UNJUSTIFIED SENTENCE DISPARITY: A CASE STUDY OF THE LEVELLING EFFECT OF PAROLE Robert A. Weninger* Of all the systems that have been affected by the recent sentencing reform movement in the United States, none has been more vulnerable than "indeterminate" sentencing, the practice whereby judges and parole boards share vast discretion in fixing prison terms within broad ranges set by statute.' Critics vary in the reforms they propose, but complain in unison about arbitrary uses of discretion, race and class discrimination, and widespread disparity in the treatment of offendem2 * Professor of Law, Texas Tech University. B.B.A. 1955, LL.B. 1960, University of Wiconsin; LL.M. 1964, University of Chicago. This research was supported by a Walter E. Meyer Research Grant of the American Bar Foundation, the Texas Tech Law School Foundation, and the Texas Tech University Office of Research Services. The writer thanks the members, commissioners, and staff of the Texas Board of Pardons and Paroles for their assistance in the data collection effort and Michael R. Gottfredson, Professor of Criminal Justice a t the Claremont Graduate School, for having reviewed an earlier draft of this article. Any errors that remain are, of course, the writer's alone. The views expressed are his, and are not necessarily those of the ABF. 1. See Messinger, Introduction to A. VONHIRSCH& K. HANRAHAN, THE QUESTIONOF PAROLE a t xi-xii (1979). Indeterminate systems vary from state to state but exhibit certain common characteristics. Initially, the judge prescribes a sentence that almost everyone understands to be much longer than the offender is expected to serve. Id. a t xi. Through its parole release decision, a c e n t r a l i d board determines the actual period of confinement within the constraints of maxima or minima iixed by statute or a requirement that offenders serve a fixed proportion of their sentences before parole eligibility. Ordinarily, the board has discretion to set widely varying terms of imprisonment for offenders appearing before it. Moreover, most boards can delay making term-fixing decisions until the offender has spent a substantial period of time in confinement. Id. a t xii. Nor does parole release end the board's authority over the offender. Following release on parole, an offender remains under "parole supervision," usually until the court-imposed sentence is discharged. During this period the parolee is subject to reimprisonment at the discretion of the board for the failure to abide by the rules i t has imposed. Id. 2. Criticisms of indeterminate sentencing can be categorized. One k i d concerns rehabilitation as the primary goal of sentencing; investigators have questioned whether it is pos& J. WILKS,THEEFFECTIVENESS sible to rehabilitate offenders. See D. LIPTON,R.MARTINSON, OF CORRECTIONAL TREATMENT: A SURVEYOF TREATMENT EVALUATION STUDIES(1975); Greenberg, The Correctional Effects of Corrections: A Survey of Evaluations, in CORRECTIONS Heinonline - - 36 Syracuse L. Rev. 715 1985-1986 Syracuse Law Review [Vol. 36:715 This article reports the results of a study of whether parole operates to reduce sentence disparity in a system where sentencing is indeterminate,3 and where the parole release decision is not regulated by the use of explicit decisionmaking guidelines.' The inves- AND PUNISHMENT (D. Greenberg ed. 1977). Another kind of criticism concerns the procedures used to determine sentence types and length. The proposals for reform of the sentencing process range from relatively modest measures, such as sentencing councils or appellate review of sentences, to major structural reforms, including the abolition of parole release and the establishment of determinate sentencing systems. The literature is vast. See generally AMERICAN FRIENDS SERVICE COMM.,STRUGGLE FOR JUSTICE: A REPORTON CRIMEAND PUNISHMENT IN AMERICA (2d ed. 1971); D. FOGEL,WE ARE THE LIVINGPROOF:THEJUSTICE MODEL FOR CORRECTIONS (1975); M. FRANKEL, CRIMINAL SENTENCES: LAWWITHOUT ORDER(1973); D. STANLEY, PRISONERS AMONG US (1976); E. VAN DEN HAAG,PUNISHING CRIMINALS (1975); k VON HIRSCH,DOINGJUSTICE: THECHOICEOF PUNISHMENTS (1976); J. WILSON,THINKING ABOUTCRIME(1975); Harris, Disquisition on the Need for a New Model for Criminal Sanctioning Systems, 77 W. VA.L. REV.263 (1975). These proposals have a common objective: to structure and confine discretionary sentencing power. 3. The statute on indeterminate sentences in Texas provides: If the verdict fixes the punishment a t confinement in an institution operated by the Department of Corrections for more than the minimum term, the judge in passing sentence shall pronounce an indeterminate sentence, fixing in such sentence as the minimum provided by law as the lowest term in an institution operated by the Department of Corrections and as the maximum the term stated in the verdict. TEx. CODECRIM.PROC.ANN.art. 42.09, $ 1 (Vernon 1979). Prisoners convicted of any certain aggravated offenses become eligible for release on parole only after they have served one-third of their maximum terms. Depending on the amount of their good-conduct time, prisoners convicted of other offenses may become eligible for parole before they have served one-thud of their terms. The principal provision of the statute on parole in Texas provides as follows: A prisoner under sentence of death is not eligible for parole. If a prisoner is serving a sentence for the offenses listed in Section 3f(a) 2 of this Article or if the judgment contains an affirmative finding under Section 3a(f)(2) of this Article, he is not eligible for release on parole until his actual calendar time served, without consideration of good conduct time, equals one-thud of the maximum sentence or 20 calendar years, whichever is less, but in no event shall he be eligible for release on parole in less than two calendar years. All other prisoners shall be eligible for release on parole when their calendar time served plus good conduct time equals onethird of the maximum sentence imposed or 20 years, whichever is less. TEX.CODECRIM.PROC.ANN.art. 42.12, $ 15(b) (Vernon 1979). Good conduct time is the product of a complex of rules and discretion administered by the Texas Department of Corrections a t a varying rate of from zero to 25 days for every month served in the institution, depending on the prisoner's classification as determined by his conduct. Good conduct time reduces the maximum sentence and advances parole eligibility by the reduction of the maximum sentence. See TEx. REV.CN. STAT.ANN,art. 6181-1, $8 2-5 (Vernon Supp. 1985). See also V. O'LEARY& K. HANRAHAN, PAROLE SYSTEMS IN THE UNITEDSTATES:A DESCRIPTION OF THE^ S T R U ~ U R AND E FUNCTION 300 (3d ed. 1976). 4. Effective August 1,1983, the Texas parole board adopted, on an experimental basis, a numerical selection assistance scale that provides a guideline for parole release decisions. See TEXASBOARDOF PARDONS AND PAROLES, EXPERIMENTS GUIDELINES FOR PAROLEDBCI- Heinonline - - 36 Syracuse L. Rev. 716 1985-1986 19851 Parole and Sentence Disparity 717 tigation is based on observation of the behavior of participants in the parole process in Texas, interviews with parole officials, and a statistical analysis of a random sample of cases involving 1183 adult offenders released on parole in that state in 1981.5 The question of disparity reduction is examined from the perspective of the interplay between the parole board and the judiciary. Contingency tables present the results of an analysis of quantitative data on sentence length and time served for comparable groups of offenders. Examination of the data in these tables shows whether parole board decisions were less disparate than judicial sentencing decisions. Observation and interview data were used to investigate whether disparity reduction is a latent function of the parole process, or whether it is a central concern of administrative decisionmaking. This study suggests that parole can mitigate disparity even in a large and diverse state where sentencing and parole authorities exercise unregulated discretion. It makes two tentative findings. First, there appears to be greater disparity in sentences imposed by judges on offenders with more serious offenses and extensive criminal histories than on offenders with less serious offenses and prior records. Second, the worse the offense and prior record, the more likely it is that parole will reduce sentence disparity. Although these findings suggest that administrative release decisions can enhance equity in sentencing, the study concludes that in s r o ~ (Aug. s 1983); infra note 142. The study reported in this article, however, is based on a random sample of cases involving offenders released on parole in 1981, before adoption of the new system, when the board acted without explicit guidelines. 5. The writer based his study on both qualitative and quantitative data. In 1983 and 1984, he met with an interviewed most of the six members and nine commissioners of the Texas Board of Pardons and Paroles, and many of the employees comprising the staff of the agency, including the Executive Director, Director of Staff Services, planners, and institutional parole officers. In May 1984, the writer observed parole hearings conducted a t state prison facilities located in Gatesville (Riverside Unit), Huntsville (Ferguson Unit), and Palestine (Beto I1 Unit). The parole board furnished the writer with a tape containing data regarding a random sample of 2500 offenders released from prison during 1981, of when 1183 were released on parole. The data concerning each offender included sentence length, actual time served, case attributes, and scores relevant to the guidelines system then the subject of research by the parole board. See text accompanying infra notes 144-145. The procedure used for the random sample was that provided for in the software for the Burroughs computer used by the agency. The board's researchers compared the relevant characteristics of offenders in the sample with those of the population of inmates in the state prison system and found the sample to be representative as to age, race, sex, offense type, and sentence length. Heinonline - - 36 Syracuse L. Rev. 717 1985-1986 718 Syracuse Law Review [vol. 36:715 Texas, parole's capacity to mitigate sentence disparity has been insufficiently used, and that a system of structured discretion for parole officials would be preferable to one that is largely unguided. The decade of the 70's produced a variety of innovations in American sentencing laws and pra~tices.~ Indeterminate sentencing, in widespread use for more than half a century, lost ground to statutory determinate sentencing, mandatory minimum sentences, and sentencing and parole guidelines.' Reformers were motivated, in large part, by a desire to introduce greater certainty into sentencing and to reduce disparity by narrowing the discretion of judges and parole board^.^ The results of these reforms have been 6. The dynamic changes during the 1970's were noted by the Panel on Sentencing Research (PSR), See infra note 9, noted some of the changes: Indeterminate sentencing systems were in widespread use until the 1970's and had not changed materially for 50 years . . . By 1982, however, most jurisdictions had made dramatic changes in their sentencing practices and institutions. Parole release had been abolished for the majority of prisoners in as many as 10 states, and parole guidelines had been established in a t least 8 others. Determinate sentencing statutes, under which prisoners could predict their release dates a t the time of sentencing assuming good behavior in prison, were in effect in more than 10 states, and mandatory minimum sentence laws were in effect for some offenses in more than 30 states. Several stat- had adopted statewide sentencing guidelines, and local sentencing guidelines had been establihsed in more than 50 jurisdictions. I RESEARCH ON SENTENCING: THESEARCH FOR REFORM 2 (A. Blumstein, J. Cohen, S. Martin & M. Tomy eds. 1983) [hereinafter cited as PANEL REPORT]. 7. See PANEL REPORT,supra note 6, a t 60-61. PSR said that dissatisfaction with interdeterminate sentences was precipitated by six major factors: (1) Prison uprisiiga in the 60's that led to an analysis of the purposes and institutions of punishment from the perspective of prisoners, (2) concern over the broad discretion residing in judges and parole boards, (3) a demand for accountability in official decision makiig, (4) disbelief that rehabilitation, the goal of interdeterminate sentencing, was possible, (5) studies showing widespread disparity and racial and class discrimiiation in sentencing, (6) public concern that lenient judges and parole officials returned dangerous offenders to the community. See id. at 2-3. 