 0645, a case involving sale of cocaine to an undercover officer, prosecutor SW was not ready when the case was called, and CR then asked if the defendant wanted a fine. Defender LZ accepted, and the case was disposed of as a stipulated trial.Google Scholar

139 No. 0514. Defender LZ considered the offer a great deal because the defendant had a record “a mile long.” Nonetheless, the defendant refused to accept the $500 fine, and the case was listed for trial.Google Scholar

140 E.g., No. 2715, a prostitution case (Judge JM). Similarily, in No. 1789, a drug case before Judge GA, private attorney BT announced that the case would be a negotiated plea so as “not to waste court time.” The court reporter then interjected, “A stip. trial would be faster,” and BT replied, “We'll do a stip, then.”Google Scholar

141 Similarly, in No. 2594, an indecent exposure case, a guilty plea colloquy was begun, but the defendant, who had obvious psychiatric problems, had difficulty with the colloquy, so Judge JM persuaded the parties to proceed by a stipulated trial instead. Judge HL, in contrast, approached stipulated trials very cautiously. He conducted very complete colloquies (e.g., No. 2315, a drug case), and he told us that he does not like the stipulated trial format because he fears that defendants do not realize what they are doing and what rights they are forfeiting. Ironically, however, Judge HL once permitted a prostitution defendant to waive her guilty plea colloquy. See supra note 132.Google Scholar

142 No. 2061, a lotteries case. In No. 1040, a shoplifting case, Judge GP made very clear his preference for the stipulated trial format. Prosecutor PT insisted on proceeding by guilty plea because she did not want to leave open the possibility of an appeal. Although GP said he did not want to hear the colloquy and made plain his irritation, she stuck to her position. (She told us afterward that she had a job to do and was “not in court at [GP's] discretion.”) When PT promised that the colloquy would be short, GP replied, “No guilty plea colloquy can be done correctly in a short time.” To prove his point, GP then made PT conduct her colloquy with extreme care and often interrupted her to call for more detail. (“Explain how the jury is selected.”) After what proved to be a 28-minute colloquy, GP said, “You see, if we did that in every case we'd be here until 10:00 at night.”Google Scholar

143 The mean was 7.9 minutes (versus 9.9 minutes for a guilty plea). The median was 6 minutes (versus 8 minutes for a guilty plea). One explanation for the relatively small difference in average times is the more frequent resort to mass processing in guilty plea cases. Although some stipulated trials were also conducted en masse (once with a group of ten unrelated defendants), 79% of the stipulated trials were conducted individually, compared with only 46% for the misdemeanor guilty pleas. The room assignment procedure, which channels large numbers of guilty pleas to a single room, makes mass processing more likely for the guilty plea cases.Google Scholar

144 For example, in No. 2569, a drunk driving case, the police report stated that the officer observed the defendant standing outside a crashed vehicle and smelled alcohol on his breath. Private attorney SG stipulated to the facts, and Judge GP granted a complete acquittal because there was no evidence that the defendant had been driving the car. Similarly, in No. 3077, a drug possession case, the police report stated that an officer had entered a house pursuant to a warrant and seized three marijuana cigarettes on a living-room table. Defense attorney BT stipulated to the facts, and Judge GE acquitted because, with several people in the room, there was insufficient proof of possession by the defendant.Google Scholar

145 For example, in No. 2849, a weapons possession case, the police report listed the weapon as a shotgun with a 30-inch barrel. Because the statute defines “firearms” as including shotguns only when the barrel is less than 24 inches, see 18 Pa. Cons. Stat. Ann. § 6102 (Purdon 1983), defender FT argued that the particular weapon was not covered by the statute. Judge KA adopted a broad “plain meaning” interpretation and rejected this argument, but FT immediately gave notice of intent to appeal.Google Scholar

146 E.g., No. 1890, a theft case (Judge KJ). The case rested heavily on statements made by the defendant to a private investigator. A motion to suppress the statements, directed primarily to the question whether the investigator was required to give Miranda warnings, consumed 29 minutes. Judge KJ finally ruled the statements admissible on the ground that regardless of the state action problem, the interrogations had not occurred in custody. Defender CA then stipulated to the facts but argued intensely over the grading of the offense. Altogether this stipulated trial consumed one hour and 19 minutes.Google Scholar

147 In Los Angeles, where stipulated trials were a frequently used alternative to the guilty plea, Professor Lynn Mather found that those trials similarly had significant adversarial elements on occasion. Lynn M. Mather, Plea Bargaining or Trial? 55–56 (Lexington, Mass: D. C. Heath & Co., 1979).Google Scholar

148 Conceivably, a defense victory can occur when the D.A. realizes that there is no case, the parties stipulate, and the case is submitted to the judge for acquittal. Such essentially uncontested cases are, in effect, “slow dismissals.” See Schulhofer, supra note 21, at 1076 n.142; Pamela J. Utz, Settling the Facts 59–60 (Lexington, Mass: Lexington Books, 1978). But we did not see this kind of behavior in the stipulated trials we observed.Google Scholar

149 E.g., No. 1890, supra note 146; No. 2849, supra note 145. Altogether there were four instances in which unsuccessful suppression motions were made in cases recorded as stipulated trials.Google Scholar

150 If a pretrial motion has been filed, this will be heard and decided first, with legal argument and a full-dress evidentiary hearing when necessary. See supra text accompanying notes 116–21. But relatively few dispositions involve a motions hearing on the same day as the trial. In our sample, there were 107 cases involving motions, and 51 of these were continued or dismissed without trial. The remaining 56 motions cases represent 14% of all the cases disposed of by trial.Google Scholar

151 The attorneys sometimes expressed a view about sentencing (often after the judges indicated their initial position), but presentence investigations were infrequent except in drunk driving cases (for which the statute mandates explicit consideration of prior record). Altogether, presentence investigations were ordered in 17% of the cases ending in conviction.Google Scholar

152 The mean time per trial was 26 minutes, and the mean time per defendant (allowing for joint trials) was 25 minutes, excluding recesses. The median time was 21 minutes for both trials and defendants.Google Scholar

