9901-4169/79/7001-0077$02.00/00

THE JOURNAL OF CRIMINAL Law & CriminoLOCY Copyright © 1979 by Northwestern University School of Law

Vol 70, No. 1 Printed in USA.

1

INTRODUCTION

Few aspects of criminal justice are currently undergoing more critical appraisal than the incarceration process. Nearly every survey of the field undertaken in the last five years recommends major reform of current methods for determining which convicted persons to imprison and what length of confinement should be imposed. Most recommendations for reform recognize the interdependence of the parole and sentencing decisions. Consequently, proposals for reform typically affect both decisons. Nonetheless, much of the concern surrounding the incarceration process centers on the concept of parole and its contemporary correlate, the indeterminate term. Although there is a growing consensus about the necessity of either abolition or substantial modification of the parole function, there is no consistency in the basis for believing that such reform is required. In fact, proponents of change cannot agree on even the most fundamental effects of the parole process. We are told, for example, that on the one hand, parole leads to decreases in terms of confinement2 and, on the other, to increases in time served.3 A lack of rigorous research in the area encourages such antithetical claims.

One issue that permeates the sentencing-parole field is concern for disparity-dissimilar treatment of equally situated offenders. Numerous reform proposals concentrate on disparity, including suggestions for sentencing councils and appellate review of sentences as well as legislatively fixed mandatory terms and the abolition of parole. To many, the indeterminate term itself accounts for the

1 See D. FOGEL, " WE ARE THE LIVING PROOF" (1975); N. Morris, The Future of ImprISONMENT (1974); A. VON HIRSCH, Doing Justice (1976).

2E. VAN DEN HAAG, PUNISHING Criminals, 6 (1975). Cohen, Abolish Parole: Why Not? 46 N.Y. ST. B. J., 51 (1974).

See also the proposal to reform the Federal sentencing structure in S.1437, 95th Cong., 1st Sess. (1977).

existence of great disparity in incarceration time." The lack of adequate guidelines for the type of information to be considered by the judge in making a disposition, judicial variation in punishment philosophies, and wide discretion in setting the term of confinement are often seen as culminating in gross sentencing disparities.

Under many sentencing structures once the decision to incarcerate has been made the determination of length of confinement is shared between the judiciary and the paroling authority. Within such a structure, the sentencing judge sets the outside boundaries of incarceration time, either by specifying a maximum term, a minimum term or both. Often, the parole board then determines, within these confines, the actual time served in prison. Thus, disparity in the time actually served in prison, for those jurisdictions that employ such sentencing-parole structures, is an issue that is relevant to both decision points.

Although reduction of disparity is not one of the stated goals of most parole systems', it has been argued by some that parole boards do serve to reduce judicially created disparity through the exercise of their discretion in determining the length of confinement. Recent efforts to change the sen

Sentence disparity can, of course, take a variety of forms, including decisions not to prosecute, the incarceration or probation decision, and the decision as to length of confinement. It is only the decision as to length of incarceration that will be of concern here because parole boards, generally, have only had direct influence over this decision.

77

tencing-parole process grant increased importance to ascertaining the validity of such claims about latent functions of the parole process. According to the National Advisory Commission on Criminal Justice Standards and Goals:

8

Though it is seldom stated openly, parole boards often are concerned with supporting a system of appropriate and equitable sanctions. This concern is reflected in several ways, depending upon a jurisdiction's sentencing system. One of the most common is through decisions seeking to equalize penalties for offenders who have similar backgrounds and have committed the same offense but who have received different sentences. Similarly, the United States Board of Parole reports that "to a very real degree, the Board of Parole tends, in practice, to equalize [sentencing] disparity whenever it is not bound to the one-third maximum time required in 'regular sentencing.' But the claim that parole boards do serve a sentence disparity-reduction function still is questioned. After studying parole practices in New York State, one commission has concluded that there is "no hard evidence" that the parole board "reduces sentence disparities by paroling those whose of fenses are similar after they have served comparable amounts of time."

9910

11

Arguments that parole boards do in fact reduce unwarranted variation in sentences are often based on the notion that when one central body makes decisions in every case, it tends naturally to make more homogeneous decisions than would numerous de-centralized decision-makers.' Conversely, the idea that parole boards reduce time-served disparity is often questioned. This is due to the lack of reliable empirical evidence showing such a reduction, and a belief that parole boards make timeserved decisions in an arbitrary fashion in an attempt to achieve ends like rehabilitation that are beyond current capabilities. There does exist some empirical evidence suggesting that state parole boards may modify disparities arising from plea negotiations, but the question of whether and to

a

12

NATIONAL ADVISORY COMMISSION ON CRIMINAL JUSTICE STANDARDS AND GOALS, CORRECTIONS 394 (1973).