8. See id. at 3-4. PSR reported that "despite the growing diversity of factors considered and the increasing methodological sophistication of statistical analysis of sentencing, large portions-two thirds or more--of variance in sentence outcomes remain unexplained." Id. at 80. The panel distinguished between four types of disparity, stating that each needed to be evaluated differently. "First, there may be only the appearance of disparity. This occurs when cases seem alike to an outside ovserver but differ materially in case attributes observed by the judge." Id. at 75. For example, one defendant might exhibit remorse while another, otherwise a like offender, does not. PSR said that this seeming disparity may be reduced by improved observation of variables affecting sentencing. Id. The second type of disparity is that which is deliberately introduced as a matter of social policy. For example, it has been decided that it is sufficient to single out and punish only one of several tax evaders. Such disparity may be unjust to one who espouses equal . Heinonline - - 36 Syracuse L. Rev. 718 1985-1986 Parole and Sentence Disparity the subject of a growing body of social science research: and investigators have questioned whether particular innovations reduced unjustified disparity in sentencing.I0 This article reviews some of their findings and describes how sentence disparity is sought to be treatment for like offenders, but may be tolerated by one who believes that " 'like' offenders are entitled to only an equal opportunity of receiving a particular sentence." Id. Third, there is "interjurisdictional disparity such as that found between urban and rural courts in the same state." Id. Such geographical disparity may be the product of differences in community attitudes toward crime and punishment. Whether such disparity is warranted depends on one's "concern for evenhandedness or uniformity of standards versus the value of preserving local community control." Id. a t 75. A fourth type of disparity relates to individual judges whose differences in sentencing are explained by their varying philosophies, experiences, and backgrounds. Some observers might tolerate such disparity as a reflection of acceptable variation in attitudes toward sentencing within a community. Others might deplore such variation as inequitable, inconsistent with the rule of law, and having undesirable consequences, like "judge shopping." See id. at 76-77. PSR also recognized the concepts of "diicrimination" and "disparity" and distinguished between them in terms of the legitimacy of the criteria for determining sentences and the consistency with which those critera are applied to similar cases: Discrimination exists when some case attribute that is objectionable (typically on moral or legal grounds) can be shown to be asssociated with sentence outcomes after all other relevant variables are adequately controlled. . . .Race is the clearest example of an illegitimate criterion. . . . Disparity exists when "like cases" with respect to case attributes-regardless of their legitimacy-are sentenced differently. For example, this might occur when judges place different weights on the various case attributes or use different attributes in their sentencing decisions. See id. a t 72 9. See PANELREPORT.supra, note 6. The Panel on Sentencing Research was convened to review the research on sentencing, assess its quality, and suggest directions for further inquiry. PSR was established by the National Research Council in 1980 as a panel of the Committee on Law Enforcement and the Administration of Justice. Its members represent a variety of academic disciplines and methodological approaches in the criminal justice system. The panel's report synthesizes much of the recent research, explicates methodological issues that limit the validity of that research, and identifies future research possibilities. See id. a t xi-xii. 10. Whether variations in sentences are unjustified may depend on the assumptions made about the normative goals of sentencing. If these goals are the utilitarian ones of preventing crime, (deterrence, incapacitation, and rehabilitation) sentences are justified on the basis of predictions of future crime and rehabilitative potential. Individualized sentencing is appropriate in pursuit of utilitarian aims. The rehabilitative goal, for example, prescribes sentence variation based on offender characteristics. But if the purpose of sentencing is retributive, sentences are determined on the basis of the seriousness of the offender's criminal conduct, his personal culpability, and the harm done or risked. The retributive aim requires similar punishment for cases. See A. VONHIRSCH, supra note 2, a t 66-67. PSR reports that there currently is widespread disagreement over the goals of sentencing and that the decision makers-legislators, judges and parole officials-are rarely "purely utilitarian or purely retributive." See PANELREPORTsupra note 6, a t 4-5. Heinonline - - 36 Syracuse L. Rev. 719 1985-1986 720 Syracuse Law Review [Vol. 36:715 mitigated in systems where judges or parole boards are regulated by guidelines or where the parole function has been eliminated or greatly modified. It then offers some new data on the question of the reduction of sentencing disparity by parole boards acting without explicit decision making guidelines. A. Statutory Determinate Sentence Laws-Disparity Reduction in California At least seven states have recently attempted to regulate sentencing power by abolishing parole release and enacting statutory determinate sentence laws:ll Alaska, California, Colorado, Illinois, Indiana, New Mexico, and North Carolina.12 Two other states, Arizona and New Jersey, enacted such statutes while retaining parole release.lS Typically, the laws in these states reduce judicial discretion by dividing crimes into a few groups, and specifying a sentence of imprisonment-or range of sentences-for each category." Except for a class of serious offenders identified by statute, the standards govern the decision to fix the duration of the prison sen- 11. Determinate sentencing laws vary widely from state to state, but their common characteristics are that: (1) they structure discretion with explicit and detailed standards for duration of the prison sentence if the judge chooses imprisonment, and (2) they establish procedures for the early notification of prisoners of their expected release dates. See von Hirsch & Hanrahan, Determinate Penalty Systems in America: An Overview, 27 CRIME & DBLINQ. 289, 294 (1981) Under this definition of determinacy, various bodies would formulate the explicit standards: the legislature, a sentencing commission, or the parole board. See id. at 294. 12. See ALASKA STAT.§§ 12.55.125--12.55.175 (1984); CAL.PENALCODE§§ 1170 (West Supp. 1984); COLO. REV.STAT. 16-11-304,18-1-105(Supp. 1984); ILL. ANN. STAT.ch. 38, 1005-5-1, 1005-5-3, 1005-5-3.1, 1005-5-3.2, 1005-8-1, 1005-8-2 (Smith-Hurd 1982 & Supp. 1983-84);IND.CODEANN. 3 s 35-50-2-1 to 35-50-2-10 (Burns 1979 & Supp. 1984); N.M. STAT. ANN. $3 31-18-15, 31-21-10 (1978 & Supp. 1984); N.C. GEN.STAT.$8 15A-1340.1-7, 15A1370.1 to 15A-1380 (1983). See also von H i c h & Hanrahan, supra note 11, at 299. In some of these states, indeterminate sentences and parole release were retained for certain categories of cases. See, e.g., People v. Wasley, 133 Cal. App. 3d 352, 184 Cal. Rptr. 25 (1982) (noting that legislature chose to exempt mentally disturbed sex offenders from determinate sentencing). 13. See AHIZ.REV. STAT.ANN. $8 13-701b to 13-709, 13-411 to 13-412, 41-1604.06 (1978 and Supp. 1983-84); N.J. STAT.ANN.§$ 2C:43-6,2(=:44-1 (West 1982 & Supp. 1983); $8 30:4123.45--.69 (West 1981 & Supp. 1984-85); see also von Hirsch & Hanrahan, supra note 11, a t 299.28. 14. See von Hirsch & Hanrahan, supra note 11, a t 299. Heinonline - - 36 Syracuse L. Rev. 7 2 0 1985-1986 19851 Parole and Sentence Disparity 721 tence, not the decision whether to imprison.lS If, under determinate sentencing, the judge selects incarceration as a penalty, the presumptive sentence must be imposed unless a departure is justified by the circumstances of the case.16 For example, the California Uniform Determinate Sentencing Act provides that, upon committing offenders to prison, judges must choose one of three specified sentences as the "base term" for defendants convicted of a particular offense,'? such as 2,3, or 5 years for robbery;18 and 2, 4, or 6 years for first degree burglary.l9 The middle term is the presumptive sentence, the term to be imposed in the ordinary case.20The upper or lower term may be imposed only where aggravating or mitigating circumstances are found and set forth on the record.21 The law also provides for increasing the base term by a variety of increments, termed "enhancements," for prior imprisonment, use of a weapon, serious bodily injury, or large property loss, if alleged and proven by the pro~ecutor.2~ In California, like most states with legislative schemes, parole authorities supervise offenders released from ~onfinement,2~ but no longer make release decisions and therefore play no role in determining the actual period of imprisonment?' The term announced by the court represents the period the offender must actually serve, less a deduction for good-time credits e ~ n e d . 2In~ most states, a prisoner might earn up to one-third or one-half off the sentence for satisfactory behavior in pris0n.2~ The purposes of the new sentencing law, according to the California legislature, include both punishment and the elimination of 15. In California, for example, except in unusual cases, prison terms must be imposed on persons convicted of burglary of a habitation in the nighttime, or a felony burglary in the daytime. See C u . PENALCODE$ 462 (West Supp. 1984). 16. See infra text accompanying note 22. 17. Uniform Determinate Sentencing Act, 1976 Cal. Stat. ch. 1139, 273 (codified as amended at CAL.PEN& CODE8 1170 (West Supp. 1985)). 18. See C u . PENALCODE8 213 (West Supp. 1985). 19. See id. 3 461. 20. See id. 3 1170(b). 21. See id. 22. See id. 8 1170.l(fj. 23. See id. 8 3001 (West 1982). In California, the period of parole supervision may be as short as one year following the release on parole from state prison. See id. 24. See id.§ 3000. 25. See C u . PENALCODE3 3000(a) (West 1982). 26. In California, a prisoner's sentence may be reduced be one-third for good behavior. See id. 3 2931 (West 1982 & Supp. 1984). Heinonline - - 36 Syracuse L. Rev. 721 1985-1986 722 Syracuse Law Review [Vol. 36:715 di~parity?~ two goals that are compatible under a theory of imposTo a retributivist, those whose acts are ing deserved puni~hment.2~ Inequally blameworthy deserve equal amounts of puni~hment.4~ deed, the California act is entitled the Uniform Determinate Sentencing The statute provides for mechanisms intended to achieve sentence uniformity. A presumptive sentence and a "reasons requirement" restrain the judge from unexplained deviations from prescribed penalties. If a deviation is justified, the law still strives to avoid wide differences in the lengths of sentences for the same offense by allowing only a relatively narrow range between the upper or lower term, thereby limiting the exercise of discretion upward or d o w n ~ a r d The . ~ ~ legislation includes another unique feature: automatic administrative review for disparity. The Board of Prison Terms must review each prison sentence "to determine whether the sentence is disparate in comparison with the sentences imposed in similar cases."se 27. See id. 8 1170(a)l (West Supp. 1984). The California determinate sentencing statute provides: The Legislature finds and declares that the purpose of imprisonment for crime is punishment. This purpose is best served by terms proportionate to the seriousness of the offense with provision for uniformity in the sentences of offenders committing the same offense under simiiar circumstances. The Legislature further finds and declares that the elimination of disparity and the provision of uniformity of sentences can best be achieved by determinate sentences fixed by statute in proportion to the seriousness of the offense as determined by the Legislature to be imposed by the court with specified discretion. Id. 28. See supra note 10. 29. See id. 30. See Uniform Determinate Sentencing Act, 1976 Cal. Stat., ch. 1139,s 273, (codied as amended a t CAL.PENALCODE8 1170 (West SUDD. - - 1984)). 31. One's view of the width of ranges may depend on assumptions made about the normative goals of sentencing. A utilitarian, who in particular cases might use sentencing to achieve incapacitative, deterrent, or rehabilitative objective, would favor broad ranges that permit lengthy confinement for some offenders but not others. Norval Morris argues against sentencing offenders up to the maximum permitted by the concept of desert: "To say that a punishment is deserved. ..is not to say that it ought to be imposed. The concept of desert. ..is one of a retributive maximum; a license to punish the criminal up to that point but by no means an obligation to do so." N. MORRIS,THEFUTUREOF IMPRISONM~NT 75 (1974). One who espouses retribution, the express goal of sentencing in California, see supra note 27, would favor narrow ranges, believing that they achieve equality in sentencing and reduce disparities. See PANELREPORT,supra note 6, a t 164,170. 32. CAL.PENALCODE8 1170(f)(l) (West Supp. 1985). PSR has said that administrative review may be more advantageous than appellate review because it does not depend on the Heinonline - - 36 Syracuse L. Rev. 722 1985-1986 19851 Parole and Sentence Disparity 723 The Panel on Sentencing Research reviewed a t least seven major studies that examined the impact of the determinate sentence . ~ ~ of these studies addressed the question law in C a l i f ~ r n i a Three of whether the reform has resulted in greater uniformity in senten~ing.~' The approach common to each was to compare the variability in prison terms imposed by judges under the new law with the variability in periods of actual imprisonment as determined by parole officials under the old law. Controlling for offense types, and using the standard deviation as an index of variation, Brewer, Becker, and Holt reported reductions from twenty to fifty percent in the spread of sentences for five of the seven offense types for men.s6 Measuring dispersion by initiative of the parties. See PANELREPORT,supra note 6, a t 181. The Board of Prison Terms describes its review for disparity as follows: MANAGEMENT INFORMATION SECTION. CALIFORNIA BOARDOF PRISONTERMS,SENTENCING PRACTICES UNDERTHE DETERMINATE SENTENCELAW3-5 (1984). The annual reports of the Board of Prison Terms, shows that the Board has found relatively few sentences to be disparate: Prison Disparate Disparately Disparately Year Sentences Sentences High Low - See, e.g., id. The annual report also shows that courts resentenced defendants in less than half the cases referred to them for reconsideration. See, e.g., id. PSR has said that the California review system has not yet had significant effect. See P m m REPORT,supra note 6, a t 182. For a discussion of the administrative review for disparity, see Fenili, California's Disparate Sentence Review Process: Conceptual and Practical Issues, in SENTENCINGREFORM:EXPERIMENTS IN REDUCING DISPARITY 131 (M. Forst ed. 1982). 33. See PANELREPORT,supra note 6, a t 189. 