153 One day GP opened court by telling the D.A. to “crank it up to 78 RPM.” In No. 1857, a drug case, NZ—a vigorous private attorney—announced at the end of the Commonwealth's case that he had several witnesses to call, and Judge GP replied, “Your $50 fine is going up rapidly!” NZ immediately rested, GP convicted, and the promised sentence was imposed. More simply, GP sometimes just interrupted a defender in midargument and ruled, “Guilty.” In No. 1376, GP was perplexed by a police witness's testimony, so he called a recess and had a private discussion with the officer in chambers.Google Scholar

154 See also No. 1040, discussed supra note 142.Google Scholar

155 Such judges included KA, GE, HL, GO, CR, and OT. In contrast, Judges GA, KJ, and JM often showed signs of impatience, and Judge MN, though usually careful in his rulings, occasionally cut off relevant questioning or chided attorneys unfairly for wasting time.Google Scholar

156 Similarly, in No. 1669, a theft case, an apparently inexperienced private attorney frequently failed to initiate proper objections to evidence. During the prosecutor's case, Judge MN often interrupted to say, “sustained.” Defense counsel then said (belatedly), “objection!”. Other formal matters, such as the competency of witnesses or the qualifications of experts, were also litigated, though less frequently. In No. 1633, a case involving charges of indecent assault on a 3 1/2-year-old child, Judge GO heard extended voir dire and intense, thoroughly prepared arguments before ruling that the child was not competent to testify. In No. 2364 the charge was tampering with a motorcycle identification number; defender CA challenged the qualifications of the police witness to testify as an expert in the matter, and when prosecutor BS was unable to qualify the officer, Judge JM dismissed the case.Google Scholar

157 GE and CR were exceptions; both were very patient, but they tended to sit passively through the trials and rarely asked questions.Google Scholar

158 Similarly, in No. 1565, where the charge was illegal inhaling of solvents (18 Pa. Cons. Stat. Ann. § 7303 (Purdon 1983)), defender LZ showed Judge JM a case suggesting that the Commonwealth must prove specific intent to produce intoxicating effects. JM studied the case and argued with LZ about whether it was factually distinguishable before he finally rejected the defender's position.Google Scholar

159 E.g., No. 1128, a case before Judge KA in which the charge was obstructing a governmental function (the defendant had delayed a police officer who was pursuing another man). In No. 0262, a firearms case, Judge KJ offered prosecutor OT a chance to submit a brief on the question whether a possession charge was proper when the weapon was found in a car the defendant had driven; OT declined, and KJ acquitted.Google Scholar

160 Similarly, VA seemed very difficult to persuade on reasonable doubt questions; and CR, though quite patient and normally a rather fair judge, seldom rejected dubious police testimony.Google Scholar

161 The extent of any penalty on appeals remains unclear, however, and defender GL maintained that there is, in general, no penalty for taking an appeal. The problem appears to be one of comparing the sentence actually imposed with the sentence that might be expected from any one of a number of different common pleas judges. Sometimes a common pleas appeal can result in a sharply reduced sentence. See supra note 37. And some of the municipal court judges, notably GE, impose very stiff sentences even in cases that defenders view as involving considerable doubt. Such cases are frequently appealed.Google Scholar

162 No. 0031, a case before Judge GO, involving a charge of endangering the welfare of a child. For the judge's efforts to protect the defendant in this case, see supra text accompanying note 156.Google Scholar

163 See supra text following note 62.Google Scholar

164 Judge MN's standard approach was to tell an unrepresented defendant to step back so that he could see the defendant's feet. Then MN would say, “Good, you have your sneaks on. If you're not back here in one hour with a lawyer, bring a toothbrush, because that's all you're allowed in jail.” Similarly, Judge JM told one defendant that if he didn't have a lawyer at next listing, his bail would be increased to $5,000. When the defendant said he didn't want a lawyer, JM replied “You are trying my patience,” and warned him about the increase in bail. Judge GE was even more direct; he told one defendant that he had half an hour to get an attorney or his bail would be revoked.Google Scholar

165 Examples include No. 0914, a drunk driving case (Judge VA); Nos. 2242 and 2245, related shoplifting cases (Judge CR); No. 0820, a criminal mischief case (Judge CR).Google Scholar

166 In No. 2091, an assault case growing out of a neighborhood scuffle, defender BN argued that the mentally retarded defendant was incapable of forming the intent required to support an assault charge; Judge KA rejected this argument, but placed the defendant on psychiatric reporting probation. See also No. 1565, discussed supra note 158; No. 1128, discussed supra note 159.Google Scholar

167 342 U.S. 246 (1952).Google Scholar

168 Nos. 2416 and 2417 (Judge MN). Officers on routine patrol had noticed the defendants taking compressed steel bricks from a vacant lot. After the officers determined that the bricks had not been abandoned, the defendants were charged with theft; the property was valued at $70. At an early point in the trial the judge said that he thought the defendants knew the steel bricks had not been abandoned, and he also cut short a cross-examination designed to show (for grading purposes) that the value of the metal was below $50. But there was careful probing of the circumstances indicating guilty knowledge, and in the end, on grounds of reasonable doubt, the judge granted an acquittal on all charges.Google Scholar

169 A typical example is No. 1991, a case before Judge JM. Defendant was charged with reckless endangerment and resisting arrest; allegedly, he had ridden his motorcycle out of control through a crowd of people and then slugged the officer who tried to arrest him. There was close cross-examination of the arresting officer, and the defendant himself testified in detail about the incident: He claimed in particular that the officer had approached with his gun drawn; that he—the defendant—had never thrown a punch; and that several officers had beaten him badly. The defense also presented the testimony of two eyewitnesses who provided detailed corroboration of the defendant's version of the episode. The trial lasted 44 minutes. Similarly, in No. 1049, a case before Judge GE, a woman who allegedly pulled a department store guard away from her boyfriend was charged with assault and hindering apprehension; there was extended cross-examination of the guard and defense testimony by the woman herself and one eyewitness; the trial consumed 47 minutes.Google Scholar