UNITED STATES Board of PAROLE, ANNUAL REPORT (1975).

10 Citizen's Inquiry on Parole and Criminal Justice (1976). Report on New York Parole: A Summary, 11 CRim. L. BULL at 297.

"Of course, not all paroling authorities are entirely centralized. They do, however, consist of a small number of decision makers (in comparison to judges).

12J SHIN, ANALYSIS OF CHARGE REDUCTION AND ITS OUTCOMES, (1972).

what extent parole boards reduce judicial incarceration-time disparity is unknown.

A corollary issue with considerable significance in light of contemporary sentencing-parole reform proposals is the extent to which post-sentencing factors influence time-served decisions by parole boards. A major historical argument for the large grant of discretion given paroling authorities to determine the length of incarceration, was to provide an opportunity to observe the inmate's behavior while in prison. In theory, evidence of prison adjustment, as indicated by compliance with institutional regulations and lack of disciplinary actions, and participation in appropriate treatment programs, would permit the parole board better to gauge the inmate's prognosis for successful release from prison. The question of the extent to which institutional behavior factors influence time-served decisions is of considerable import. If factors unknown at the time of sentencing are critical in determining actual time served for a large proportion of inmates, it is obviously important to determine the validity of these factors in assessing post-release success. If post-sentencing factors are not important in time-served decisions, then the practice of providing paroling authorities such wide discretion in time-served decisions is questionable. One purpose of this study, therefore, will be an assessment of the extent to which selected post-sentencing factors are important in determinations by parole boards of time actually served in prison.

13

Problems in Disparity Research

The empirical evidence concerning disparity-reduction on the part of parole boards is sparse and the topic is fraught with difficult analytic and measurement issues. Numerous theoretical and empirical complications attend any disparity research, perhaps partially accounting for the lack of evidence on the effect on disparity of decisions made

13

Keeping control within the institution is an additional rationale for sentencing-parole structures that permit parole boards to consider institutional behavior in time-served decisions (see, eg., N. MORRIS, supra note 1). Regardless of the rationale invoked for the use of institutional behavior factors in time-served decisions, one important question-and the sole concern of the research reported here is the extent to which such factors actually influence time-served decisions. It should be stressed that if it is found that such factors do, in fact, exert a significant influence on time-served decisions, the validity of their use, for either institutional control or for prognosis regarding post-release success, is a separate

issue.

90-231 0 - 82 -- 25

at various points in the criminal justice system. One of the principal impediments to research in the area is the absence of an agreed upon definition of disparity. For example, it might be argued that disparity can only be measured against the specific goals of the sentencing decision. Thus, the factors that legitimately should be considered in arriving at a disposition when general deterrence is conceived of as the goal of the sentencing decision may differ from the factors legitimately relevant if retribution is the sole concern. To the extent that different decision-makers employ different goals upon which to base their judgments, it might be argued that rather than reflecting disparity, differences in dispositions simply reflect differences in the goals of persons making those decisions.

A second impediment to disparity research is the difficulty inherent in operationalizing important concepts. For instance, equity in sentencing might require that offenders with similar offenses and prior records be afforded similar treatment. The measurement difficulty, of course, is how to operationalize these concepts. Categories of conviction offense are most often extremely heterogeneous in the offense behavior that they encompass. Thus, when conviction offense is "held constant" in disparity research, there may still be much uncontrolled variance within categories. Similar problems of scaling and measurement are associated with prior record variables.

Additional complications arise when the aim is to examine the influence of several decisions on disparity. For example, design constraints impede research in the area of disparity reduction by parole boards. Optimally, what is required for an empirical assessment of this problem is an experimental design including random allocation to groups with and without subsequent parole board review as to length of incarceration.

Further complicating the problem for researchers is the lack of comparable data available to both the sentencing judge and the paroling authorities for the same offenders. Thus, a special data collection effort is required for a study of the problem. Finally, major complications in the empirical study of disparity reduction are introduced by the myriad legal restrictions imposed on both the judiciary and the paroling authorities that can make simple comparisons extremely misleading.

