Wolff at the 14th Annual Justice Brennan Lecture on State Courts & Social Justice

Wednesday, February 20, 2008 - 2:45pm

On Feb. 20th, Missouri's Justice Michael A. Wolff visited the Brennan Center for Justice. After an entertaining and informative lunch with the staff, Wolff gave an address on recidivism and sentencing to the 14th Annual Justice Brennan Lecture on State Courts and Social Justice. Wolff sits on the Missouri Supreme Court and is the Chairman of the state's Sentencing Advisory Commission. The following is his edited speaking draft. The Brennan Center for Justice greatly appreciates Justice Wolff's participation.

Justice Wolff


The Dwight D. Opperman Institute of Judicial Administration
The Brennan Center for Justice
New York University School of Law
February 20, 2008

The 14th Annual Justice William J. Brennan Jr. Lecture
on State Courts and Social Justice

Evidence-Based Judicial Discretion:
Promoting Public Safety through State Sentencing Reform


Michael A. Wolff
Judge of the Supreme Court of Missouri and
Chairman of the Missouri Sentencing Advisory Commission

Speaking draft


Americans put more people behind bars per capita than any country in the western world. But this rate of incarceration is not necessarily helping to reduce crime.  In fact, when we put the wrong people in prison, we make them - and the problem of crime - worse. As we come to realize this, hopefully a new way of thinking about sentencing will emerge that will focus on sentencing outcomes as a way to ensure that public safety is a top national priority.  

Sentencing is a complex topic that needs to be approached with humility, an open mind and common sense.  I believe we have the analytical tools available to help create a system that minimizes recidivism and maximizes public safety.

Let me begin with a case from my state that seems to be the product of old thinking.  The defendant was a 37-year-old man who earned a comfortable living as a construction worker in a rural Ozark Mountain community in southern Missouri.  He was a landlord of some rental property.  He had sole custody of his two small children after his wife had moved to another state.  The sheriff came to the defendant's apartment after one of his tenants called to complain about an altercation.  While in the apartment, the sheriff noticed the remains of a marijuana cigarette and arrested the defendant.  He was charged with possession with intent to distribute.  Despite the fact that there was no evidence of such intent, his defense attorney persuaded him to plead guilty.  He was sentenced to 120 days in prison, with the remainder of the sentence on probation.  Every actor - sheriff, lawyers, judge - in that case exercised discretion.

This was one of the cases we studied in a workshop of Missouri judges and law students that my friend Professor Robert J. Levy of the University of Minnesota Law School and I conducted at Saint Louis University School of Law for ten years, starting in the early 1990s. The participating judges from urban areas were surprised that the sheriff made the arrest, shocked that the prosecutor issued a charge, dismayed at the role of defense counsel and amazed by the sentence. The consensus of the judges and students was that this defendant should not have been sent to prison.

Statistic support their view. Research shows that   probation and community sentences generally have low rates of recidivism, that recidivism rates following prison are two to three times that of probation (depending on the offense) and that the 120-day "shock" sentence this man received has recidivism rates only slightly lower than regular prison sentences.

Over the years, I often have wondered about the outcome of this sentence: Who took care of the defendant's children while he was in prison? Was the defendant employable after prison? 

Lately, when I think about this and similar cases, I think:  enough about this defendant, What about the community's interests ... our interests? 

Specifically, did this defendant commit other offenses? What effect did his imprisonment have on the life outcomes of his children?  Are we safer or less safe as a result of the punishment he received?

[mis-spent resources?]Justice Kennedy's Challenge

Nearly five years ago, in a speech to the American Bar Association, Justice Anthony Kennedy noted the extraordinary rate of incarceration in this country, one in 143 persons, compared with European nations' rate of about one per 1,000.  He summed up the sad state of American sentencing in just a dozen words: "Our resources are misspent, our punishments too severe, our sentences too long."

The over-reliance on prison as punishment makes us less safe, not more. When offenders are sent to prison, they are more likely to re-offend than if they serve probation or community-based sentences.  Most offenders we send to prison, moreover, are sentenced for nonviolent offenses.