34. See D. BREWER,GE. BECKETT & N. HOLT,DETERMINATE SENTENCING IN CALIFORNU: THEFIRSTYEAR'SEXPERIENCE (1980);J. CASPER, D. BREREMN,& D. NEAL,THE IMPLEMENTATION OF THE CALIFORNIA DETERMINATE SENTENCING LAW(1982)[hereinafter cited as CASPER); R. Ku,IV AMERICAN PRISONS AND JAILS:CASESTUDIES OF NEWLEGISLATION GOVERNING SENTENCING AND RELEASE (1980)These studies and others were reviewed in Cohen & Tonry, Sentencing Reforms and Their Impacts, in 11 RESEARCH ON SENTENCING: THE SEARCHFOR REFORM 305 (A. Blumstein, J. Cohen, S. Martin & M. Tonry eds. 1983) This volume contains several papers which were commissioned by PSR to investigate research in particular areas. While those papers represent the views of the individual authors and not necessarily those of the panel, they were published because PSR believed they made a valuable contribution to the literature on sentencing research. See Blumstein, Preface to PANELREPORT, supra note 6, a t xii. I t may be noted that J. Cohen and M. Tonry served as consultants to the PSR. See id. a t xiii. 35. See Cohen & Tonry, supra note 34, a t 409. Standard deviations increased for rape and assault with a deadly weapon, but decreased for robbery, forgery and checks, first-de- Heinonline - - 36 Syracuse L. Rev. 723 1985-1986 724 Syracuse Law Review [Vol. 36:715 the breadth of a mid-range around the median, Casper, Brereton, and Neal (Casper) observed less variability in time served for bur~~ researcher, Richard Ku, glary and robbery o f f e n ~ e s .Another made similar findings for burglary, robbery, and all felony offenses and concluded that "the intent to narrow variation in prison terms was largely met."s8 These findings do not, however, show a reduction in sentence disparity in the sense that the term is used in this article. As defined here, disparity means that "equally situated offenders" are "treated unequally"89 or, a t the decision point of interest here, that they receive unequal periods of imprisonment. For reasons discussed below, "equally situated offenders" are viewed in this article as persons with similar criminal histories and similarly serious offense behavior.'O The California studies used only offense types, such as burglary or robbery, as a measure of offense seriousness, legal classifications that subsume a vast heterogeneity of offender conduct.41 Without more precise indicators of offense seriousness, gree burglary, second-degree murder, and receiving stolen property. See Cohen & Tonry, supra note 34,a t 453.The five offense typeswith decreased standard deviations experienced increases in means, thus providing a more reliable indication of decreased variability. See id. at 409. 36. See CASPER,supra note 34, a t 254. The following table derived from CASPER, supra note 34, a t 154, shows narrower mid-ranges under the new law: Months of Confinement 1975 1978-79 Releases* Releases Robbery Median 80% Mid-range** 43 29-66 32 24-56 Burglary Median 33 80% mid-range 21-61 *Actual terms imposed less one-third for maximum good time credit. **The 80% mid-range is the range of prison sentences for 405 of the sentences above the median and 40% below. These decreases in mid-ranges must be viewed with caution because they are associated with decreased in medians. "Any decrease in the mean or median increasingly contains the possible distribution of prison terms below that 'midpoint,' thus limiting the range of potential variation." Cohen & Tonry, supra note 34,at 409. 37. See R. Ku, supra note 34, at 67. Ku observed reductions under the new law in the 50 and 75 96 mid-ranges for burglary, robbery, and all felony offenses combined. Id. 38. See id. a t 71. 39. See infra text at notes 135-36. 40. See infra text at notes 139-40. 41. See Forst, Sentencing Disparity: An Overview of Research and Issues, in SEHFBNC- Heinonline - - 36 Syracuse L. Rev. 724 1985-1986 19851 Parole and Sentence Disparity 725 such as presence of victim, harm to victim, or use of weapon, and without any controls for the prior record available, the offenders in the California studies cannot be viewed as "equally situated," and any conclusions regarding sentence disparity, as that term is deh e d here, are suspect. The Casper study also found that after the enactment of the new law, the range of actual sentences within conviction classes remained quite large.'= Despite the relatively narrow range across the base terms for each offense, sentences imposed for robbery ranged from one to twenty years and from one to nine years for burglary/8 These wide ranges were explained by the discretionary use of enhancements and consecutive sentences." If acts of enhancement are charged and proven, the judge may impose or stay an additional term on the enhan~ement.'~If the offender is convicted of multiple counts of the same offense, the judge may impose concurrent or consecutive sentence^.'^ Even though the new law appears to reduce judicial latitude by requiring the selection of one of three base terms, such choices by judges can significantly expand the range of penalties faced by offender^?^ Thus Casper concluded that "in all but the simplest single-count, no-enhancement cases (in which prison is often not seriously considered anyway), the judge and prosecutor possess a great deal of discretion over the term to be served by the defendant."4s As noted, the potential for disparity exists in determinate penalty systems not only because of judicial discretion, but also because of the pivotal role of the prosecutor. Since the charge a t conviction determines the sentence, the prosecutor has a large measure of control over sentence outcomes through charging and plea bargaining decisions.'@ This discretion lends itself to inequali- REFORM: EXPERIMENTS IN REDUCING DISPARITY 9, 29 (M. Font ed. 1982). 42. See CASPER, supra note 34, at 155. 43. See id. Casper noted that "the data indicate that a good deal of sentence disparity among those convicted of similar crimes still remians. The DSL, despite its expressed concern for increasing uniformity, has not by any means produced a narrow range of relatively equal sentences." Id. 44. See id. 45. See CAL.PENAL CODE8 1170.l(h) (West Supp. 1984). 46. See, e.g., People v. Wiener, 91 Cal. App. 3d 238,154 Cal. Rptr. 110 (1979). People v. Seaman, 86 Cal. App. 3d 815, 150 Cal. Rptr. 430 (1978). 47. See CASPER, supra note 34, at 23. 48. Id. at 23. 49. See Zimring, Making Punishment Fit the Crime: A Consumer's Guide to Sentenc- ING Heinonline - - 36 Syracuse L. Rev. 725 1985-1986 726 Syracuse Law Review [Vol. 36:715 ties because of variations among prosecutors in charging practices and disagreements over issues of sentencing Franklin Zimring argues that under presumptive sentencing schemes it is inevitable that disparities will persist between different areas and different prosecutors, and between sentences following a guilty plea and those following a jury trialP1 Albert Alschuler predicts that, unless plea bargaining is abolished or the charging decisions of prosecutors regulated, determinate sentencing standards will be nullified through the exercise of prosecutorial discretiona6= The Panel on Sentencing Research recognized that prosecutors might subvert determinate sentence penalty standards through charge bargains,6s and urged policy makers to devise . ~ apmethods to counterbalance the effects of such p r a ~ t i c e s One and proach, recommended by the American Bar Asso~iation~~ may be appellate review found useful in some European system~,6~ of sentences for disparity. While such review has not proven to be ing Reform, in HAWINGSCENTERREP. (Dec. 1976). 50. See Alschuler, Sentencing Reform and Prosecutorial Power: A Critique of Recent Proposals for "Fixed" and "Presumptive" Sentencing 126 U. P a L.REV.550,563-76(1978). 51. See Zimring, supra note 49, a t 12. 52. See Alchuler, supra note 50, a t 563-76. The regularization of plea bargaining through the use of administrative rulemaking procedures and the formulation of internal guidelines has long been demanded. See, e.g., K. DAVXS, DISCRETIONARY JUSTICE: A PRELIMINARY INQUIRY (1969); Vorenberg, Narrowing the Discretion of Criminal Justice Oflichls, 1976 DUKEL. J. 651,680-82;Vorenberg, Decent Restraint of Prosecutorial Power, 94 HAW. L. REV. 1521, 1543 (1981). But the diiculties of imposing the task of self-regulation on prosecutors are substantial. Even if they welcomed the dilution of their present powers, it is hard to imagine that prosecutors could draft useful guidelines that took into account the myriad factors relevant to sentencing decisions. See Alschuler, supra note 50, a t 575-76. Almost " any standards one can imagine would refer to the strength of the evidence or other factors so vague as to impose no effective constraint upon the prosecutor's sentencing power." Schulhofer, Due Process of Sentencing, 128 U. P a L.REV.733,756 (1980).Professor Schulhofer concluded that the long-term solution to disparity in sentencing may lie in the development of comprehensive standards over prosecutorial decisionmaking, but that the immediate obstacles to their formulation and implementation make it preferable to concentrate on less drastic reforms. See id. a t 767. 53. PANELREPORT,supra note 6, a t 26. PSR posited that "[ulnder sentencing guidelines and statutory determinate sentencing laws with presumptive authority and under mandatory sentencing laws, prosecutors and defense attorneys may be able to circumvent applicable standards through charge bargains." Id. 54. See id. a t 26. 55. See AMERICAN BARASS'N PROJECTON MINIMUM STANDARDS FOR CRIMINAL JUSTICE, STANDARDS RELATING TO APPELLATE R E ~ OP W SENTENCES (1976). 56. See Andenaes, The Choice of Sanction: A Scandanavian Persepctive, in REFORM AND PUNISHMENT 3 (M. Tonry & F. Zimring eds. 1983); Weigand, Sentencing in West Germany, in REFORMAND PUNISHMENT 21 (M.Tonry & F. Z i i i n g eds. 1983). Heinonline - - 36 Syracuse L. Rev. 726 1985-1986 19851 Parole and Sentence Disparity 727 effective in the United States under indeterminate senten~ing:~its prospects for success in this country may be greater under the more definite standards of determinate sentence laws.6s Even under these new systems, however, the prevalence of plea bargaining raises questions concerning the efficacy of appellate review as a mechanism for reducing d i ~ p a r i t y . ~ ~ Various commentators have proposed another line of attack on the disparity resulting from prosecutorial subversion of penalty standards in a regime of determinate sentencing, namely, legislative specification of a discount to follow entry of a plea of The proposals vary as to the form such discounts might take:* but 57. See PANELREPORT,supra note 6, a t 57. Systems of indeterminancy allowed vast discretion to sentencing judges and provided no meaningful standards by which appellate judges could measure the appropriateness of sentences. See id. a t 180. Moreover, parole boards, rather than appellate courts, were seen as providing sentence review. Id. a t 57. Indeed, in view of parole release, board determinations of sentence length were more important than sentencing judges' decisions. See id. See also Zeisel & Diamond, Search for Sentencing Equity: Sentence Review in Massachusetts and Connecticut, 1977 AB. FOUND. RESEARCH J. 881; Note, Appellate Review of Primary Sentencing Decisions: A Connecticut Case Study, 69 YALEL. J. 1453 (1960). 58. See PANELREPORT,supra note 6, a t 180-81. 59. Professors Coffee and Tonry question whether a just and effective system of appellate review can exist in a world of plea bargaining. They pose the following questions: (1) will either party have an incentive to appeal a negotiated sentence, (2)has the defendant who pled guilty in consideration of the dismissal of other charges been denied the benefit of his bargain when the prosecutor appeals the sentence, (3)should the plea agreement, to avoid deceiving the defendant, contain an express reservation by the prosecutor of a right to appeal the sentence, (4) is there a danger that the defendant's willingness to appeal may be chilled by the threat of a cross-appeal by the prosecutor, (5) or chilled by the prospect of resentencing by the original sentencing judge upon remand by an appellate court, (6) can procedures be devised to guard against these dangers, (7)should the prosecutor have the right to appeal, (8)does the double jeopardy clause bar an appeal by the prosecutor, (9)if either prosecutor of defendant lack the incentive to appeal should a thud party have that right, (10) should the victim, (11)who are the victims, (12)is a parole board precluded from looking behind the conviction offense to the "real" offense when the prosecutor accepts a plea to a lesser charge and agrees not to appeal the sentence, (13)wouldn't sentence review be best handled on a de facto basis by the parole board alone. See Coffee & Tonry, Hard Choices: Critical Trade-offs in the Implementation of Sentencing Reform through Guidelines, in REFORM AND PUNISHMENT 155 (M. Tomy & F. Zimring eds. 1983). 60. See, e.g., D. GOTPPREDSON, L. WILKINS& P. HOFFMAN, GUIDELINES FOR PAROLE AND SENTENCING 125-26 (1978) [hereinafter cited as GOTTPREDSON, WILKINS& HOFFMAN]; Alschuler, supra note 50, a t 575-76;Schulhofer, supra note 52, a t 791-94;Note, Restructuring the Plea Bargain, 82 YALEL.J. 286, 301-02(1972). 61. One proposal is for a fixed discount applicable across-the-board in all guilty-plea cases. See YALENote, supra note 60, a t 301-02.Another is that every cell in a guidelines matrix contain one sentence for offenders convicted a t trial and another for those who plead guilty. See GOTTPREDSON, WILKINS& HOFFMAN, supra note 60, a t 125-26.Still another pro- Heinonline - - 36 Syracuse L. Rev. 727 1985-1986 Syracuse Law Review all provide for an explicit differential between guilty-plea and trial defendants. Arguably disparity would be reduced because the concessions granted defendants who plead guilty would be unaffected by extraneous factors that might operate where prosecutors find covert ways to reward guilty pleas.s2 Legislative specification of guilty-plea discounts might help to achieve equity in sentencing and undoubtedly would increase the candor of our system.6gConstitutional uncertaintys4 aside, however, sentence differentials present policy and operational problems so difficult as to render these proposals of doubtful feasibility.s6 posal is that the discount be adjustable and geared to the burden created for the state by the defendant. See Coffey & Tonry, supra note 59, a t 188-90. 62. Professor Alschuler suggests that legislatively prescribed discounts would reduce the unevenness in sentencing under our present system of unregulated plea bargaining: Under such a statute, coupled, of course, with the elimination of plea bargaining by prosecutors, the "break" that follows the entry of a guilty plea would not depend upon the prosecutor's whim. The extent of this "break" would not be affected by a prosecutor's feelings of friendship for particular defense attorneys, by his desire to go home early on an especially busy day, by his apparent inability to establish a defendant's guilt a t trial, by his (or the trial judge's) unusually vindictive attitude toward a defendant's exercise of the right to trial, by the race, wealth or bail status of the defendant, by a defense attorney's success in threatening the court's or the prosecutor's time with dilatory motions, by the publicity that a case has generated, or by any number of other factors-irrelevant to the goals of the criminal process-that commonly influence plea bargaining today. Alschuler, supra note 50, a t 575. 