170 Examples include No. 0440, a theft case before Judge OT (defendant allegedly took $40 from an open cash register; defender SR probed how clearly witness had seen incident; trial lasted 12 minutes); No. 2050, a marijuana possession case before Judge KA (police observed defendant smoking a joint and found small quantity of marijuana in paper bag on table beside him; defender BN brought out that bag could have belonged to other people at table and that quantity of drugs was insufficient to establish more than a summary offense; trial consumed 10 minutes); No. 2040, a prostitution case before Judge JM (defendant solicited plainclothes officer; private defense attorney GU vigorously developed entrapment argument; trial lasted 12 minutes).Google Scholar

171 Attorney preparation before trial is described supra at text accompanying notes 110–13.Google Scholar

172 For example, in No. 2541, a theft case, prosecutor TF seemed to have unexpected difficulty establishing the defendant's possession of the stolen goods and knowledge that they were stolen; after an awkward 53 minute trial, Judge OT acquitted, saying to the defendant that he was doing so “because the Commonwealth did not put on its case right, but I still think you did it.” A more technical slip-up occurred in No. 2612, a gambling case. After the arresting officer began to testify, prosecutor EE realized that he had failed to instruct the officer to bring the lottery slips with him to court; Judge GE (normally an extremely pro-prosecution judge) refused to grant a recess and ordered an acquittal. Inexperience and poor preparation sometimes caused similar problems for prosecutors on motions to suppress. In No. 1899, a gambling case, an intern seeking to establish probable cause was unable to elicit nonconclusory testimony from the arresting officer; Judge JM granted a motion to suppress, and prosecution was then withdrawn. In No. 0866, a drug case, prosecutor TU admitted that he was unfamiliar with the particulars of the defense motion and fumbled badly in his efforts to refute several of the claims raised; Judge KJ suppressed the evidence and entered an acquittal on all charges.Google Scholar

173 As previously noted, novice defender TS was sometimes unable to prepare all her cases fully despite her intense efforts to do so. See supra text accompanying note 111. In a few instances, defenders agreed to assume responsibility for cases on the very day of trial, when appointed counsel had failed to appear. E.g., No. 1324, discussed supra note 138. Similarly, in No. 3482, a drunk driving case, defender BN accepted appointment on the day of trial and had to read the prosecutor's file in court; nonetheless, BN appeared confident in pursuing the standard lines of defense in such a case, and she was able to show on cross-examination that the arresting officer (who had responded to an accident call) had not seen or personally determined that the defendant was the actual driver of the car. Judge CR found reasonable doubt and granted an acquittal.Google Scholar

174 For both prosecutors and defenders, municipal court trials are among the first assignments; as they become seasoned, the attorneys eventually graduate to more difficult and important responsibilities. At the Defender Association, however, more experienced attorneys typically rotate back into municipal court from time to time.Google Scholar

175 E.g., No. 2541, discussed supra note 172.Google Scholar

176 See, e.g., cases discussed supra note 170.Google Scholar

177 No. 1857, discussed supra note 153; No. 0842, a shoplifting case (Judge MN); No. 0792, an indecent assault case (Judge GO); No. 0283, a theft case (Judge VA); No. 2230, an assault case (Judge GE); No. 1350, a shoplifting case (Judge GE); No. 1447, a hit-and-run accident case (Judge GE).Google Scholar

178 It bears emphasis that degrees of adversariness are inherently resistant to quantitative measurement and that the data in this section are presented solely for purposes of supplementing, from an arguably more “objective” perspective, the essential qualitative material already considered. For discussion of the value and limitations of the quantitative material, see Schulhofer, supra note 21, at 1075–76.Google Scholar

179 The acquittal rate in the nonstipulated bench trials is also roughly comparable with reported acquittal rates in jury trial cases. For Philadelphia, the jury trial acquittal rate for 1981 was 27.5%. Schulhofer, supra note 21, at 1063 n.98. In one national sample, 30.3% of the jury trials ended in acquittal. Harry Kalven & Hans Zeisel, The American Jury 56 (1966). The judges who presided at the trials studied by Kalven and Zeisel reported that if the verdict had been left to them, they would have acquitted in only 16.7% of the cases. Id. The acquittal rate in the Philadelphia misdemeanor bench trials is thus substantially higher than the acquittal rate that judges alone would have generated in the presumptively adversarial jury trials in the Kalven and Zeisel sample.Google Scholar

180 For discussion, see Schulhofer, supra note 21, at 1069–70.Google Scholar

181 United States v. Cronic, 104 S. Ct. 2039, 2050 (1984).Google Scholar

182 Id. Even in minor cases on which a novice attorney may appropriately begin, however, there ought to be room, despite Cronic, for considering the adequacy of the novice's training and supervision. In some jurisdictions (including Philadelphia) it is still possible for relatively inexperienced attorneys without significant trial experience to be appointed as counsel in capital cases.Google Scholar

183 See supra note 161 and accompanying text.Google Scholar

184 See supra text accompanying note 31.Google Scholar

185 As indicated supra table 3, 25.8% of all the adjudications (i.e., decisions by trial or plea) were on pleas of guilty or nolo contendere, and 24.4% were by stipulated trial. Thus, 50.2% of the adjudications could be treated as functionally equivalent to pleas of guilty. For discussion of the sampling procedure and confirmation that the sample is representative of the court's caseload as a whole, see infra appendix 1.Google Scholar

186 See supra text accompanying notes 144–49.Google Scholar

187 See supra text accompanying notes 148–49.Google Scholar

188 See supra text following note 180. This conclusion seems consistent with the various quantitative measures of defense effort. See supra text accompanying notes 179–80.Google Scholar

189 Based on the figures in table 3, “slow pleas” would represent 3% of the nonstipulated trials, which in turn represent 49.1% of all adjudications (decisions by plea or trial); thus, slow pleas would constitute 1.5% of all adjudications. On a similar basis the nonadversarial stipulated trials (80–85% of all the stipulated trials) would account for 19.5–20.7% of all adjudications. The functional equivalents of the guilty plea (slow pleas, nonadversarial stipulated trials, and formal guilty pleas) would represent, altogether, 46.8–48% of all adjudications.Google Scholar

190 As indicated supra table 3, successful defense motions followed either by continuance pending possible appeal or by outright dismissal occurred in 31 cases or 4.9% of all the dispositions in the sample.Google Scholar