Are we better off now - in terms of public safety - than we were five years ago when Justice Kennedy spoke?  I would say we are not.  Today, there are even more offenders in prison than in 2003.  In state and federal prisons and local jails, there are more than two million inmates.  When those of us in the legal academy think of sentencing, we are tempted to think primarily of the federal system. This is odd, of course, because only about six percent of felony sentencing in the United States occurs in the federal courts.  Frankly, for all the attention paid to it, federal sentencing is not a big factor in the day-to-day dispensing of justice in the United States.

In state courts, there are a million felony sentencings per year, of which three-fourths are for nonviolent offenses.  About 97 percent of those sent to prison eventually come back to our communities. That amounts to more than 600,000 persons each year. These are daunting numbers. 

[Prisons and Criminogenisis] Focus on Recidivism

Let us look at recidivism. The Missouri Department of Corrections statisticians recently analyzed ten years of data on the 25 most frequently sentenced crimes for 1995 to 2005.   Most are nonviolent offenses.  Take felony stealing, for example.  Of the 13,000 offenders sentenced to probation or a community sentence, 19.1 percent (one in five) committed another offense. Of the 1,000 or so offenders sent to prison on 120-day sentences, 45 percent committed another offense. And of the 1,900+ [1,921] offenders who went to prison for longer periods, nearly half (48 percent) re-offended.

So, does prison cause recidivism?  The data looks ominous. We need to know out what these numbers mean.  If prison is criminogenic - that is, if it encourages or teaches offenders to commit further offenses - then we will need to find effective punishments that do not make the problem worse.

Still you may wonder: are we safer for having incarcerated a large number of dangerous criminals?  Maybe, at least as far as violent offenders are concerned. But we have not reserved the spaces in our prisons for the most dangerous and most likely to repeat.  We have, in fact, thrown the net far more widely and included many more offenders who, after prison, will be more likely to commit crimes than they were before they went to prison.  Nearly all of them will be back in our communities, and many will resume committing similar offenses or, perhaps, will have graduated to worse forms of crime based on their experiences behind bars, or their diminished life prospects upon leaving prison.

There is a strong case to be made that the large increase in prison population has made us less safe. Looking back to 1980, the worst prisoners already were being incarcerated; the expansion of prisons since then has resulted in incarcerating large numbers of nonviolent, "marginal" offenders who then become recidivists in greater numbers than if punished outside of prison.  People learn from each other:  if we put nonviolent offenders in prison with violent offenders, the nonviolent do seem to learn from the violent. And the other way around? Not so much.

"To be sure," Justice Kennedy said, "the prisoner has violated the social contract; to be sure, he must be punished to vindicate the law, to acknowledge the suffering of the victim, and to deter future crimes."  We must acknowledge that the reason for sentencing is to punish, but the statistics show that if we choose the wrong punishments, we make the crime problem worse, punishing ourselves as well as those who offend.

Unfortunately, most offenders are repeat offenders.  So if we are to think rationally about what is in our own best interests - i.e., public safety - we should focus on what reduces recidivism, and should pay particular attention to which sentences work to reduce recidivism and which programs and punishment-treatment regimens have the best outcomes.

What Does the Public Think?

When we examine issues of crime and sentencing, we are well advised to pay attention to public perceptions and attitudes, for the public seems to be wiser than the politicians who exploit these issues.  A recent survey for The National Center for State Courts found:

  • 1. The public consistently favors "a much tougher approach in sentencing those convicted of violent crimes than ... in sentencing non-violent offenders."
  • 2. "Americans think rehabilitation is a more important priority than punishment and overwhelmingly believe that many offenders can, in fact, be successfully rehabilitated.  But most see America's prisons as unsuccessful at rehabilitation."
  • 3. There are "high levels of public support ... for alternatives to a prison sentence like probation, restitution, and mandatory participation in job training, counseling or treatment programs, at least for non-violent offenders.  The public is particularly receptive to using such alternatives in sentencing younger offenders and the mentally ill."

Ready for Reform?

As a society, we may be near    yet another point at which"sentencing reform" will be on the national agenda or the agenda of some of the states.  Recent U.S. Supreme Court decisions in Gall v. United States and Kimbrough v. United States make it plain: federal sentencing guidelines are advisory, not mandatory. These cases brought federal sentencing into focus. The American Law Institute is fairly well along in a revision of the Model Penal Code on sentencing, though its final product is some years away.  The ABA Kennedy Commission made a number of recommendations based on the following principles:

"(1) lengthy periods of incarceration should be reserved for offenders who pose the greatest danger to the community and who commit the most serious offenses.