63. See Coffey & Tonry, supra note 50, a t 575. 64. In Corbitt v. New Jersey, 439 U.S. 212 (19781, the Court, by a five-member majority, upheld a statutory scheme that imposed substantially more severe penalties on trial defendants than on those who plead no contest. Professor Schulhofer, analyzed this decision and concluded that a properly developed system of guilty-plea discounts would probably survive constitutional challenge, a t least if there were overlapping between separate sentencing ranges for contested and uncontested cases. See Schulhofer, supra note 52, a t 780-83. 65. Different views may be taken of a system of guilty-plea concessions: A sentencing . . system that awards an explicit guilty-plea discount can be seen in three very different lights. First, it is an incentive to plead guilty and so disfavors those defendants who go to trial. Second, it is a limitation on the amount of credit that can be given for such a plea; that is, it is as much a ceiling on the permissible reward as a disincentive to profess innocence. Third, it is a form of consumer protection for offenders. Offenders sometimes plead guilty in return for a valueless consideration: . . . the difference between the statutory ceiling for most offenses and for their lesser included offenses may be irrelevant, because the likely sentence for either crime is below the lower ceiling; similarly, when collateral counts are dropped, they may not have much impact on the sentence (as when 99 counts of mail fraud, each based on a different use of the mails, are reduced to 15). Underlying such one-sided negotiations may be either a conflict of interest between defense counsel and the client (so that the attorney does not explain to the client the limited value of the benefit) or simply intransigent self-deception on the part of a . Heinonline - - 36 Syracuse L. Rev. 728 1985-1986 19851 Parole and Sentence Disparity 729 B. Sentencing Guidelines and Sentencing Commissions-Minnesota as a n Example The idea that judges might use guidelines in sentencing is an outgrowth of the concept developed by the Federal Parole Board in the 1970's' that parole officials should follow set standards in making release decision^.^^ A valuable function of either parole or sentencing guidelines is that their writing requires the formulation of explicit policy by articulation of the most important factors considered in sentencing and how they are ~ e i g h t e d . ~ ? Three states, Minnesota, Pennsylvania, and Washington, have created sentencing commissions and empowered them to develop detailed sentencing criteria that judges are required to consider in ~ e n t e n c i n gMinnesota .~~ abolished parole release and a highly indeterminate sentencing system.6s Its new guidelines specify the exact clearly guilty defendant, which requires that he be able to tell himself that he got something in return for his admission of guilt. In contrast to such illusory consideration, a specific sentence discount. ..would ensure the defendant that he gets something in return. Coffee & Tonry, supra note 59, a t 187. These writers are also troubled by policy and operational problems associated with overtly rewarding guilty pleas. They discuss the following questions: (1) if discounts survive constitutional challenge, can they be structured to encourage guilty pleas from the guilty without coercing them from the innocent, (2) if the decision is made to give discounts, how should they be calculated, (3) should a discount be adjustable and geared to the difficulty and cost of a trial to the state, (4) should a discount be given following conviction a t a bench trial on the theory that a jury trial would have placed a greater burden on the state, (5) should a discount be adjusted to the time when the defendant plead guilty, so that he who pleads guilty a t the outset receives a larger discount than he who waits until the day of trial, (6) should a discount be inversely proportional to the strength of the state's cases- he weaker the state's evidence, the greater the discount-or would that coerce guilty pleas from those who have the best chance of acquittal, (7) should a discount be directly proportional to the strength of the defendant's case--the stronger the defense, the greater the discount-so that the least pressure to plead guilty is put on those who have the best chance of acquittal, (8) will discounts be distorted by "double counting" when the determined defense lawyer insists upon a charge reduction in addition to a discount, (9) wouldn't "double counting" be prevented a t the sentencing stage if discounts were combined with "real-offense" sentencing. See id. a t 185-90. 66. See generally GOTTFREDSON, WILKINS& HOFFMAN, supra note 60. For a discussion of conceptual and methodological problems associated with the construction of sentencing guidelines, see Sparks, The Construction of Sentencing Guidelines: A Methodological Critique, in I1 RESEARCH ON SENTENCING: THE SEARCH FOR REFORM 194 (A. Blumstein, J. Cohen, S. Martin & M. Tonry eds.. 1983). 67. See generally GOTTPREDSON, WILKINS& HOFFMAN, supra note 60, a t 13-39. 68. See MINN.STAT.ANN. 8 244.09 (West Supp. 1985); Pa STAT.ANN. tit. 42, § 2151 (Purdon 1981); WASH.REV.CODEANN. § 9.94A.040 (West Supp. 1985). 69. See MINN.STAT. ANN.3 244 app. (West Supp. 1985) (complete text of sentencing Heinonline - - 36 Syracuse L. Rev. 729 1985-1986 730 Syracuse Law Review [vol. 36:715 time, less a one-third deduction for good behavior, to be served by the offender.?O Pennsylvania adopted sentencing guidelines while retaining an indeterminate system.?' Its guidelines specify a minimum term to be served before parole eligibility and a maximum term beyond which the prisoner cannot be held?2 The actual period of imprisonment is determined by the parole board through Washington abolished parole release7' and esits release deci~ion.?~ tablished determinate guideline sentence^.?^ In each state, judges must impose the sentence recommended by the guidelines or provide an explanation of why some other sentence was imp~sed.?~ Because the guidelines in these states are legislatively mandated and have the force of statutes,'? they are referred to as "presumptive" guideline^.?^ The guidelines in the three states are also referred to as "prescriptive" guidelines because they are designed not to codify existing sentencing practices, but to depart from the past and develop new sentencing p0licies.7~In Minnesota, for example, the sentencing commission made the explicit policy choice that greater emphasis ought to be placed on achieving sentence uniformity guidelines and commentary) [hereinafter cited as MINNESOTA GUIDELINES]. 70. See id. 71. See 204 PA. ADMIN.CODE3 303. (Shepard's 1984) (complete text of sentencing guidelines). 72. See id. 73. See id. 74. See WASH.REV.CODEANN.3 9.94A130, 9.94A310-9.94A450 (West Supp. 1984-85) (complete text of sentencing guidelines). 75. See id. 3 9.94A120. 76. See MINN.STAT.ANN. 3 244.10 (West Supp. 1985); P a STAT.ANN.tit. 42, 3 2155 (Purdon Supp. 1984-1985); WASH.REV.CODEANN. 3 9.94A120 (3) (West Supp. 1985). 77. See citations a t supra note 68. 78. Guidelines are classified as either presumptive or voluntary, depending on whether they are legislatively mandated to have the force of law or whether they are merely hortatory and create no rights in the parties. See PANELREPORT,supra note 6, a t 135-39. Under presumptive guidelines, such as those in Minnesota, Pennsylvania, and Washington, both prosecutors and defendants are entitled to appellate review of the sdliciency of an explanation for a departure from the sentence recommended by the guidelines. See id. a t 136. 79. See id. a t 135. Guidelines are also classified as either prescriptive or descriptive, depending on whether they are intended to develop new sentencing policies (prescriptive) or whether they are designed to codify past policies and make them explicit (descriptive). See id. a t 135-39. The two are not always distinguishable because most guidelines prescribe policy, and most are empirically based on past sentencing practices. See id. at 135. The guidelines in Minnesota, Pennsylvania, and Washington are both presumptive and prescriptive. Guidelines that are both voluntary and descriptive have been adopted a t the local level in a t least 11 states. See id. a t 138-39. Heinonline - - 36 Syracuse L. Rev. 730 1985-1986 19851 Parole and Sentence Disparity 731 within categories defined by the guideline^.^^ In addition, the commission said that sentences should be neutral with respect to race, sex, and socioeconomic status of offenders.*' Sentencing guidelines, like parole guidelines, attempt to reduce sentence disparities by assuring that similar offenders are dealt with in similar ways in similar situation^.^^ The guidelines matrix is the primary mechanism for achieving sentence uniformity in any guidelines context. The matrices of the states with parole or sentencing guidelines have a similar format and their operation can be illustrated by the two-dimensional chart constructed by the Minnesota sentencing commission. 80. See MINNESOTA GUIDELINES, supra note 69, at 251. 81. See id. 82. See G ~ R E D S O WILKINS N, & HOFFMAN, supra note 60, at 13. Heinonline - - 36 Syracuse L. Rev. 731 1985-1986 [Vol. 36:715 Syracuse Law Review FIGURE I Minnesota Sentencii Guidekm Grid Presumptive Sentence Length in Montbs I t a l i c i d nunbers within the grid denote the range within which a judge may sentence without the sentence b e i i deaned a departure. Offenders with nonirnprisonment felony sentences are subject to jail tirne m r d i n g to law. CRIMINAL HISTORY SCORE SEVERITY LEVELS OF 0 1 2 CONVICTION OFFENSE 4 3 5 6armm Unauthorized Use of Motor Vehicle Possession of Marijuana I 12* 12* 12* 13 15 17 19 18-20 Theft Related Crimes ($260%2600) Aggravated Forgery I1 12* 12* 15 13 17 19 ~$260-$2MX)) 21 20-22 1 Theft Cn'mes ($250-$2500) I11 12* 13 15 17 Nonresidential Burglay IV Th& Crimes (over $2500) 12* 15 18 21 22 26 19 18-20 21-23 24-26 25 , V 18 23 27 30 29-3l Assault, 2nd Degree VI 21 26 30 Aggnnxzted Robbery VII 32 41 24 23-26 30-34 38-44 Crimiml S e a l Conduct, VIII 54 1st Degree 43 Assault, 1st Degree 41-45 50-58 Murder, 3rd Degree Murder, 2nd Degree (fklony murder) Murder,2nd Deprce (with intent) IX 105 119 65 a-70 127 41 -- 24-26 30-34 37-46 1 Residential Burglary Simple Robbery 32 38 46 54 36-40 43-49 60-68 34 44 54 65 33-35 42-46 50-68 6Y)-70 49 45-63 65 @-70 76 95 81 97 76-87 90-104 113 132 71-8l 80-101 106-120 124-140 149 176 205 230 102-108 116-122 124-130 143-155 la-184 196-216 218-242 X 120 140 162 203 243 2&1 324 116-124 133-147 153-171 192-214281-255270-208 SY)O-S39 1st Degre M d a is excluded from the guidelines by law and continues to have a mandatary life senteaee. *one yea and one day (b Eff. .8/1/81; 11/1/85) Heinonline - - 36 Syracuse L. Rev. 732 1985-1986 19851 Parole and Sentence Disparity 733 On the grid's vertical axis is a ranking of conviction offenses according to their seriousness. On the horizontal axis is a criminal history score for rating the extent of the offender's prior record. Within the grid are cells that show the recommended sentences for offenders whose scores intersect in any row or column. The assumption underlying the grid as a device for reducing disparity is that it classifies cases into subgroups on the basis of key determinants of sentences-criminal history and seriousness of offenseS3--so that offenders falling within the same cell may be seen as fairly homogenous and deserving of similar sentences.s4 Sentence disparity also exists when one similarly situated offender is incarcerated and another is not. The Minnesota sentencing matrix addresses this issue by specifying the length of a sentence for each guidelines category and, uniquely, whether that sentence should be served in prison. Cells above the bold line on the chart-called the dispositional, or "in/out7' line-prescribe the length of a nonincarcerative sentence. Cells below this line prescribe the length of a prison sentence.86In specifying the variation in sentences across the cells on the grid, particularly in locating the "inlout" line on the matrix, the commission was guided by its adoption of retribution as the primary goal of s e n t e n ~ i n g . ~ ~ 83. See MINNESOTA GUIDELINES, supra note 69, a t 251; see akro infra text accompanying note 138. 84. See Cohen & Tonry, supra note 34, a t 412. 85. Despite the Minnesota commission's policy objective of reducing disparities, use of the guidelines may result in dramatically different sentences for similarly situated offenders whose cases fall in adjacent cells on the grid, but who are on opposite side of the "inlout" line. See PANELREPORT,supra note 6, a t 165. PSR cited the example of two offenders convicted of second-degree assault, who have criminal history scores of 2 and 3 respectively. The difference between the offenders could be but a single felony conviction which adds one point to a criminal history score. Nevertheless, the guidelines prescrible a nonincarcerative sentence for one offender, but a 34-month prison term for the other. The irony would be all the greater if the offender with the lower criminal history score had been initially charged with first-degree assault but had been the beneficiary of a charge reduction in exchange for a guilty plea. See id. 86. The commission explicitly incorporated the retributive principle of proportionality into the guidelines' statement of purposes: . . that the severity of sanctions increase in direct proportion to increases in the severity of criminal offenses and the severity of criminal histories of convicted felons." MINNESOTA GUIDELINES, supra note 69, a t 251. In drawing the "in/out" line, the commission gave offense severity high emphasis and criminal history very low emphasis, reflecting a "just deserts" dispositional policy. See id. a t 252-54. Previously, the philosphical basis of sentencing in Minnesota was highly utilitarian. The Minnesota Corrections Board, for example, had incorporated the diverse goals of deterrence, incapacitation, and rehabilitation in its parole decision making guidelines. See MINNESOTA ". Heinonline - - 36 Syracuse L. Rev. 733 1985-1986 734 Syracuse Law Review [Vol, 36:715 The discretion that judges retain under a system of guidelines is restricted in that they may depart from the prescribed sentence only if they state in writing reasons sufficient to justify a deviation. In .Minnesota, departures are subject to a relatively high standard-that the individual case involves not merely "substantial circumstances," but "substantial and compelling circ~mstances."