191 The 21% figure for Philadelphia reflects the assumption that genuine adversary trials represent 52% of the adjudications on the merits and the fact that such adjudications constituted 40% of all dispositions for the first half of 1984. The trial rate (as a percentage of total dispositions) was found to be 0% in New Haven, 3% in Columbus, and 3.5% in Austin. See supra note 32.Google Scholar

192 With respect to felonies, see Schulhofer, supra note 21, at 1083–84.Google Scholar

193 Id. at 1085–86.Google Scholar

194 Id. at 1056–57 & n.74, 1085.Google Scholar

195 Id. at 1085–86.Google Scholar

196 See supra text accompanying note 50.Google Scholar

197 See supra note 151.Google Scholar

198 This calculation assumes that of 100 cases, the 90 resolved by plea would require 900 minutes (10 minutes each) and the 10 resolved by trial would require 250 minutes (25 minutes each). On this basis the total court time required would be 1,150 minutes at a 90% plea rate; it would rise by 13% to 1,300 minutes (800 minutes for 80 pleas and 500 minutes for 20 trials) at an 80% plea rate. Using median disposition times, see supra notes 132, 152, the required court time would rise by 14%.Google Scholar

199 This calculation assumes that among 100 cases, the 26 resolved by guilty plea require 260 minutes (10 minutes each), the 24 resolved by stipulated trial require 192 minutes (8 minutes each), and the 50 resolved by nonstipulated trial require 1,250 minutes (25 minutes each). On this basis the total court time required is 1,702 minutes at the current disposition pattern; it would rise by 29% to 2,200 minutes (200 minutes for 20 pleas and 2,000 minutes for 80 trials) at a 20% plea rate. Using median disposition times, see. supra notes 132, 143, 152, the required court time would rise by 31%.Google Scholar

200 See Schulhofer, , supra note 21, at 1061.Google Scholar

201 Recent case law rejects the notion that a constitutionally valid guilty plea must be preceded by any particular form of colloquy and holds that the plea is valid so long as it can be shown that the waiver was knowing, intelligent, and voluntary under the totality of the circumstances. E.g., Morgan, Henderson V., 426 U.S. 637 (1976). In Pennsylvania, Pa. R. Crim. P. 319 (a) requires the judge, before accepting the plea, to determine “after inquiry of the defendant that the plea is voluntarily and understandingly tendered,” but again no particular litany is required and precise understanding of all the rights waived is unnecessary. E.g., Commonwealth v. Anthony, 475 A.2d 1303 (Pa. 1984) (guilty plea not defective where defendant not informed during colloquy that jury verdict must be unanimous; knowledge of the nature and consequences of the plea may be found from the totality of the circumstances). From the perspective of such decisions, the typical municipal court colloquy would seem to go well beyond the required minimum, and even a mass colloquy of 10 or 12 defendants probably would withstand scrutiny, since normally each defendant is at some point addressed personally and asked to acknowledge his understanding of what the group has been told. The objection to such a proceeding, rather, must be based on the assembly-line atmosphere that it creates and the danger of confusion inherent in any process that seeks simultaneously to establish a factual basis and obtain an acknowledgment of guilt in a dozen unrelated cases. The typical mass plea hearing is quite a different procedure from the one that the Supreme Court seems to have had in mind when it ratified bargained pleas as solemn admissions of guilt in open court. Yet, so far as I can determine, no court has considered this problem. Mass notification of rights (itself a very dubious procedure) has been approved in, e.g., Mills v. Municipal Court, 10 Cal. 3d 288, 515 P.2d 273 (1973); City of Cleveland v. Whipkey, 29 Ohio App. 2d 79, 278 N.E.2d 374 (1972); Crew v. Nelson, 216 N.W.2d 565 (S.D. 1974). The court rejected a challenge to a joint colloquy procedure in People v. Hailey, 67 Mich. App. 540, 241 N.W.2d 282 (1976), but the colloquy there involved three co-defendants jointly charged in the same case. The propriety of a joint guilty plea colloquy for 10 defendants charged with unrelated acts, and even unrelated kinds of offenses, is of course an entirely different matter.Google Scholar

202 The question has arisen in a number of jurisdictions, and nearly all have agreed that where a stipulation is “tantamount to” or “the functional equivalent of” a guilty plea, it must be preceded by the same waiver colloquy mandated in guilty plea cases. E.g., Brookhart v. Janis, 384 US. 1 (1966); People v. Levey, 8 Cal. 3d 648, 504 P.2d 452 (1973); People v. Smith, 59 Ill. 2d 236, 319 N.E.2d 760 (1974); Sutton v. State, 289 Md. 359, 424 A.2d 755 (1981); see Annot., 17 A.L.R. 4th 61 (1982). Pennsylvania law is in accord with this view. Commonwealth v. Davis, 474 Pa. 194, 322 A.2d 103 (1974); cf. Commonwealth v. Williams, 443 A.2d 338 (Pa. 1982) (conviction reversed on grounds of ineffective assistance of counsel where stipulation entered without proper waiver colloquy). The courts recognize that although a fully stipulated trial is often the functional equivalent of a guilty plea, such a trial may under some circumstances constitute a vigorously contested proceeding, and the decisions explore the factors that would indicate which of these is the case. See, e.g., People v. Phillips, 31 Cal. App. 3d 483, 107 Cal. Rptr. 386 (1973); People v. Young, 25 Ill. App. 3d 629, 323 N.E.2d 788 (1975). One court has apparently adopted a per se rule that a stipulation will never be deemed tantamount to a guilty plea. State v. Johnson, 38 Crim. L. Rep. (BNA) 2012 (Wash. Oct. 2, 1985).Google Scholar

203 In Philadelphia felony guilty pleas, which are almost always taken individually, the average time consumed is 22 minutes per case. See Schulhofer, supra note 21, at 1056 & n.71.Google Scholar

204 These calculations assume that among 100 cases, the 90 resolved by plea would require 1,350 minutes (15 minutes each) and the 10 resolved by trial would require 250 minutes (25 minutes each). On this basis the total court time required would be 1,600 minutes at a 90% plea rate; it would rise by 6% to 1,700 minutes (1,200 minutes for 80 pleas and 500 minutes for 20 trials) at an 80% rate and would rise by 25% to 2,000 minutes (750 minutes for 50 pleas and 1,250 minutes for 50 trials) at a 50% rate.Google Scholar