"(2) Alternatives to incarceration should be provided when offenders pose minimal risk to the community and appear likely to benefit from rehabilitation efforts."

If we are to apply the ABA's principles and get sentencing right, we must focus on public safety.  We should recognize the centrality of discretion and the need to inform decision makers as to the risks and needs of offenders, and we must measure the effectiveness of treatment programs and the outcomes of sentences.

In terms of sentencing reform, we should try to keep up with the public.

Big Ideas and Smaller Ideas

In the past, sentencing reform has been characterized by "big ideas," including  mandatory minimum sentences, prescriptive sentencing guidelines, "truth in sentencing" (whatever that is) and abolition of parole.   Each seems animated by a preference for incarceration and a distrust of discretion.

I suggest instead, with hope and with some evidence, that there are some smaller ideas we should pursue.  Each speaks to a  particular category of offenders. Just as many of us involved in state sentencing have done, we should differentiate between what works and what doesn't work to produce positive outcomes, i.e., the lack of recidivism.   Each time we find a successful approach, we should be prepared to defend its efficacy with statistics.

The Dilemma of Disparity

The preference for incarceration seems loosely linked to another big idea - that disparity is bad.  The federal sentencing scheme seems designed to address the issue of disparity. The catchphrase of the 1980s was: "If you do the crime, you do the time."  Same crime, same time.  This emphasis on minimizing disparity, in retrospect, may be considered an example of the simple truth that if you ask the wrong question, you are not likely to get a satisfactory answer.

In the real world of state crimes, however, every crime is different, though two factual situations may carry the same legal definition.  More importantly, every person who offends is different. Because we have a 97 percent chance of seeing a particular inmate back on our streets, we should - for all of our sakes, as well as his - be attuned to his needs as they relate to the chances he will offend again. This seems more satisfactory than worrying about whether he got the same sentence as another offender who violated the same statute.  If disparity is based on the severity of the offense and the offender's criminal history - which is part of the way we calculate risk - the disparity seems tolerable.

The federal sentencing system that took effect in 1987 attempted to eliminate disparities by minimizing judicial discretion and making federal sentencing a rule-based system. This was largely the product of one of the big ideas of the time - that the disparities created by judicial discretion made sentencing an essentially lawless activity.

When Missouri established its first sentencing commission in the late 1980s, its mission was simply to study sentencing to determine whether there were disparities.  Money was appropriated and spent.  Disparities were found.

One of the statutory goals of the Missouri Sentencing Advisory Commission, which I chair, is to reduce disparity.   The recent "crack vs. powder" cocaine debate exposed a dramatic problem in disparate sentencing.  We need to fix this kind of specific problem in our sentencing schemes, but also to develop analytical tools that will help ensure that race, gender and location are not the factors that account for disparities.

 "Nothing Works"

The preference for incarceration followed rehabilitation's fall from favor during the 1970s.  Can offenders be rehabilitated?  "Nothing works," was the answer of the time.   That was a big idea, which remains influential to this day.  There is a "large body of rigorous research conducted over the last 20 years of treatments and programs" that are effective in reducing offender recidivism.  And, importantly, the public no longer believes, if it ever did, that nothing works.

                In the 1990s, when a predecessor to our current sentencing commission promulgated advisory "guidelines," I was privileged to enjoy a first hand insight into Missouri judges' attitudes in the course of the workshop Professor Levy and I conducted at St. Louis University.  Here I learned that judges had incomplete and sometimes inaccurate information about programs in the community, or in prison, for addressing the needs of offenders and their families, and they similarly were unaware of what the parole board standards were.  No judge had any idea about risk assessment, though the parole board at the time was using risk assessment to guide its discretionary release decisions.

Informed Discretion

                When the Missouri Sentencing Advisory Commission was reconstituted in 2004, we faced the prospect of doing the same thing over again, that is, promulgating guidelines or recommendations, and somehow expecting different results. But, unlike sentencing commissions in other states, we set out not to restrict judicial discretion, but to better inform its exercise. 

Our discussions, early on, centered on how little each of the various actors involved in sentencing knew about what the others were doing. The parole board offered to share its risk assessment and release guidelines, as well as the data on its actual decisions.  This methodology formed the foundation of the commission's work.  Disclosing the parole board's risk assessment methods and practices to the trial judge at the time of sentencing proved to be very popular with our judges.