~~ The adequacy of the judge's reasons for a departure are subject to review by the state supreme court upon the request of either the prosecutor or the defendant.88 The cells that call for incarceration on the Minnesota chart specify a range within which a prison sentence can vary to reflect The judge may sentence aggravating or mitigating circ~mstances.~~ anpwhere within the range without providing an explanation. This range varies by only plus or minus seven to eight percent from the guideline sentence, which for most cases represents a range of approximately two to four months?O The narrowness of this range reflects an intent by the commission to decrease variability in sentence lengths for similar cases?' If judges generally sentence within this range and depart from it only infrequently, it may be expected that the guidelines will reduce sentence disparity?%But if judges SENTENCING GUIDELINES COMM'N., PRELIMINARY REPORTON THE DEVELOPMENT AND IMPACT OF THE MINNESOTA SENTENCING GUIDELINES 9 (1982) (discussing D. PARENT,MINNESOTA CORRECTIONS BOARD, FINALREPORTON PAROLE DECISION MAKINGPROJECT(1977)). [hereinatter cited as MINNESOTA REPORT]The shift in goals was a deliberate policy choice by the commission. It believed that offenders should be punished for their offenses, that sentencing was an ineffective mechanism for controlling crime, and that the simultaneous pursuit of multiple goals maximized discretion and produced excessive disparity. Although the commission did not entirely forsake utilitarian goals, i t deemphasized them substantially. See MINNESOTA REPORT, supra at 9-12. GUIDELINES, supra, note 69, at 251; see also MINNESOTA REPORT, 87. See MINNESOTA supra note 86, a t 13 (stating that the stronger standard for departures was adopted to promote the goal of uniformity in sentencing). 88. See MINN.STAT.ANN. 5 244.11 (West Supp. 1985) (providing that review may be had to determine whether sentences are "unjustifiably disparate"). The commission also eatablished an internal monitoring system to encourage judges to follow the guidelines; but the procedure consists only of receiving departure reports from judges whose sentences deviate from the guidelines and does not appear to constitute any credible control upon decisionmaking. See MINNESOTA MORT, supra note 86, a t 13. GUIDELINES, supra note 69, at 252. 89. See MINNESOTA REPORT,supra note 86, a t 13. The statute establishing the commis90. See MINNESOTA sion authorized a range in sentence length up to 30 % for each guidelines classification. See MINN.STAT.ANN.5 244.09 (West Supp. 1985). REPORT,supra note 86, a t 13. 91. See MINNESOTA supra note 6, a t 164. 92. See PANELREPORT, Heinonline - - 36 Syracuse L. Rev. 734 1985-1986 19851 Parole and Sentence Disparity 735 find the range too narrow and often seek to justify a sentence not within the guidelines, it might also be expected that the narrowness of the range will increase d i ~ p a r i t y . ~ ~ Thus far, there have been no independent studies of the effect of the sentencing guidelines on sentence disparity in Minnesota. The commission's own analysis of sentencing practices before and after the guidelines, however, showed a substantial reduction of disparity in "in/out7' decisions under the guidelines. To measure uniformity in the imprisonment of offenders (dispositional as opposed to durational uniformity), the commission compared actual sentencing decisions made by judges in one year under the guidelines with an estimate of sentencing decisions made by judges in one year under the former indeterminate system. Assigning preguidelines offenders to categories within the sentencing grid, the commission calculated a measure of variance or uniformity of imprisonment for each cell and ultimately for the entire grid.94 It then calculated a comparable measure from actual sentencing practices during a post-guidelines year. Comparing these measures, the commission found a fifty-two percent overall increase in the uniformity of commitment decisions under the guidelines, with increases in uniformity for males, females, and every racial group except a residual minority Unfortunately, the potential for disparity resulting from manipulation of penalty standards under a sentencing guidelines sys- 93. See id. 94. See MINNESOTA REPORT,supra note 86, AT 21-22. The commission measured dispositional uniformity by first calculating a measure of variance for each cell in the grid. Cell variance is computed by multyplying p by (1-p). To illustrate, let p = probability of imprisonment in a particular guidelines category (e.g., .20 or 20 96 go to prison): if p = 0, no one in the category goes to prison (complete uniformity) if p = 1, everyone in the category goes to prison (complete uniformity) if p = .5, half of the offenders in the category go to prison (there is less uniformity than under any other circumstance) The commission then calculated a measure of variance for the entire grid by multiplying each cell variance by the number of cases in that cell, summing the weighted variances, and dividing by the total number of cases in the grid. The grid variance, like the cell variance, ranges from 0, representing complete uniformity, to .25, representing the least uniformity possible. See id. 95. A comparison of grid variances shows substantial increases in dispositional uniformity under the guidelines: Heinonline - - 36 Syracuse L. Rev. 735 1985-1986 736 Syracuse Law Review [vol. 36:715 tem is the same as under a statutory system like Calif~rnia's?~ According to the Minnesota commission, "[tlthrough initial charging and charge reduction practices, a prosecutor can determine the presumptive sentence . . . and thereby subvert the intent of the guideline^."^^ In support, it cited disuniformity in patterns of charge reduction in aggravated robbery The guidelines presume a prison sentence for aggravated robbery offenders, but a nonprison sentence for simple robbery offenders with less serious prior records. The commission reported that after the guidelines prosecutors reduced charges from aggravated to simple robbery-thereby moving the offense to a presumptive nonprison sentence-disproportionately more for defendants with less serious records than defendants with more extensive records.9s In doing so, prosecutors made case-processing decisions on the basis of an offender's criminal history, a factor explicitly rejected by guidelines policy. Reporting that this pattern repeated itself in assault and 1978 preguidelines 1980-81 postguidelies Percent Change in Uniformity Overall .0499 .lo41 52% White .loo0 .0408 59A Black .0779 .0674 13% Native American Other Minority (mostly Hispanics) Male Female MINNESOTA REPORT,supra note 86, a t 22,table 1. 96. See generally Coffee & Tonry, supra note 59. 97. MINNESOTA REPORT,supra note 86, a t 62. 98. Id. a t 62-63. 99. See id.; Cohen & Tonry, supra note 34, a t 426. Under the guidelines there were fewer charge reductions for aggravated robbery defendants with higher criminal history scores than for defendants with lower criminal history scores, as follows: Percent of Defendants Convicted of Aggravated Robbery Among Defendants Charged with Aggravated Robbery Criminal history scores 0 1 2 3 4 59 75 65 54 58 1978,preguidelines 1980-81,postguidelines 49 60 66 70 70 See id. a t 426 table 7-42. Heinonline - - 36 Syracuse L. Rev. 736 1985-1986 19851 Parole and Sentence Disparity 737 criminal sexual conduct cases, the commission warned, "the potential for undermining legislation and commission policy clearly exists and must be monit~red."'~~ Disparities caused by the adaptive responses of prosecutors under presumptive sentencing schemes might also be reduced through real-offense sentencing-the practice whereby judges impose sentences that are based on actual offense behavior rather than on the charge at convi~tion.~~' Real-offense sentencing may be seen as a means of mitigating variations among prosecutors in charging practices, and of lessening the differential between sentences following a guilty plea and those following a jury trial.'02 Because of this potential, use of the practice in a guidelines context has been endorsed by the United States Parole Cornmis~ion,'~~ the parole boards of several states,'04 and the National Conference of Commissioners on Uniform State Laws.lo6 Unless its employment is admitted, however, real-offense sentencing tends not to reveal itself in systems where discretion is broad and unstructured, and some suspect that it has been practiced much more widely than recent acknowledgements would indicate.loe While real-offense sentencing may be viewed as a means to 100. See MINNESOTA REPORT,supra note 86, a t 63. 101. For discussions of real-offense sentencing, see generally Coffee & Tonry, supra note 59, a t 171-85; Schulhofer, supra note 52, a t 757-72; Schwartz, Options in Constructing a Sentencing System: Sentencing Guidelines under Legislative or Judicial Hegemony, 67 Vn L. REV.637, 680-84 (1981). 102. See Schulhofer, supra note 52, a t 759-60. 103. See 40 Fed. Req. 41,328 (1975). The federal guidelines express this idea by referring to "offense characteristics-severity of offense behavior9'-rather than to the charge of conviction. The Parole Commission's real-offense policy is based on the idea that emphasis on the offense of conviction "would place excessive reliance on convictions obtained more often by negotiation of pleas than by trial of facts. Neither justice nor uniformity of treatment could be achieved with such a system . . . ." See id. a t 41, 330. 104. For example, the rules of the Oregon parole board specifically permit consideration of the actual criminal conduct of the defendant as a basis for departing from the presumptive sentence for the offense of conviction. See OR ADMIN.R. 255-35-035 (1982); see also PANELREPORT,supra note 6, a t 152 (stating that provisions for real-offense sentencing are common in parole guidelines systems). 105. See MODELSENTENCING AND CORRECTIONS A m 3 3-206(d) (Nat'l Conference of Comm'rs on Uniform State Laws 1979). ("the court shall consider the nature and characteristics of the criminal conduct involved wiithout regard to the offense charged"). 106. See Foote, Deceptive Determinate Sentencing, in NAT'L INST.OF LAWENFORCEMENT AND CRIMINAL JUSTICE, LAWENFORCEMENT ASSISTANCE ADMIN.,U.S. DEP'T OF JUSTICE, DETERMINATE SENTENCING: REFORMOR REGRESSION? 133, 134-35 (1978); See a k o Coffee & Tonry, supra note 59, a t 171. Heinonline - - 36 Syracuse L. Rev. 737 1985-1986 738 Syracuse Law Review [Vol. 36:715 offset plea bargains, it might also be seen as a stratagem so unfair as to impugn legal processes.lo7It is argued, first, that critical fact determinations are shifted from the trial to the sentencing state, where adjudication is less formal, less reliable, and where the tribunal will be more favorable to the state. Second, it is claimed that defendants are deprived of the benefits of their bargains in exchange for which they waived their constitutional right to trial, and judges are caused to renege on implied promises to not take dismissed charges into account in sentencing.loSThe issue is genuand attended by due process problems.l1° Apart inely 107. Use of real-offense sentencing has been explicitly rejected in Minnesota. The sentencing commission chose to base guidelines on the offense of conviction because it thought "that serious legal and ethical questions would be raised if punishment were to be determined on the basis of alleged, but unproven, behavior, and prosecutors and defenders would be less accountable in plea negotiation." MINNESOTA GUIDELINES, supra note 69, a t 252; see also Schulhofer, supra note 52, a t 765-72. 108. See Schulhofer, supra note 52, a t 765-72. 109. Professor Schwartz writes that proposals for real-offense sentencing arise amidst principles that push in opposite directions: Principle number one: To punish a person for crimes not charged and proved beyond a reasonable doubt is a denial of the rights to formal specific accusation, public trial, confrontation of accusers, and due process of law. Principle number two: Procedural rights relate to the process of convicting and sentencing on the count of which defendant was found guilty, not to the elements that may be taken into considertion in sentencing on that count within statutory limits fixed for that offense. Principle number three: Due process should govern sentencing procedures. Principle number four: It is consistent with due process in sentencing that many relevant circumstances may be taken into account without being proved beyond a reasonable doubt; informal proof, including hearsay when corroborated, will suffice, a t leaat if the defendant has been given an opportunity to contradict it. Principle number five: Sanctions against an accused, so far as practicable, should be determined by a judge (or parole authority) rather than by the litigating strategy of the prosecutor. Principle number six: Once plea bargaining had been accepted as part of the penal system, it should not be undercut by allowing a judge to take into account offenses dismissed by the prosecutor and thus elude the implications of the bargain. This is the principle of consistency: the law cannot claim respect if it takes away with one hand what it pretends to give with the other. Principle number seven: The law should not command where it cannot secure compliance: if judges will know of other offenses and take them into account, whether or not they do so explicitly, it would be baneful to drive the practice underground by a prohibition that would be obeyed by some judges and disobeyed by others, resulting in sentence disparities. Schwartz, supra note 101, a t 681. For an analysis of the strengths and weaknesses of real-offense sentencing, see Coffee & Tonry, supra note 59, a t 171-85. 