205 This calculation assumes that 100 cases resolved in accordance with current disposition patterns and current average times would require 1,702 minutes of court time. See supra note 199. The required time would rise by 17.5% to 2,000 minutes if the 50 pleas or stipulated trials consumed a full 15 minutes each and the 50 trials took 25 minutes each.Google Scholar

206 The 2,000 minutes of court time required at a 50% trial rate, see supra note 205, would rise by 15% to 2,300 minutes (300 minutes for 20 pleas and 2,000 minutes for 80 trials) at an 80% trial rate.Google Scholar

207 See supra text accompanying note 42.Google Scholar

208 Feeley, supra note 16, at 199–243.Google Scholar

209 See supra text accompanying notes 68–73 and supra note 112.Google Scholar

210 Indeed, Philadelphia's trial rate has risen markedly over recent years as the severity of sentencing has increased. See supra note 42. It is not clear, however, whether the proportion of genuinely contested cases has in fact increased or whether the statistics reflect a shift from guilty pleas to stipulated trials.Google Scholar

211 See supra text accompanying notes 69–74.Google Scholar

212 We were unable to collect systematic data on the number of court appearances required for each case in our courtroom sample.Google Scholar

213 Feeley, supra note 16, at 291; Alschuler, supra note 26, at 951–52.Google Scholar

214 Further research on this question would require systematic interviews with affected defendants.Google Scholar

215 Hart, The Aims of the Criminal Law, 23 Law & Contemp. Probs. 401, 404 (1958).Google Scholar

216 See Becker, Howard S., Outsiders (1963); Edwin M. Schur, Labeling Deviant Behavior (1971).Google Scholar

217 The collateral consequences of conviction have received virtually no rigorous attention in comparative studies of courts. Indeed, the law and economics research, which in other respects brings so much precision to the study of sanctions, has tended to ignore this dimension by conflating crime and tort so that the distinctive collateral features of the criminal sanction disappear from the analysis. See, e.g., Becker, , Crime and Punishment: An Economic Approach, 76 J. Pol. Econ. 169, 201 (1968) (“when crimes are punishable by fines, the analytical differences [from tort law] virtually vanish”); for discussion, see Schulhofer, , Is There an Economic Theory of Crime 26 NOMOS 329, 336–39 (1985). With respect to collateral consequences in New Haven, Professor Feeley noted the elusive nature of the measurement problem. He cautioned that despite the relatively cavalier attitude toward criminal sanctions displayed by most of his informants: the impulse for formality, even with its manifest shortcomings, cannot be so quickly dismissed. While lower courts sentence very few people to terms in jail, in theory almost all of those appearing before them face a slim possibility of incarceration. While creating a record of petty criminal offenses may not significantly affect the future of most people … it can have a long-lasting and unpredictable impact on some. Citizenship can be placed in jeopardy, careers destroyed, aspirations dampened, delinquent propensities reinforced. Such problems may be few in number but they do occur. And it is impossible to tell in advance which cases may precipitate these more serious consequences, since the specific impact of a record may not make itself felt until much later in life. Feeley supra note 16, at 242.Google Scholar

218 See supra text accompanying notes 69–71.Google Scholar

219 See supra text accompanying note 70. The defenders' relative disinterest in trying to persuade clients to plead guilty is underscored by the fact that defenders at crash court frequently do not even bother to present the prosecutor's offer to their clients. See supra note 64 and accompanying text.Google Scholar

220 See Alschuler, , The Defense Attorney's Role in Plea Bargaining, 84 Yale L.J. 1179 (1975).CrossRefGoogle Scholar

221 See id. at 1281. See also Mills, I Have Nothing to Do with Justice, Life Magazine, Mar. 12, 1971, at 56: “When will [trial] be?”. “In a couple of months. Maybe longer.” Santiago has a grip on the bars. “You mean if I'm guilty I get out today?”“Yes.”“But if I'm innocent, I have to stay in?”“That's right.”Google Scholar

222 Although my argument focuses on factors that may explain different perceptions of sanctions, when the sanctions themselves remain constant, it is also possible that a movement toward a more informal processing system could lead the concrete, measurable component of the sanction to become “objectively” less severe. Malcolm Feeley has suggested similar possibilities: “[T]here is a chance that increased formalism would increase the harshness of the criminal process for a great many defendants… While [stricter formalities] might result in a higher appearance rate in court and vindication of the innocent in some cases, and could conceivably lead to higher standards of police conduct, they could also lead to a tendency to invest a great many cases with a degree of seriousness that few people [in New Haven] currently give to them.” Feeley, supra note 16, at 292–93.Google Scholar

223 In the Boston Municipal Court, the reported guilty plea rate is even lower than Philadelphia's—only 6% of all misdemeanor dispositions by trial or plea in 1983. Annual Report of the Massachusetts Trial Court—1983, at 27 (1984). The high trial rate is partially explained by the Massachusetts procedure of “admission to sufficient facts.” Under this procedure the defendant pleads not guilty but stipulates to facts sufficient to warrant a finding of guilt, in the expectation of either claiming a trial de novo in the jury trial session or obtaining a “continuance without finding.” The latter procedure results in imposition of conditions of probation, with provision for dismissal of all charges on successful completion of the probationary term. See Smith, Kent B., 30 Massachusetts Practice: Criminal Practice and Procedure § 1202, at 676–77 (St. Paul: West Publishing Co., 1983); Shapiro, Should a Guilty Plea Have a Preclusive Effect? 70 Iowa L. Rev. 27, 42 n.75 (1984). The procedure for continuance without finding bears many similarities to a pretrial diversion program, but it requires formal admissions, and if the defendant violates the probation, a finding of guilt can be entered without further trial. (The defendant, however, can then claim a trial de novo.) Under these circumstances the Boston statistics clearly overstate the proportion of fully contested cases. Indeed, a Massachusetts study committee estimated that statewide roughly 40–65% of all lower court cases could fall in the “continuance without finding” category. Administrative Office of the [Massachusetts] District Court Dep't, Report of the Committee on Juries of Six: Elimination of the Trial De Novo System in Criminal Cases 63 (Jan. 1984). (I am indebted to Professor David Shapiro for providing me with a copy of this report.) It would appear, nonetheless, that a substantial number of the Boston misdemeanor cases may be the subject of reasonably adversarial trials; the same study committee estimated that the average time for a nonstipulated misdemeanor bench trial would be about 30 minutes. Id. at 47, 51. Plainly, the Boston situation is one that would repay careful study. Very high misdemeanor trial rates also appear to be suggested by statewide statistics reported for Delaware (100%!) and Iowa (52%). see id. appendix C, but the available figures lack detail sufficient to warrant specific conclusions. The contrasting situation is epitomized in New York City, where trials account for only 0.8% of all dispositions by trial or plea in misdemeanor cases (those involving a fine less than $1,000 or imprisonment less than one year). Id.Google Scholar