                The guiding principle of the current commission's work is that "judicial discretion is the cornerstone of sentencing in Missouri courts."  Coupled with the central idea of discretion, of course, is the smaller idea of enhancing such discretion by data that can help shape the correct placement of offenders. To shape our recommended system, we went through bottom-up process that involved trial judges, prosecutors, and especially, probation officers. Since the system of recommended sentencing was implemented two short years ago, the state's prison population already has dropped by nearly 700 inmates. Concurrent with the statewide implementation of the new pre-sentence information system were the Department of Corrections' improvements in community supervision centers and greater efforts at strengthening re-entry programs. 

                As we have achieved results in sentencing outcomes, I believe it is time to re-think the label "judicial discretion," which is probably much misunderstood to mean that judges get to do whatever they like.  I suggest we re-brand our central concept and call it evidence-based sentencing, for that is what it is: sentences by judges who have considered the evidence that informs their discretion.

Analyzing Risk Factors

Our process involved redesigning the pre-sentence investigation report, which was renamed the Sentencing Assessment Report.  We first held focus groups involving the various actors in the system, and the actual redesign was conducted by a team of probation officers from around the state.  The content of the new Sentencing Assessment Report is driven by the risk assessment factors.

What are risk assessment and needs assessment?  Risk assessment may be defined as predicting who is likely to behave criminally in the future.  Needs assessment may be defined as attempted reduction in criminality by assigning different treatments to different offenders.  The significance of risk and needs assessments is that, taken together, they are the means by which we can try to ascertain what sanctions and which programs are appropriate for an individual offender.

These instruments are far from perfect.  For this reason, it is inappropriate to base the severity of a punishment on a risk-assessment prediction.  But prediction - however imprecise - is often  part of the rationale on which judge's base sentences they impose.  Judges routinely express the view that the are sentencing a defendant to prison in order to protect the public.  Likewise, judges are often willing to give an offenders who seem likely to stay out of trouble a break when it comes to sentencing.

Though  prediction seems inherent in sentencing decisions, experienced trial court judges often express humility about their predictive abilities. Such humility squares with empirical data that suggests actuarial predictions are consistently superior to clinical or human judgments in predicting future criminal behavior.

At the very least, judges should use these statistical measures as a check on their own intuitions and judgments in sentencing. The current draft of Model Penal Code: Sentencing comes to the same conclusion; it recognizes the use of risk-assessment instruments, especially to identify low-risk offenders who should be diverted from prison.

What does risk assessment include? The excellent statisticians of the Missouri Department of Corrections identified 11 factors that correlate with re-offending. Each is scored according to how strongly the factor correlates with re-offending.  Six of the 11 factors are prior criminal history.  Other factors include age, employment status, education and substance abuse.  Being over the age of 45 rates well. (Good news for some of us.) Being under 21 is a minus.  The commission noted, however, that following the statistics strictly with respect to age would result in overly harsh treatment of some youthful offenders who may need to catch a break. When the total scores are calculated, offenders are classified for risk as "good," "above average," "average," "below average" and "poor."

Various states have developed risk-assessment instruments.  Virginia was the first to develop a risk assessment tool for the purpose of diverting low-risk offenders from prison to community sanctions. This effort appears to have produced positive results.  Now only 20 percent of Virginia's inmates are in prison for nonviolent offenses - a substantial contrast to the national figures that show about three-fourths of inmates are in prison for nonviolent crimes. About half of Missouri's inmates are nonviolent offenders.

The questions that risk assessment may help to answer are these: First, are we using prison appropriately? And second, are we using community-based programs appropriately?

If we put people in prison who do not belong there, we may well destroy their lives (and the life prospects of their children) - beyond what their own conduct has done - and make them worse individuals. And if we put people in the wrong kind of community program, we can be wasting our money.

The punishment should fit the offender as well as the crime, to reduce recidivism. But, of course, we should not blindly follow recidivism rates as a sentencing determinant; the types of crimes one may be at risk of committing in the future are also important.  Persons who may commit a violent felony in the future concern us more than, say, a person who commits a petty theft. 

We changed the information made available to sentencing judges, to prosecutors and defense attorneys, and to those who supervise offenders in the community by basing our information on the risk assessment factors.