The centrality of the disparity issue to the contemporary call for fundamental reform in the sentencing-parole process lends increased importance to attempts to resolve these impediments to empirical study of claims about the latent functions of

parole. A more solid empirical basis for discussion than that currently available is thus required. The major purpose of this paper is to begin an exploration of whether and to what extent one parole board has reduced time-served disparity arising from the dispositions from several courts. Unlike much prior research on the topic of sentencing disparity, emphasis will be placed on the interdependence of the sentencing and parole decisionmakers and the effect that multiple decisions have on incarceration disparity.

THE STUDY

The Sample and the Data

14

To explore the questions of whether and to what extent parole boards reduce judicial disparity in incarceration length and to consider the influence of institutional behavior on time-served decisions, a sample of adult parole cases was obtained from the United States Board of Parole. This sample, which was drawn in conjunction with a larger study, consists of random samples of releases by the Board of Parole in 1970, 1971 and 1972. Because of substantial differences in statutory limitations on maximum and minimum allowable sentences between juvenile and adult cases, only adult cases are studied here. Additionally, persons sentenced under specialized statutes, such as the 1966 Narcotic Addict Rehabilitation Acts were excluded from this sample. Finally, only new court

16

The data used in this study were collected as part of a parole decision-making project directed by Don M. Gottfredson and Leslie T. Wilkins in collaboration with the United States Board of Parole. Their permission to use the data is greatly appreciated. The coding procedures and definitions of terms are reported in D. Gottfredson & S. Singer, Parole Decision-Making Coding Manual, Supplemental Report Two, (Research Center, National Council on Crime and Delinquency, Davis, California, 1973). The proportion of cases drawn by year are as follows: For 1970, 50% of the cases between January and June and 20% of the cases between July and December were randomly selected resulting in 2,497 cases; for 1971, 30% of the cases between July and December were randomly selected resulting in 1,138 cases; for 1972, 30 % of the cases between January and June were randomly selected resulting in 1,011 cases. The sample was drawn prior to the adoption by the Parole Board of the guideline system (see Gottfredson, Hoffman, Sigler & Wilkins, Making Paroling Policy Explicit, 21 Crime and DelinquENCY 34 (1975)), and, therefore, the results of this study may have greater applicability to other jurisdictions that have not established similar discretion structuring mechanisms. 18 U.S.C. §§ 4251-54 (1976).

15

17

18

The federal sentencing structure allows the judge to select among several types of incarceration dispositions for adults. In this sample, the most commonly used alternative (sixty-five percent of the cases) is the "regular adult” sentence, in which the inmate becomes eligible for parole after serving one-third of the full sentence." The full sentence is selected by the judge within statutory confines. Alternatively, the judge may sentence under a section that allows the judge to set the maximum term within statutory confines and to set the date of eligibility for parole at some time earlier than one-third of the maximum. 19 This option accounts for two percent of the cases in the sample. Additionally, an offender may be sentenced under a section that permits the judge to set the maximum term and the parole board to set the earliest parole eligibility date. This option accounted for thirtytwo percent of the cases in the sample. Persons sentenced under these options and who were either paroled, mandatorily released or released by virtue of the expiration of their sentence as of 1973, were defined as the study group for the research. Thus, not only is the judicially set sentence length known for each inmate, but the actual time served as determined by the parole board is known as well. The final study sample consists of 2,833 persons. For each person in the sample, a wide variety of personal characteristics, prior record information, current offense information, and prison experience variables were collected. The reliability of the individual items in the data set was found to be acceptably high with reliability coefficients for most items well above 0.8.21

20

These data are suited for exploring the issue of parole board disparity reduction for several reasons. Both sentencing data from numerous Federal Districts and decision-makers and time-served data are known for each case. Thus, the two types of decisions relevant to the question-judicial determination of sentence length and parole board determination of time served-can be studied for each person in the sample. Also, this data set

17 See Project, Parole Release Decisionmaking and the Sentencing Process, 84 Yale L. J. 810 (1975), for a description of Federal sentencing practices.

18

18 U.S.C. § 4205 (a) (1976).

19 Id. at § 4205 (b) (1) (1976).

Id. at § 4205 (b) (2) (1976).

21 See J. Beck, S. Singer, W. Brown, & G. Pasela, The Reliability of Information in the Parole Decision-Making Study (National Council on Crime and Delinquency 1973).