224 Referring to table 12 supra, one can observe that cases ending in substantial charge reduction, acquittal, or dismissal accounted for 37% of the cases tried. Moreover 28 cases (10%) ended in complete conviction but involved third-party defense witnesses, a strong indication that the cases presented seriously triable issues. Of course this attempt at estimation ignores all the conviction cases in which significant defense issues were developed solely by cross-examination or defendant testimony; presumably many cases in those categories were also far from “open and shut.” For discussion of the problem of estimating “dead bang” cases, see Schulhofer, supra note 21, at 1081–82.Google Scholar

225 See cases discussed infra notes 228–31. A Boston defense attorney quoted by Albert Alschuler underscored the same point: “There may be no objective ground for optimism at the beginning, but I've won many apparently hopeless cases. The evidence only looks open and shut.” See Alschuler, supra note 220, at 1282 (emphasis in original).Google Scholar

226 E.g., Utz, , supra note 148, at 129–48; Enker, Arnold, Perspectives on Plea Bargaining, in President's Comm'n on Law Enforcement & Admin. of Justice, Task Force Report: The Courts 112 (1967); Church, In Defense of “Bargain Justice,” 13 Law & Soc'y Rev. 509 (1979); Easterbrook, Criminal Procedure as a Market System, 12 J. Legal Stud. 289, 316–17 (1983).Google Scholar

227 For an overview of the issues presented, see Eisenberg, , Private Ordering Through Negotiation: Dispute-Settlement and Rulemaking, 89 Harv. L. Rev. 637 (1976); and sources cited in Schulhofer, , supra note 21, at 1039 n.1.Google Scholar

228 E.g., No. 0684, involving a charge of carrying a firearm in public. The defendant claimed that the gun was in his apartment when the police seized it, but the police report stated that the defendant was outside, walking away from an argument, when the gun fell out of his waistband. When pressed on cross-examination about the police version, the arresting officer admitted, “That's not quite what happened.” Ultimately, Judge JM (a very pro-prosecution judge) acquitted. In No. 0280, the defendant was charged with resisting arrest and assaulting a police officer, but the facts indicated that the defendant himself had not suffered any physical injuries. Judge MN rejected the police version of the event and convicted only of disorderly conduct: He reasoned that the officer's testimony couldn't possibly be true because the defendant had never received “a PT—police tune-up.” Other examples include No. 0228, an assault case before Judge KA (complainant's testimony differed from police report); No. 1161, a theft case before Judge JM (testimony revealed presence of another suspect at the scene and thus opened up possibility of motion to suppress); No. 2759, a drug case before Judge CR (officer refused to take stand under circumstances suggesting that he knew his testimony would be rejected as false); No. 3244, a drunk driving case before Judge OM (defendant contradicted himself on stand and finally conceded he “might” have done acts alleged).Google Scholar

229 In Nos. 1448 and 1449, the defendants were charged with resisting arrest and assaulting an officer. At trial the arresting officer spoke in a very low voice, and when the judge asked him to speak up, the officer became anxious and angry; Judge KJ decided that the officer was lying and acquitted on all charges. In No. 1092, a drunk driving case, the defendant had refused to take the breathalizer test, and the case rested in part on evidence of his slurred speech at the time of arrest. Testifying at trial, the defendant showed convincingly that he had a natural stutter, and Judge KA, finding reasonable doubt, acquitted. Similar examples include No. 2632, an assault case (complainant appeared drunk and incoherent during testimony; Judge GP acquitted); No. 0703, a drunk driving case (breathalizer reading was borderline and arresting officer gave hesitant answers to crucial questions; Judge GA found reasonable doubt and acquitted).Google Scholar

230 This was the recurring situation when neighborhood scuffles produced cases riddled with conflicts over who produced which weapon when, and who was acting in self-defense. E.g., No. 2091, an assault case (Judge KA); No. 1986, an assault case (Judge JM); No. 0423, a case involving charges of assault and resisting arrest (Judge GA). Other examples include No. 1128, discussed supra note 159; Nos. 2416 and 2417, discussed supra note 168; No. 1092, discussed supra note 229; No. 0228, discussed supra note 228; No. 3244, discussed supra note 228.Google Scholar

231 E.g., No. 1352, an assault case (Judge OT); Nos. 2416 and 2417, discussed supra note 168. See also cases discussed supra notes 229 and 230.Google Scholar

232 A fruitful, and more scientific, alternative would be to take the files in a sample of Philadelphia cases and to ask defense attorneys accustomed to plea bargaining (in New York, for example) to provide an evaluation of those files as they stood at various stages of the pretrial process—would a plea be advisable? on what terms? would more information be needed? Conversely, Philadelphia defenders could be asked to evaluate files on which guilty plea agreements were reached in a city such as New York. My guess is that such research would reveal enormous disparities in case evaluation standards from one court culture to another.Google Scholar

233 See Alschuler, , supra note 220; id., The Prosecutor's Role in Plea Bargaining, 36 U. Chi. L. Rev. 50 (1968). Plea bargaining proponents often dismiss the conflict of interest problem on the ground that conflicts can affect trial performance as well. See, e.g., Easterbrook, supra note 226, at 309. This response ignores the fact that the conflicts of interest identified, for example, by Alschuler, severely distort the attorney's preferences concerning whether to go to trial, but have much less, if any, impact on preparation and trial performance once the trial route is chosen.Google Scholar