The risk-assessment score simply informs the judge of the likelihood that an offender with this person's characteristics will re-offend.  The kind and severity of the sentence, as contained in the sentencing commission's recommendations, are based on the severity of the crime and the offender's criminal history.

Missouri's probation officers recently adopted for statewide use an actuarial instrument for assessing the risk of recidivism in sex crimes. This is important at the time of sentencing because a substantial number of men pleading guilty to sex offenses do not receive prison sentences; these non-prison sentences often appear to be as a result of plea negotiations based on the weakness of the prosecution's evidence.

Missouri's adoption of risk-assessment measures is at an early stage.  More refined and more sophisticated use of such instruments will, I hope, develop over time.

Sharing Information about Risk

Risk assessment is an appropriate aid for those involved in sentencing.  I call it a small idea, because it is not the complete answer, for, as noted, no predictive system works anywhere near perfection. 

When we organize the information at sentencing to address an individual's risk factors, we may be able to address more precisely the individual needs of offenders and minimize the risk to public safety in allowing them to serve their sentences in the community rather than in prison.  Using the risk factors as a way of organizing the pre-sentencing information in the Sentencing Assessment Reports, I believe, has improved the results of risk analysis.   

Risk-assessment methodology - whatever its components - ought to be shared by all of those who exercise discretion or judgment in the sentencing process.  This includes prosecutors, defense attorneys, probation officers, judges, prison officials, parole boards and parole officers.  The important attribute of risk assessment is that it does provide a statistically valid measure of factors that affect the outcome of particular sentences.  But the use of risk assessment has a larger purpose, that is, to get all of these actors in the system to think along the same lines - the community's safety and the offender's needs. 

All in the system should have the same risks and needs information, for judges are not the only ones who make discretionary decisions.  As our case from the Missouri Ozarks shows, the law enforcement officer exercises discretion in making an arrest, the prosecutor exercises discretion as to the charge, and the prosecutor and defense attorney exercise some amount of judgment in negotiating a plea.  The probation officer who prepares a pre-sentence report also exercises a good deal of discretion about what to include and what not to include even though the report - as we have reframed it - is governed by the various components of the risk scale. 

To help ensure the success of our new Sentencing Assessment Report, it was important to have the redesign process be driven by some of the 1,200 probation officers who would be using the methodology who do the work on a day-to-day basis. 

Managing the Offender

The probation officer's expertise is essential in assessing the offender's needs.  A very important part of the Sentencing Assessment Report is a section called the "offender management plan," which gives the officer's best judgment as to whether a particular program or treatment option is available and whether it will be appropriate for the particular offender.  The Sentencing Assessment Report contains the commission's recommendations about the proper sentence, but the probation officer's recommendations as to management strategies often are just as important. 

Judges, on their own, do not have the resources to keep up with both the availability of programs and the alternatives to incarceration.  Nor do judges tend to keep up with whether such programs and alternatives are effective.  But probation officers who write Sentencing Assessment Reports and who supervise offenders can develop expertise in what kinds of supervision strategies, restraints and programs will be most effective at reducing the offenders' chances of re-offending. 

In my view, probation officers and judges are becoming more sophisticated at targeting what are called "criminogenic" needs.  These are the particular factors that influence whether the offender will be inclined to re-offend, for example, employment prospects, substance abuse and education.  Prison, as we have seen, is a negative criminogenic factor because it can cause recidivism by exposing some offenders to serious criminals, by diminishing employment prospects, by breaking up families and by traumatizing the prisoner while also increasing the criminogenic risk for other family members. 

A particular sentencing option or treatment should target those needs to reduce the risk of re-offending. 

Examples of community-based programs abound, both through corrections department and private groups.  Our sentencing commission's Web site tries to keep track of what programs are available in each county - from community service to restorative justice to mediation to transcendental meditation. With every program, of course, should come the challenge of showing - by careful and neutral examination of the data - which programs are effective and with what categories of offenders.

Programs that do not reduce recidivism should not be supported.  Remember "boot camps?"  We thought that young offenders would be improved and steered away from lives of crime by undergoing a military-style program of rigorous physical activity and disciplined living.  Most of those boot camps have closed.  The reason: They simply did not work.  What we ended up with was a more physically fit group of offenders who still had not undergone the kind of educational changes needed to move them away from criminal behavior.