110. In Specht v. Patterson, 386 U.S. 605 (1967), the Supreme Court invalidated a Colorado statute permitting a defendant to be sentenced to life for a crime normally carrying a much shorter statutory maximum, if the judge concluded that the defendant constituted a "threat of bodily harm to members of the public, or is an habitual offender and mentally Heinonline - - 36 Syracuse L. Rev. 738 1985-1986 19851 Parole and Sentence Disparity 739 from policy and constitutional questions, Professors Coffee and Tonry see the problems of implementation so great as to make real-offense sentencing feasible only through the vehicle of a parole board."' C. Parole Release Guidelines Following the lead of the federal parole board in the 1 9 7 0 ' ~ , ~ ~ ~ the parole boards of a t least fifteen states have attempted to regulate parole discretion by adopting written standards or guidelines.l13 The concept of guidelines for use in parole decisionmaking ill." Id. a t 607-08.Arguably, real-offense sentencing is comprehended by the Court's language that the Colorado statute "makes one conviction the basis for commencing another proceeding . . [requiring] a new finding of fact that was not an ingredient of the offense charged." Id. a t 608. Professor Schulhofer, however, contends that real-offense sentencing may be critically different than the factual question in Specht, where the elevation in punishment was greatly disproportionate to the seriousness of the crime of conviction. If realoffense sentencing generates penalties within the normal statutory range for the offense of conviction, Specht should not invalidate the practice. See Schulhofer, supra note 52, a t 76065. Even if real-offense determinations do not call for egregiously disproportionate penalties, however, procedural due process almost certainly requires that there be notice and opportunity to be heard, including confrontation, cross-examination, and the use of witnesses, if actual offense findings are sought to be made the basis of sentencing. Id. a t 763-64. See Gardner v. Florida, 430 U.S. 349, 358 (1977) (Petitioner denied due process when death sentence imposed on basis of report which defendant had no opportunity to deny or explain); see also Coffee & Tonry, supra note 59, a t 178-80. 111. The initial determination of the actual offense behavior will likely have been made by the probation department a t the time it prepared its presentence report. Professors Coffee and Tonry question who will present the state's case if the defendant challenges this report and a hearing is held. The prosecutor who agreed to a charged reduction would appear to be undermining his own agreement if he later argues that the real offense is more serious than the charge a t conviction. The probation officer might be assigned that responsibility, but he is a lay person and no match for defense counsel and, further, would appear to be cast in the unfortunate role of both witnesses and advocate. The probation department might be given its own counsel, but since the chief judge in most jurisdictions controls the probation department, i t is argued that even his most neutral agents would be hesitant to disrupt plea bargains approved by the judges. The writers conclude that only a mechanism independent of the sentence work group, such as a parole board, could effectively implement real-offense sentencing. See Coffey & Tonry, supra note 59, a t 183-85. Such a role is already performed, apparently successfully, by the United States Parole Commission. 112. See G ~ R E D S O WILKINS N. & HOFFMAN, supra note 60. 113. Fourteen states and the District of Columbia are reported to have adopted systemwide parole guidelines. See BUREAU OF JUSTICE,STATISTICS, U.S. DEP'T OF JUSTICE, SEITING PRISONTERMS. BUREAUOF JUSTICE STATISTICS BULLETIN(August 1983). The interest in guidelines may be measured not only by the number of jurisdictions actually employing such systems, but also by the number engaging in research leading to the development and possible adoption of parole guidelines. In 1983 planners in the research division of the Texas Parole Board conducted a survey of all 50 states, the District of Columbia, and Puerto Rico, . Heinonline - - 36 Syracuse L. Rev. 739 1985-1986 740 Syracuse Law Review [vol. 36:715 is inconsistent with the notion that sentencing should be wholly indeterminate and that parole officials should have total, unrestricted discretion in deciding individual cases. The guidelines concept is also a t odds with the notion that release dates should be prescribed by statute and that parole officials should have no discretion in their decisionmaking. A guidelines system assumes that parole officials should have discretion, but that their discretion should be structured and controlled without eliminating the opportunity for individual case consideration."' Most parole boards use matrix-type guidelines to articulate their policy as to customary lengths of incar~eration,"~ usually for cases with specific offense (severity) and offender (parole prognosis) characteristics. Decisions departing from the recommended penalties are permitted, but each case requires specific written reasons for such a departure. In general format, the parole guidelines of most states are patterned after those which had been used by the United States Parole Commission before its abolition in 1984.116 The federal matrix was two-dimensional. On the vertical axis of the grid, hearing examiners rated the seriousness of the offender's present offense behavior. This scale enumerated a large number of offenses and ranked them according to severity in seven categories, ranging from "Low," which included such offenses as simple possession of illicit drugs, to "Greatest 11," which included such crimes as murder and aircraft highjacking."' This scale was developed by the United States Parole Commission on the basis of its own subjective evaluation of the seriousness of the criminal behavior underlying an offense.lls On the horizontal axis was an actuarial device, termed the "salient factor" score, which rated the offender's parole prognosis. Based primarily on information concerning criminal history, but and determined (with 48 jurisdictions responding) that 36 jurisdictions were then employing parole guidelines or were actively engaged in research that could result in the adoption of such systems. Telephone interview with Pablo Martinez, Planner, Texas Board of Pardons and Paroles (Nov. 5, 1984). 114. See GOTITREDSON, WILKINS & HOFFMAN, supra note 60, at 1. 115. See PANEL REPORT, supra note 6, at 139. 116. See id. 117. See Paroling Policy Guidleines, 44 Fed. Reg. 26, 542-548 (1979) (codified as amended at 28 C.F.R. 8 2.20 (1984)). WILKINS & HOFFMAN, supra note 60, at 69-80. 118. See GOTTFREDSON, Heinonline - - 36 Syracuse L. Rev. 740 1985-1986 19851 Parole and Sentence Disparity 741 also on items related to the offender's social history, it conveniently summarized knowledge on the relationship between offender characteristics and the likelihood of recidivi~m."~For each combination of offense and offender score, the federal matrix prescribed a fairly narrow range of months of confinement before release on parole. In contrast to the sentencing guidelines adopted by the sentencing commission in Minnesota, parole guidelines inherently concern only the length of imprisonment, not the critical decision of whether to incarcerate an offender.120 Because "in/out" decisions are always beyond the authority of parole boards-they decide only when to release prisoners-much disparity can stand in the shadow of the unregulated discretion judges have in deciding whether to impose confinement. An asserted advantage of parole guidelines, however, is that because a parole board sets standards and carries them out, disparity reduction can be a central decisionmaking concern.121In setting penalty standards, arguably a small, centralized body can achieve greater consistency in policy than a large and heterogeneous group of sentencing judges.122In applying such policy in individual cases, consistently and on a state-wide basis, parole boards can level out dissimilarities in the lengths of sentences imposed on similarly situated offenders.12=Such a system preserves individualized sentencing by judges, a t least symbolically, but protects against disparity through the board's performance, in effect, of a sentence review function.12' The claim that parole guidelines can reduce sentence disparity appears to have been validated in recent studies of guidelines in the federal and some state systems.12sIn most of these studies, researchers examined actual periods of imprisonment before and after adoption of a parole guidelines system, and compared the variability of prison terms set by parole officials acting without 119. See id. at 23,41-67;Hoffman & Adelberg, The Salient Factor Score: A Nontechnical Overview, 44 FED.PROBATION 44 (1980); see also infra note 144. 120. See PANELREPORT,supra note 6, at 140. 121. See id. 122. See id. 123. See id. 124. See id. See also Coffee & Tonry, supra note 59, at 202. REPORT,supra note 6,at 218-19;Cohen & Tonry, supra note 34,at 437125. See PANEL 38. Heinonline - - 36 Syracuse L. Rev. 741 1985-1986 742 Syracuse Law Review [Vol. 36:715 guidelines with the variability in prison terms determined by authorities acting under explicit guidelines.le6 Mueller and Sparks studied the impact of guidelines in Oregon, where the parole board in 1975 began experimenting with a guidelines matrix patterned after the federal mode1.1a7 Controlling for offense severity and offenders' criminal historylrisk assessment scores, they concluded that "the variability of prison terms was less in 1976 and 1978, under the guidelines, than in 1974 before the guidelines were implemented."les Similarly, Arthur D. Little, Inc., and Goldfarb, Singer, and Austern (ADL) examined the impact of parole guidelines in Minnesota, a state where a system of parole guidelines preceded sentencing guidelines.lee Controlling for criminal history, they found less variability in the sentences of aggravated robbery offenders released in 1979, under parole guidelines, than in 1974, before parole guidelines were adopted."O In a study of federal parole, Michael Gottfredson examined a sample of cases classified simultaneously by salient factor score and offense severity, the two scales on the federal guidelines matrix.lS1This scheme of classifying cases was assumed to create categories of "equally situated offenders," much like cases falling in the same cells on a guidelines chart. Using a statistic known as the 126. See, e.g., MINNESOTA REPORTsupra note 86, a t 22 table 1 (table appears supra note 95). 127. See PANELREPORT,supra note 6, a t 218-19; Cohen & Tonry, supra note 34, a t 438 (both citing Mueller & Sparks, Strategy for Determinate Sentencing--Some Statewide Statistical Results (Oregon), in Report on Strategies for Determinate Sentencing (S. Messinger, A. von Hersh & R. Sparks eds. 1982) (unpublished report prepared for the Nat's Inst. of Justice, U.S. Dep't of Justice)). 128. Cohen & Tonry, supra note 34, a t 438 (citing Mueller & Sparks, Strategy for Determinate Sentencing-Some Statewide Statistical Results (Oregon), in Report on Strategies for Determinate Sentencing (S. Messinger, A. von Hersh & R. Sparks eds. 1982) (unpublished report prepared for the Nat's Inst. of Justice, U.S. Dep't of Justice)); see also PANEL REPORT,supra note 6, a t 218-219 (same). 129. See PANELREPORT,supra note 6, a t 219; Cohen & Tonry, supra note 34, a t 437-38 (both citing Arthur D. Little, Inc. and Goldfarb, Singer & Austern, An Evaluation of Parole Guidelines in Four Jurisdictions (1981) (unpubliihed report prepared for the Nat'l Inst. of Corrections)). 130. See PANELREPORT,supra note 6, a t 219; Cohen & Tonry, supra note 34, a t 437-38 (both citing Arthur D. Little, Inc. and Goldfarb, Singer & Austern, An Evaluation of Parole Guidelines in Four Jurisdictions (1981) (unpublished report prepared for the Nat'l Inst. of Corrections)). 131. Gottfredson, Parole Guidelines and the Reduction of Sentencing Disparity, 16 J. RESEARCH CRIME& DELINQ.218 (1979). Heinonline - - 36 Syracuse L. Rev. 742 1985-1986 19851 Parole and Sentence Disparity 743 coefficient of variation,ls2 Gottfredson compared the variability in time-served decisions by parole officials acting under explicit guidelines with the variability in sentence length decisions by judges for the same cases. For each of the eighteen categories of "equally situated offenders" permitting comparison, he showed that determinations by parole officials were less disparate than those by judges.lsS While his findings by themselves did not establish a "guidelines effect," an ADL study of federal parole found less dispersion among periods of actual time served by robbery offenders released in 1979, after guidelines were adopted, than in 1970, before the guidelines were adopted.ls4 Disparity typically implies that defendants in like cases are sentenced differently; that is, "equally situated offenders" are "treated uneq~ally."'~~ I t is essential in any study of disparity that the two elements of this idea receive operational definitions. It is difficult to define operationally how offenders may be "treated unequally," largely because sentences have many different qualitative and quantitative dimensions and there is no single sentence type. The sentencing tribunal must choose among a variety of kinds of sentences, including split sentence, suspended sentences, supervised probation, fines, and confinement, and then must decide on the size of the sentence type selected. This research focuses on variations in the magnitude of only one sentence type-prison terms.ls6 For purposes of this study, "unequal treatment" is defined solely on the basis of the length of incarceration. Disparity exists when "equally situated offenders" receive unequal periods of imprisonment. It is also difficult operationally to define "equally situated of- 132. See infra text at note 153. 133. See Gottfredson, supra note 131, at 226. 134. PANEL REPORT, supra note 6, at 219; Cohen & Tonry, supra note 34, at 238 (both citing Arthur C. Little, Inc. and Goldfarb, Singer & Austern, The Parole Guidelines of the United States Parole Commission: An Analysis of Disparity, in An Evaluation of Parole Guidelines in Four Jurisdictions (1981) (unpublished report prepared for the Nat'l Inst. of Justice, U.S. Dep't of Jusitce)). 135. See PANEL REPORT, supra note 6, at 72. 