22

both judicial and parole decisions as to incarceration time-from several indicators of prior record and offense type to prison behavior items—an essential requirement if adequate controls on factors influencing sentence are to be exercised in defining disparity. Additionally, the sentencing-parole structure in existence in the Federal system at the time these data were collected is similar to that found in many jurisdictions, although the types of offense may be dissimilar. Finally, the number of cases available is sufficiently large to permit multivariate analyses of the problem. The question of the influence of institutional behavior factors on time-served decisions is also capable of study, as an effort was made to collect such items that are presented to the parole board for their consideration. Thus, both time-served and some indicants of institutional behavior are known for each person in the sample.

It should be stressed that the data studied here were collected prior to the implementation by the United States Parole Commission of new guidelines23 and that, therefore, these results should not be viewed as reflective of the current practices of that board. The current practices of the United States Parole Commission depart significantly from the practices during the period of time reflected in these data in ways that could exert a considerable influence on the disparity question. Although these findings will not be indicative of disparity reduction in the current Federal System, the operation of the Federal Parole Commission at the time these

22 The following is a list of offense and prior record items used in the study. Precise definitions of terms and coding instructions may be found in D. Gottfredson & S. Singer, supra note 14.

1. Type of Sentence-simple, consecutive, concur

data were collected was similar to the current operation of most parole boards.

Method of Assessing the Disparity Reduction Hypothesis

As usually understood, disparity means that equally situated persons are treated differently at some stage of the criminal justice process. At the decision point of interest here, disparity means that equally situated offenders are sentenced to different lengths of incarceration. There are numerous potential sources for such disparity. It may arise from inconsistency due to individual judges over time, from inconsistency among different judges within a district, from inconsistency among judges in different districts, or from all three. There are two components of the concept of disparity that require operational definitions for the purpose of this study: One considers what is meant by inconsistent or different treatment; and the other questions what is meant by "equally situated offenders". The first component will simply be indicated by variation in the maximum sentence length (for judicial decisions) and in time actually served in prison before first release (for parole board decisions). When "equally situated offenders" receive equal maximum sentence lengths in months, there is no judicial disparity as measured here and when "equally situated offenders" serve equal amounts of time in prison there is no parole board disparity as the term is used here.

The second component of disparity-what is meant by "equally situated offenders"-is, as noted above, much more difficult to operationalize in a satisfactory fashion. There are obviously numerous factors that may be considered by both judges and parole board members in setting length of confinement. Depending on the goal of incarceration that is being pursued by the decision-maker (i.e., general deterrence, incapacitation, retribution, or treatment) the offender's prior record, the seriousness of the conviction offense, the offender's family situation, and the prognosis for recidivism are but a few of the factors that may influence the disposition and, hence, be applicable in defining "equally situated offenders." As noted above, the numerous potential aims of incarceration, with their concommitant differences in "legitimate" sentencing criteria, have led some to argue that disparity cannot be measured and, therefore, cannot be studied empirically.

24 The maximum sentence length was chosen as the most appropriate indicant of judicial disposition since under most sentencing alternatives it defines the range of feasible incarceration time

It is unsatisfactory to argue, however, that because two judges differ in the goals that they employ in fixing sentence, disparity cannot be measured. The point of reference for disparity should be the dispositions given to two or more equally situated offenders. If they receive different dispositionsregardless of differences in purpose for these dispositions-they have been treated differently. If two judges have identical cases and one, operating so as to maximize deterrent goals, incarcerates for two years and the other, operating so as to maximize rehabilitative goals, incarcerates for five years, even though these separate terms may be legitimately related to the goal of the decision, a disparate result, from the point of view of the offenders at least, has been achieved.25

Although there is considerable disagreement over which factors should not be considered in sentencing decisions, there is a good deal of consensus that the characteristics of the offense and the prior criminal record of the offender should be influential. That is, although scholars differ somewhat in the extent to which other factors are seen as permissible in setting punishment, there is a growing body of sentencing literature that suggests that the seriousness of what the offender has done and the extent and nature of the offender's prior criminal conduct should determine the sanction received. There is considerable rationale, therefore, for operationalizing the concept of “equally situated offenders" in terms of these factors. The concept, “equally situated offenders" was thus operationally defined for the purpose of this study as persons with similar current offense and prior record statuses. The data used are relatively rich in the amount of information concerning the prior criminal history of the offender and the factors making up the current offense. 27

Two phases of the research were designed to address the disparity reduction hypothesis. First, multiple linear regression was used to assess the