234 See Schulhofer, , Effective Assistance on the Assembly Line, N.Y.U. Rev. L. & Soc. Change (forthcoming). Judges sometimes engage in this tactic as well. The New York Times recently reported a vivid example from the New York City criminal courts. When a burglary case was called, one of the two defendants did not have a lawyer, and rather than continue the case, the judge appointed an attorney from among those waiting in court. The D.A. then offered a plea, with a 2 to 4 year sentence. The presiding judge, Harold J. Rothwax, added, “After today, it's 3 to 6, after that it's 4 to 8. If they're ever going to plead, today is the time to do it.” When the defendants rejected the offer (after all, one of the lawyers had been appointed only moments before), Judge Rothwax declared, “We'll make it very easy. It's 4 to 8 after today. Let's play hardball.” N.Y. Times, Apr. 29, 1985, at 13, col. 1, 3.Google Scholar

235 See Yale Kamisar et al., Modern Criminal Procedure 1153–59 (5th ed. 1980).Google Scholar

236 Id. at 1158–59. A proposal to require federal prosecutors to disclose their witness lists was explicitly rejected by Congress in 1975. See id. at 1159 n.g.Google Scholar

237 Id. at 1163–67. Kamisar et al. state, “Without doubt, the single item most commonly held not to be discoverable is the statement of a prospective government witness that was given to a police officer or prosecutor.”Id. at 1159. The prosecution is constitutionally obligated to turn over to the defense, on request, any materially exculpatory evidence. United States v. Brady, 373 US. 83 (1963). Brady and its progeny make clear, however, that there is no general disclosure obligation; for example, witness statements vaguely inculpating the accused need not be disclosed, even though gaps and inconsistencies in such statements could be of crucial value in establishing a defense. See United States v. Agurs, 427 U.S. 97 (1976).Google Scholar

238 See Kamisar et al., supra note 235, at 1176–91.Google Scholar

239 Id. at 1151.Google Scholar

240 Id. at 1168–70. In most jurisdictions, the defendant can take a witness's deposition only for purposes of preserving testimony; therefore, the defendant must be able to show that the person to be deposed is unlikely to be able to attend the trial. Id. at 1168. Only a few states authorize depositions for purposes of general discovery. E.g., Ariz. R. Crim. P. 15.3 (a); Fla. R. Crim. P. 3.220 (d); Vt. R. Crim. P. 15 (a); Wash. Ct. R. 4.6 (a).Google Scholar

241 In that event, the defendant can be held for trial on the basis of grand jury testimony taken without any opportunity for cross-examination. Moreover, the transcript of the witness's grand jury testimony is generally subject to the same discovery limitations that apply to other witness statements; in some jurisdictions, discovery of the transcript is even more limited. See Kamisar et al., supra note 235, at 1167. Often the defense is not permitted to inspect the transcript until after the witness has testified at trial. E.g., 18 U.S.C. § 3500 (1982).Google Scholar

242 The victim of a rape or assault is seldom anxious to talk to the attorney representing the person who she thinks is her attacker, and the prosecutor will explain that she is not obliged to do so. Police officers, likewise, seldom agree to be interviewed before trial by defense investigators.Google Scholar

243 See Kamisar, et al., supra note 235, at 1187–88. In some jurisdictions, the defense is required to provide the names of particular kinds of defense witnesses—e.g., psychiatric experts or alibi witnesses. See, e.g., Wardius v. Oregon, 412 U.S. 470 (1973); Williams v. Florida, 399 U.S. 78 (1970); Fed. R. Crim. P. 12.1, 12.2.Google Scholar

244 A thoughtful proposal to provide a better informational foundation for plea bargaining is developed in Arenella, Reforming the Federal Grand Jury and the State Preliminary Hearing to Prevent Conviction Without Adjudication, 78 Mich. L. Rev. 463 (1980).Google Scholar

245 In a study of practice in Vermont, where general discovery depositions are authorized, the author found that depositions were probably taken in less than 8% of all cases. Langrock, Vermont's Experiment in Criminal Discovery, 53 ABA J. 732 (1967).Google Scholar

246 See, e.g., No. 0280, discussed supra note 228.Google Scholar

247 For examples of this phenomenon in plea bargaining jurisdictions, see, e.g., Vera Institute of Justice, Felony Arrests 37–42 (N.Y.: Vera Institute of Justice, 1981) (on New York City); Utz, supra note 148, at 101–12 (on Oakland, California).Google Scholar

248 Enker reports this view, supra note 226, which many practicing attorneys appear to hold simultaneously with their opinion that the overwhelming majority of cases are open and shut. Heumann, , studying the Connecticut courts in the early 1970s, reported that experienced attorneys describe the great majority of criminal cases as devoid of debatable questions of fact or law. Heumann, Milton, Plea Bargaining 58, 60 (1978); id., Author's Reply, 13 Law & Soc'y Rev. 650, 651 (1979). In contrast, Feeley, studying some of the same courts at about the same time as Heumann, stressed that the facts are “malleable.” Feeley, supra note 16, at 168–77.Google Scholar

249 E.g., Church, supra note 226, at 516; Easterbrook, supra note 226, at 316; Heumann, Author's Reply, supra note 248, at 652.Google Scholar

250 For example, Feeley reports: “Because they feel that the consequences in most cases are so minimal, prosecutors and defense attorneys have little incentive to scrutinize their cases carefully. This casualness leads to a great many errors.” Feeley, supra note 16, at 168–69. See also id. at 178–79 (“Only in rare cases… will the prosecutor carefully read through the file.”).Google Scholar

251 Note that any imperfections in the trial system (e.g., an “unwarranted” tendency to favor police witnesses or an “inadequate” standard of reasonable doubt) are very likely to play back into bargained outcomes by affecting negotiators' expectations about the outcome at trial. Thus, plea bargaining systems should in all probability suffer from any defects in the jurisdiction's trial system, together with the particular distortions unique to plea bargaining.Google Scholar