136. As with any analysis that restricts itself to cases of single sentence type, this study may be vulnerable to selection bias effects that could affect the strength of the conclusions drawn from the research. See PANEL REPORT, supra note 6, at 81. Heinonline - - 36 Syracuse L. Rev. 743 1985-1986 744 Syracuse Law Review [vol. 36:715 fenders," primarily because of the complexity of the factors that affect judges and parole boards in setting periods of incarceration. Although much of the research of the 1960's and 1970's concerned itself with the question of which of a wide range of factors account for the diversity in sentencing outcomes, the Panel on Sentencing Research reported that the general state of knowledge on this question remains largely fragmented.ls7 But the Panel on Sentencing Research also said: "Using a variety of different indicators, offense seriousness and offender's prior record emerge consistently as the key determinants of sentences. The more serious the offense and the worse the offender's prior record, the more severe the sentence."138 While noting that important methodological concerns attend the use of offense seriousness and prior record as variables in sentencing research, the panel argued that because researchers have consistently found that both variables have strong effects on sentences, their use was justified.lS9 Therefore, for the purpose of an initial investigation into the question of whether disparity is reduced by a parole board exercising unguided discretion, it seems appropriate to define "equally situated offenders" on the basis of these two factors.140 Thus, when equal periods of confinement are imposed on defendants who have committed offenses of like severity and who have similar criminal histories, there is no disparity as it is operationally defined in this study. The guidelines and matrices used in most parole systems reflect offense seriousness and the offender's prior record.l4l A guidelines matrix therefore becomes a useful device for identifying "equally situated offenders" for purposes of disparity research. Cases falling within the same cells on the chart may be seen as fairly homogenous with respect to these key determinants, and 137. See id. at 69. 138. Id. at 11 (emphasis added). PSR also said: The strength of this conclusion persists despite the potentially severe problems of pervasive biases arising from the diiculty of measuring - or even precisely defining - either of these complex variables. This finding is supported by a wide variety of studies using date of varying quality in different jurisdictions and with a diversity of measures of offense seriousness and prior record. Id. at 11. 139. See id. at 85-88. 140. See Gottfredson, supra note 131, at 222 (using these same operational definitions of "unequal treatment" and "equally situated" offenders in a study of federal parole). supra note 6, at 139-40. 141. See PAN= REPORT, Heinonline - - 36 Syracuse L. Rev. 744 1985-1986 19851 Parole and Sentence Disparity 745 therefore deserving of similar treatment. The present study analyzes data from the state of Texas. In 1983, on an experimental basis, the parole board in Texas adopted a numerical selection assistance scale, intended to provide a guideline for making release decisions.143A period of research preceded adoption of this measure, during which time the board drew a random sample of 2500 cases, including 1183 adult offenders released on parole in 1981."8 Although release decisions were entirely unregulated by the use of guidelines in 1981, the board, for research purposes only, classified the cases in the sample by salient factor and significant factor scores, two separate scales representing the dimensions of a guidelines matrix then under study. This sample was obtained from the board and used as the basis for the disparity reduction study reported here. The factors making up the salient factor score in Texas, the indicator of probability of success upon release, are identical to those used by the Federal Parole Commission, and consist essentially of information on the offender's prior record, scaled in a vari- 142. See TEXAS BOARDOF PARDONS AND PAROLES,EXPERIMENTAL GUIDELINES FOR PADECISIONS (Aug. 1983). The guidelines system presently used in Texas does not employ a matrix and does not proscribe specific terms or range of terms of confinement. Instead, a single-dimensional scale is used to assist in determining whether or not a prisoner ought to be released. "The scale combines two previously separate scales, the Salient Factor Scale and the Significant Factor Scale." Id. a t 2. The parole board described its numerical selection assistance scale as follows: "Based on a similar tool used by the Federal Parole Board, the scale provides a guidelines for Board release decisions by indicating those inmates who pose a greater risk to society and those who are most likely to succeed under parole superviBOARDOF PARDONS AND PAROLE, THIRTY-SIXTH ANNUAL STATISTICAL REPORT4 sion." TEXAS (1983). The numerical scale is calculated by an institutional parole officer after a prisoner becomes eligible for parole release. Based on their scores, prisoners are assigned to one of three risk categories. Low risk offenders are those whose release or retention requires no explanation. High risk offenders are those whose release must be explained. Telephone interview with Pablo Martinez, Planner, Texas Board of Pardons and Paroles, (Feb. 13, 1984) (notes on file with author). Research into guidelines came a t a time of disruptive growth in the prison population. Texas ranks second, behind California, among the states in the size of its prison population. See generally BUREAU OF JUSTICE STATISTICS,U.S. DEP'T OF JUSTICE, REPORTTO THE NATION ON CRIMEAND JUSTICE: THEDATA(1983). The growth in its system caught planners by surprise, and existing resources failed to meet the demand for prison space. One federal court has found the conditions of confinement to be constitutionally inadequate. See Ruiz v. Estelle, 679 F.2d 1115 (5th cir. 1982), cert. denied, 103 S. Ct. 299 (1982). As one of many steps to insure compliance with the court order, the legislature has directed the Board of Pardons and Paroles to respond to emergency overcrowding situations by advancing parole eligibility and review dates. TEXREV.CIV. STAT. ANN.art. 61840 (Vernon Supp. 1985). 143. See supra note 5. ROLE Heinonline - - 36 Syracuse L. Rev. 745 1985-1986 746 Syracuse Law Review - [vol. 36:715 ety of ways."' The significant factor score, on the other hand, is an empirically derived indicator of the probability of a favorable decision by a parole panel, intended to provide officials with a standard for evaluating their release decision^."^ A major assumption of this study concerns the use of the significant factor score as an operational definition of offense seriousness. Optimally, the variable of seriousness should be based on all important elements of offense behavior, including such matters as weapon use, harm to the victim, and role of the victim (victim provocation and prior relationship between victim and offender), scaled to reflect their combined effect on sentence out~omes."~If important elements of offense seriousness are not measured, there may be measurement errors that result in either underestimates or overestimates of the effect of this variable on sentence outcome^."^ Offense seriousness as measured by the significant factor score is based on victim harm, sentence length, and an offense severity scale that classifies offenses into three categories-low, medium, and high-independent of statutory clas~ification."~Values for these elements are combined linearly to form a single score. While a linear model assumes that the various elements of offense seriousness influence sentence outcomes in a simple additive fashion, the Panel on Sentencing Research noted that the reality is that each element does not always have the same impact on sentence outcome^."^ Nevertheless, the panel said that a linear model was the most commonly used approach to measure offense severity and that it was not clear how any alternate model might combine the 144. See supra text accompanying note 119. The salient factor score sheet used to claasify cases in the sample is reproduced in the Appendix. For a discussion of this score, see TEXAS BOARD OF PARDONS AND PAROLES, BOARD OF PARDONS AND PAROLES, STATEOF TEXAS, HANDBOOK: UTILIZATION OF SIGNIFICANT AND SALIENT FACTOR SCORES 14-24(1981)[hereinafter cited as HANDBOOK]. 145. For a discussion of the significant factor score, see HANDBOOK, supra note 144, a t 2-13.The significant score sheet used to classify cases in the sample is reproduced in the Appendix. I t is based not only on information relating to prior record, but also on offense seriousness, as measured by the length of the sentence imposed, whether the offense involved personal injury, and the ranking of the offense on a severity scale ranging from "low severity," which includes such offenses as burglary and auto theft, and "medium severity," which includes such offenses as arson and assault, to "high severity," which includes such crimes as rape, robbery, and murder. See id. a t 4. 146. See PANEL REPORT, supra note 6, a t 83-87. 147. See id. a t 83-87. 148. See HANDBOOK, supra note 144, a t 3-13. supra note 6, a t 83-87. 149. See PANELREPORT, Heinonline - - 3 6 Syracuse L. Rev. 7 4 6 1 9 8 5 - 1 9 8 6 19851 Parole and Sentence Disparity 747 various elements of this variable to reflect their relative effect on sentence outcomes.'60 Thus, while the significant factor score is an imperfect operational definition of offense seriousness, it nevertheless measures the major elements that are relevant in assessing offense seriousness. Another major assumption of this study concerns the use of the salient factor score as an operational definition of prior criminal record. Ideally, this key variable should include the number, recency, and seriousness of all prior contacts with the criminal justice system-arrests, convictions, and incarcerations-and those contacts should be measured to reflect not only the length of the record, but also its content and the relative impact it has on senThe salient factor score attempts to measure tence out~ornes.'~~ the frequency and seriousness of prior record by including the elements of parole or probation revocation, the number of prior convictions, and the number of prior incarcerations. Like the variable for offense seriousness, the various elements of prior record are combined linearly to form a single score. It should also be noted that the salient factors includes some elements that are not directly related to prior record (for example, education, employment, drug or alcohol dependence), although they may be relevant to the prognosis for success on parole. IV. THEFINDINGS A. The Quantitative Data In figures three and four, two-dimensional tables classify the cases simultaneously by variables assumed to measure offense seriousness and prior record. Within the tables are statistics measuring the variability of sentence lengths and time served for fairly comparable groups of offenders. This statistic is the coefficient of variation-the standard deviation divided by the subgroup mean times 100-a standardized measure of dispersion that is useful for comparing groups having means of a different size.'62 The upper triangle in each cell contains the coefficient of variation for sentence length, the judicial decision; the lower triangle contains the coefficient of variation for actual time served, the parole board de150. See id. at 84. 151. See id. 152. See Gottfredson, supra note 131, at 224-26. Heinonline - - 36 Syracuse L. Rev. 747 1985-1986 748 [Val. 36:715 Syracuse Law Review termination. The lower the coefficient of variation, the less disparity in the group it represents. When this value is smaller for time served than for sentence length, parole board decisions are less disparate than judicial decisions. Figure two classifies the cases by salient factor scores on one axis and by significant factor scores on the other. Both scores are divided into four categories-poor, fair, good, and very good-with the result that the chart contains sixteen cells. Classifications containing less than twenty-five cases were considered too unreliable to present. This table may be understood by viewing it as having four quadrants. It is possible, based on the nature of the salient factor and significant factor scores, to describe the general characteristics expected to be shown by the groups of offenders in each quadrant, as noted in figure two. FIGURE 2 Significant factor score Very good good fair poor very higher scores on both scales good (the "best" offenders) higher salient factor, lower signScant factor scores typically, offenders without extensive criminal histories, good incarcerated for the less serious offenses typically, offenders without extensive criminal records, incarcerated for the more serious offenses (e.g., murderers without prior records) fair lower salient factor, higher significant factors scores lower scores on both scales (the "worst" offenders) typically, offenders with more extensive criminal histories, incarcerated for less serious offenses (e.g., forgers with prior records) typically, offenders with extensive criminal histories, incarcerated for the more serious offenses Salient factor score poor The lower right quadrant represents the group of offenders who have lower scores on both scales-typically, persons with extensive criminal histories who have been incarcerated for more serious offenses, such as murder or rape. This quadrant represents the "worst" offenders. Diagonally opposite is the upper left quadrant representing the group of offenders who have higher scores on both scales-generally, persons without extensive prior records Heinonline - - 36 Syracuse L. Rev. 748 1985-1986 19851 Parole and Sentence Disparity 749 who have been coniined for the less serious offenses, such as forgery, theft, or credit card abuse. They are the "best7' offenders. The two other quadrants on the lower left and the upper right contain data for offenders who combine a higher score on one axis with a lower score on the other. The lower left quadrant represents offenders who have lower salient factor but higher significant factor scores. Typically these are people with more serious prior records who have been confined for less serious offenses. A car thief with an extensive criminal history is one example. The quadrant on the upper right is for offenders with a higher salient factor but lower significant factor scores-persons imprisoned for more serious offenses but who have less serious prior records. The first-time murderer or rapist may exemplify the group in this quadrant. FIGURE 3 Significant factor score fair good Salient factor score fair Figure three shows the coefficients of variation for both judicial and parole board sentence length, using the significant factor score as the measure of offense seriousness. The data suggests that some groups of "equally situated offenders" experience more sentence disparity than others. Observe that in figure three the measures of disparity in judicial decisions-the coefficients of variation in the upper triangle of each cell-tend to increase, from left to right, as offense severity increases. These values are lowest on the extreme left, where offense seriousness is least. Sentences in such cases appear to be relatively uniform. Observe also that disparity Heinonline - - 36 Syracuse L. Rev. 749 1985-1986 750 Syracuse Law Review [Vol. 36:715 appears to increase if the chart is read diagonally, from upper left to lower right. The coefficients of variation increase as cells containing worse cases are encountered. It is in the quadrant on the lower right, for the "wo~st"offenders, that these values are greatest and the sentences most disparate. While judges seem to be in agreement on the simpler cases, it appears that they find it increasingly difficult to reach consensus as the cases worsen, possibly because of the heterogeneity of the underlying offense behavior and because of the opportunity to sentence within a broader statutory range of punishment. The data does not show reductions in sentence disparity for all categories of offenders, but it does suggest that parole has effected substantial reductions of disparity for some of them. The cells enclosed in bold are those in which the coefficient of variation is smaller for decisions as to time served than for decisions regarding sentence length. They represent offenders for whom parole has operated to reduce disparity. While only six of the ten cells are enclosed in bold type, a reduction in disparity is shown for each of the three cells in the quadrant for the "worst" offenders. And the reductions in coefficients of variation in that quadrant are sizeable-ranging from twenty to forty-six percent. Reductions in disparity do occur for the less serious offenders, but they occur less frequently, appearing in only one cell of each of the other three quadrants. Apparently, the more serious cases present the occasion not only for judges to set sentence lengths over a broad range, but also for the parole board to significantly reduce the unevenness that attends the sentencing in such cases. The data was reanalyzed using offense type, rather than the significant factor score as the measure of offense seriousness, to see if the above-described patterns repeated themselves. Figure four reclassifies the cases by the salient factor score, which is again assumed to measure criminal history, and by the four offense types permitted by the size of the sample: forgery, theft, burglary, and robbery. This procedure produced almost identical results. Heinonline - - 36 Syracuse L. Rev. 750 1985-1986 Parole and Sentence Disparity FIGURE 4 Salient factor score good fair poor Forgery Theft Burglary Robbery For each offense type, reading the values in the upper triangle from left to right, disparity in judicial decisions increased as offenders' prior records worsened. In addition, if the upper triangles are read from top to bottom disparity in judicial decisions tended to increase with the seriousness of the offense. Of the four offense types, forgery and theft might be regarded as less serious, burglary as more serious, and robbery as most serious. Note that there was greater uniformity in judicial decisions for the less serious offenses, forgery and theft. Of the five cells for these two offenses, four contained values in the upper triangle of seventy-nine or below. Of the five cells for burglary and robbery, however, four contained values in the upper triangle of 109 or above. Judicial decisions therefore appear to be much more disparate for burglary and robbery than for forgery and theft. The reanalysis confirms that parole can mitigate sentence disparity. Parole board determinations were less disparate than judicial decisions for offenders in six of the ten cells reclassified by offense type. Again, there was greater levelling in the sentences of the more serious offenders. A reduction in disparity was shown in only one of the five cells for forgery and theft, the less serious offenses. On the other hand, for burglary and robbery, parole board determinations were less disparate than judicial decisions in four of the five cells. Furthermore, the reductions in coefficients of variation for these more serious offenses were substantial. For bur- Heinonline - - 36 Syracuse L. Rev. 751 1985-1986 752 Syracuse Law Review [Vol. 36:715 glary, reductions averaged forty percent; for robbery, forty-nine percent. B. The Qualitative Data One of the questions investigated in the field study was whether disparity reduction was a central objective of parole release, or whether it was a byproduct of the process. Investigation disclosed no official statements of policy concerning the agency's role in the elimination of excessive disparity. In the absence of any explicit policy, the writer used observation and interview data to ascertain the attitudes of individual decisionmakers toward parole's performance of a sentence equalization function. All officials interviewed had been furnished with the study's statistical finding that parole had mitigated the unequal treatment of offenders. Parole officials commented that in making release decisions they frequently encountered sentences that appeared to be uneven. Commonly cited examples were those of excessively long sentences imposed either in communities in rural West Texas or by judges known to be particularly severe. The decisionmakers were divided, however, on the question of whether they regarded disparity reduction as an objective of parole. Some responded that they made a conscious effort to reduce the unevenness of punishment, but most disclaimed any intention to perform a sentence review function. A typical reaction was that disparity reduction was a byproduct or secondary effect of parole, but not one of its primary purposes. One may question how it is that disparity has been significantly reduced by decisionmakers whose central concern was not sentence uniformity. Perhaps the answer lies in the differences in the way judges and parole officials work. Parole officials serve in three-person panels, act as a check on one another, and make sentencing decisions exclusively. Moreover, they have no dependence on other participants in the criminal justice process and are free to make independent judgments on sentencing matters. This is in contrast with decisionmaking by judges, who often seek to require the cooperation of other actors in the system and who frequently ratify bargains struck by prosecutors and defense counsel. Lastly, the investigation focused on two questions. First, did parole officials reward guilty pleas in making release decisions? Second, did the decisionmakers consider actual offense behavior or the charge a t conviction? It is widely believed, as studies in some jurisdictions have Heinonline - - 36 Syracuse L. Rev. 752 1985-1986 19851 Parole and Sentence Disparity 753 shown, that defendants who exercise their right to trial receive harsher sentences than like offenders who plead Parole officials, however, when asked whether they rewarded guilty pleas, replied that the mode of disposition was irrelevant in making release decisions. A practice against differential sentencing at the parole stage may have several explanations. It is possible that the parole decisionmakers suspect that the judges have already granted guilty plea discounts and that further concessions would amount to "double discounting." It is also possible that the parole agency, organizationally and functiondy independent of the processes by which guilt is adjudicated, fails to reward guilty pleas simply because it has no institutional stake in the manner in which a conviction is obtained. The officials were also asked whether in making release decisions they looked behind the charge of conviction to the actual offense behavior. The question was posed as follows: "Suppose the defendant had been charged with armed robbery but allowed to plead guilty to simple robbery. If the presentence report said that he had used a gun, would you regard him as an armed robber?" Parole officials responded affirmatively, saying that they would sentence on the basis of the underlying behavior and not just the This informal practice of real-offense senoffense of convicti~n.~~' tencing highlights the independence of the parole agency and reveals its capacity to make decisions based on real rather than artificial differences among offenders. Notwithstanding the above data, this study concludes that parole in Texas has been insufEciently used as a safety-net to correct for disparity a t earlier stages, principally because disparity reduction has not been central goal of parole release. The first step toward achieving a greater measure of sentence uniformity is for the legislature to declare that a primary purpose of parole is the elimination of excessive disparity. Such an announcement would not only set needed goals for administrators, but would also render the 153. See PANEL REPORT, supra note 6, at 114-15. 154. Probation officers are required to send copies of presentence reports to the institution to which the offender is committed. Tex. CRIM.PROC.CODEANN. art. 42.12 5 4(g) (Vernon Supp. 1985). Heinonline - - 36 Syracuse L. Rev. 753 1985-1986 754 Syracuse Law lieview [Vol. 36:715 agency less vulnerable to the criticism that it was overstepping its proper role by reviewing and reducing sentences imposed by judges. The research reviewed in this article throws light on the question of how parole might be structured to better achieve the goal of sentence uniformity. While this study found that parole decisions in Texas were less disparate than judicial decisions in six of ten categories of equally situated offenders, Michael Gottfredson's study of federal parole showed that decisions by officials acting under explicit guidelines were less disparate than judicial decisions in all categories of equally situated offenders.166 Moreover, other studies of parole guidelines systems, including those of Minnesota, Oregon, and the United States, demonstrated that determinations by parole officials acting under explicit guidelines were significantly less disparate than previous determinations by parole authorities exercising unguided discretion.16'j Against the background of these findings, a final specific recommendation is suggested: the legislature ought to empower the parole board to formulate sentencing policy through the adoption of written standards or guidelines that structure discretion in a meaningful way. Such guidelines ought to be patterned after the model followed by most state parole boards. These standards, unlike those that are the subject of experimentation in Texas, prescribe policy as to customary lengths of incarceration for particular classifications of cases. Unless guidelines specify recommended sentences or ranges of sentences, and require justification for departures from them, they fall short of providing officials with any real help in decisionmaking. Given the levelling effect of parole on sentence disparity demonstrated in Texas without such guidelines, it seems logical that sentence disparity could only be further reduced by the adoption of such guidelines. In a more general sense, the findings of this study suggest that parole may be a viable means of achieving uniformity of sentencing for similarly situated offenders. While parole boards do not have the power to remedy dispositional disparity, they are insulated from the systemic pitfalls that exist in determinate sentencing 155. See Gottfredson, supra note 131. 156. See supra notes 125-30 and accompanying text. Heinonline - - 36 Syracuse L. Rev. 754 1985-1986 19851 Parole and Sentence Disparity 755 schemes such as California's. The data from Texas indicates that parole has a levelling effect. Making sentence uniformity an explicit policy and structuring the decisionmaking of parole boards may be an efficient way for state penal systems to achieve the elusive goal of uniformity. Heinonline - - 36 Syracuse L. Rev. 755 1985-1986 Syracuse Law Review A. B. C. D. E. F. G. H. I. [vol. 36:715 Appendix TEXAS BOARD OF PARDONS AND PAROLES SALIENT FACTOR SCORE SHEET Prior Convictions. . . . . . . . . . . . .O Prior = 3 1 Prior = 2 (Adult or Juvenile 2 Prior = 1 3 or more = 0 Prior Incarcerations . . . . . . . . . .O Prior = 3 1 Prior = 2 (Adult or Juvenile 2 Prior = 1 3 or more - 0 Age at First Commitment . . . . .26 years or older = 2 25 - 18 = 1 17 or younger = 0 Commitment Offense . . . . . . . . .Did not involve autho theft, burglary, or forgery = 1 Involved auto theft, burglary, or forgery = 0 Parole or Probation Revoked . .Never = 2 Once = 1 More than once = 0 Drug and/or Alcohol Dependence. .......... .No history = 1 Otherwise = 0 Employment ................ .6 months last 2 years = 1 Otherwise = 0 Education . . . . . . . . . . . . . . . . . . .Completed 12th grade (GED) = 1 Otherwise = 0 Release Plan ................ .Residence and employment = 1 Otherwise = 0 TOTAL Score Success Probability 15-13 85% 12-10 60% 9-7 45% 6-0 30% SOURCE: Eisenberg, supra Indicates Very Good Probability of Success Good Probability of Success Fair Probability of Success Poor Probability of Success note 144, at 17. Heinonline - - 36 Syracuse L. Rev. 756 1985-1986 Parole and Sentence Disparity Appendix TEXAS BOARD OF PARDONS AND PAROLES SIGNIFICANT FACTOR SCORE SHEET I. Criminal Record A. Juvenile Incarcerations .......................... No =; 1 Yes = 0 B. Jails.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0-4 = 1 5 or more = 0 C. Prior Arrests ................................... 0-7 = 1 8 or more = 0 D. Probation Revoked This Incarceration ............Yes = 1 No = 0 E. Adult Prisons.. .................................. 0 = 2 1=1 2 or more = 0 F. Adult Paroles.. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 0 = 2 1= 1 2 or more = 0 G. Adult Parole Revocation.. ....................... No = 1 Yes = 0 11. Current Offenses H. Offense Severity Rating . . . . . . . . . . . . . . . . . . . . . . . . Low = 2 Medium = 1 High = 0 I. Offense Involved Personal Injury ................. No = 1 Yes = 0 J. Sentence (Years). ............................... 0-5 = 2 6-10 = 1 11 or over = 0 111. Institutional Adjustment K. Major Disciplinary Action ....................... No = 1 Yes = 0 Total . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Score: Score: F.I. Initial Second Review Review Probability Indicates 15-14 15-13 80% Favorable Action 13 12-10 50%-80% Probable Favorable Action 12-11 9-7 20%-50% Probable Postponement Postponement 20 % 10-0 6-0 SOURCE: Eisenberg, supra note 144, at 4. Heinonline - - 36 Syracuse L. Rev. 757 1985-1986 Heinonline - - 36 Syracuse L. Rev. 758 1985-1986