252 The Miami pretrial settlement procedure, observed and recorded under conditions likely to minimize care and sensitivity on the part of all participants, nonetheless involved exceedingly perfunctory, superficial discussions, and the average time consumed was only 10 minutes per case. Heinz, & Kerstetter, , Pretrial Settlement Conference: Evaluation of a Reform in Plea Bargaining, 13 Law & Soc'y Rev. 349, 357–58 (1979).CrossRefGoogle Scholar

253 For Philadelphia misdemeanors, the average trial requires 25 minutes in court, but a guilty plea requires 10 minutes in court after settlement negotiations are completed, and the typical plea would probably require about 15 minutes if mass colloquies were eliminated. See supra text accompanying notes 132, 152, 203.Google Scholar

254 In the Heinz, and Kerstetter, study of Miami, the average settlement conference required 10 minutes. Heinz & Kerstetter, supra note 252, at 358. A study of misdemeanor plea negotiation in a California municipal court found that the average case required 10 minutes of discussion. Maynard, , The Structure of Discourse in Misdemeanor Plea Bargaining, 18 Law & Soc'y Rev. 75 (1984).Google Scholar

255 See Schulhofer, , supra note 234.Google Scholar

256 In this sense, it is probably true that a trial-oriented system will be more expensive to operate than one relying primarily on guilty pleas. The resource costs arise not because individual cases require more processing time, but because authorities responsible for a trial-oriented system will likely choose to spend more on investigation, training, supervision, and personnel.Google Scholar

257 E.g., James Eisenstein & Herbert Jacob, Felony Justice (Boston: Little, Brown & Co., 1977).Google Scholar

258 If the key to the vigor of the adversary process is an independent court culture rather than dispositional formalities as such, then the adversarial Philadelphia attorneys presumably would pursue tactics of vigorous opposition in plea bargaining as well as in trials. A comparative study along these lines would be of great value. Cf. Utz, supra note 148. Nevertheless, the central point remains that for any given city, with all other things (including adversarial culture) being equal, a system relying primarily on trials appears likely to afford greater reliability in the short run and better resources, training, and personnel in the long run than a system relying primarily on guilty pleas.Google Scholar

259 See Schulhofer, , supra note 21, at 1096–1100. Much of the explanation for the high trial rate, and especially for the vigorously adversarial behavior of public defenders in those trials, can be traced to the training and supervision they receive at the Defender Association. Like younger attorneys anywhere, they are anxious to develop their skills and to make a good impression; these factors alone contribute in part to their performance. But in addition, their supervisors and more experienced colleagues impress on them that, as senior defender RR put it, “They are sent to MC [Municipal Court] with the idea of trying cases.” They learn that they can win cases (even before elected judges who could hardly be described as pro-defense) and that when they don't like the result, the office will back them up and appeal. Indeed, the office strategy has been to “uplift the MC” by making the judges follow the appellate case law. And some defenders report that as a result of this orientation, their clients feel better about the whole process. It is undoubtedly true, nonetheless, that these defenders would strongly recommend guilty pleas if the price were right. A sine qua non for the high trial rate, therefore, has to be the attitude of the judges and prosecutors. The judges refuse to exact a heavy penalty for taking cases to trial. Policy in the D.A.'s office is essentially two pronged: If guilt is clear, the defendant must face the full trial sentence; if guilt is not clear, the case should be strengthened and then proved, if possible, in court. See supra note 67 and accompanying text. Such attitudes apparently have been long-standing in Philadelphia. See Schulhofer, supra note 21, at 1098–99. Indeed, data collected by Albert Alschuler suggest that as far back as the 1920s, guilty pleas were less common in Philadelphia than elsewhere. Alschuler, supra note 26, at 1024–25. Nonetheless, it remains unclear how these attitudes took root and how deeply they are implanted. Not only can prosecutorial policy easily change, see Schulhofer, supra note 21, at 1099, but adding to the puzzle is the fact that in the local federal district court, staffed by judges and attorneys drawn from the same Philadelphia legal culture, guilty pleas and plea bargaining are commonplace, as in most other American jurisdictions. A fuller understanding of the evolution of Philadelphia's distinctive court culture plainly must await further research.Google Scholar

260 Kirp, & Jenson, , What Does Due Process Do 73 Pub. Interest 75, 90 (1983).Google Scholar

261 Id. at 75. See also Simon, Legality, Bureaucracy, and Class in the Welfare System, 92 Yale L.J. 1198 (1983).Google Scholar

262 People v. Griffin, 7 N.Y.2d 511, 516, 166 N.E.2d 684, 687 (1960) (approving the validity of guilty pleas tendered under such circumstances).Google Scholar

263 See supra text accompanying note 101.Google Scholar

264 In order to minimize the effect of our study on the behavior of courtroom participants, we began by attempting to observe proceedings from the public section without making our presence or purpose known to the judge or the attorneys. This procedure proved impossible to sustain, however, because court officers usually would not permit members of the public to take notes, and because much of the interaction between participants could not be heard clearly from the public section. We therefore obtained permission to sit close to the bench and the witness stand, though at the price of disclosing our presence as researchers to the court officer and often to the judge. Because the general tone of the proceedings did not seem in the least affected by that disclosure, we doubt that observation effects were significant. With respect to the possiblity of seasonal bias, see Schulhofer, supra note 21, at 1054 n.66.Google Scholar

265 Of course, there is no way to use matching procedures to construct a perfect replica to the actual caseload. It is, therefore, conceivable that within a given offense category, a disproportionately high number of stipulated trials could have been steered away from the courtrooms we observed. Nevertheless, the court's case-assignment procedure makes it extremely unlikely that this could have occurred. There thus appears to be every reason for confidence that the sampling procedure provides a reliable indication of the minimum rate of truly contested cases.Google Scholar

266 This figure can be at best only a rough approximation. It is obtained by adding the uncontested dispositions among the reported trials to the uncontested guilty plea dispositions. Assuming that all the stipulated trials are uncontested and that all the remaining trials are contested (but see supra text accompanying notes 186–89 for qualifications to these assumptions), one obtains this figure by multiplying the observed stipulation rate by the reported trial rate and adding this product to the reported guilty plea rate. (I am indebted to James Eisenstein for suggesting the addition of this column to the table).Google Scholar

267 See supra notes 39–40 and accompanying text.Google Scholar