Compliance with Sentencing Recommendations

Sentencing Assessment Reports written by probation officers are not used in all cases, but the commission's recommendations, based upon an offender's prior criminal history, are available on the commission's Web site. Our data show that if the sentencing recommendations are followed, recidivism actually is reduced.  Deviations tend to produce greater recidivism, especially when the deviation involves imposing a prison sentence on someone for whom probation or other community sentence is recommended.

Our most recent data, which include all felony sentencings, indicate that the sentences within the commission's recommendations are imposed more than 80 percent of the time.  In about 5 percent of cases, the sentence is more lenient than recommended; in the remaining cases (about 13 percent), the sentence is more severe. 

The statistics show that if the recommendations are followed, recidivism is minimized. This is an important piece of news, a small idea that may grow big.  Future studies, I hope, not only will make overall conclusions - as in this initial study - but also will examine the data for various categories of risk and offenses.

Drug Courts and Other Therapeutic Courts

Many states, including my own, have developed drug courts, mental health courts, DWI (driving-while-intoxicated) courts, re-entry courts and other innovative forms of what have come to be called "therapeutic" or "problem-solving" courts. Of these, drug courts are the most prominent.  Nationally, drug courts have been found effective in reducing recidivism. And drug courts have allayed some of the public's concern that we send too many people to prison for low-level drug offenses.

There are two major positive effects of the drug court movement, in my opinion.  The first is an emphasis on rehabilitation and treatment.  Drug court advocates have persuaded legislatures to fund great increases in community drug-treatment efforts.  The shift from punishment to treatment probably would not have occurred without the drug court movement. 

The second positive effect is the attention that is being paid to outcomes. Before drug courts came to prominence, there was little attention paid to sentencing outcomes, specifically, recidivism.

[[What may be missing from the analysis is a fair risk assessment of offenders participating in drug court programs.  One question is whether drug courts bring into the criminal justice system those who should not be there.  What remains to be shown are the outcomes for moderate to high risk offenders with serious addictions. If a drug court is serving only low-risk offenders - who might do just as well at avoiding recidivism with no drug court interventions  - we may not be getting our money's worth in terms of crime prevention, at the cost of unnecessarily criminalizing a large group of otherwise law-abiding persons.]] 

When we refer to "evidence-based" sentencing, one of the things we look at is whether a particular treatment is appropriate and whether it is effective. The drug court movement has given us some tools to use with respect to those who have substance abuse problems. 

In my state and many areas in the country, there are similar kinds of efforts called "mental health courts," some of which are set up to deal with the problem of nuisance crimes committed by offenders who have noticeable mental health issues.  We have found in Missouri that approximately two-thirds of offenders - probationers, parolees and prisoners - have mental health records.  More than nine out of ten offenders with known substance abuse and mental illnesses have records in the Department of Mental Health. We have a lot of work to do in addressing the mental health -criminogenic - needs of offenders and their families.

The innovative methodology of drug courts also is being adapted by some state courts that have created "re-entry courts," another model of therapeutic jurisprudence.  Re-entry has been a continuing challenge for parole officers, and there are some cases, undoubtedly, where intervention by a court, in a re-entry program, would seem better suited than parole supervision.  The need is particularly acute because the numbers are large - 600,000 released from prison each year - and about 130,000 of those have completed their sentences and are not required to report to anyone.  As opposed to drug courts, which focus on the behavior of the offender, some re-entry courts use what is called the "managerial" model, because the court functions as a manager or an advocate for getting those services to the released offender that are needed to re-adjust to life in the community. Improvements in the re-entry process -whether through re-entry courts or through improvements in parole supervision or through community-based service providers outside the criminal justice system - are essential to reducing recidivism. 

Principles and Recommendations

Let me end with a series of what I believe are simple recommendations, small ideas that I believe will help address the problems that we now face. 

  • 1. Punishment should be no more harsh than warranted.  This is a central message of the American Law Institute's Model Penal Code Sentencing revision.  Longer prison stays do not reduce recidivism.
  • 2. There should be no mandatory minimum sentences. Such provisions may seem to be politically popular, but mandatory minimums are ineffective at reducing recidivism and often have dysfunctional, unintended consequences.
  • 3. "Evidence-based sentencing" should replace the much-misunderstood phrase "judicial discretion."  As with many decisions in our courts and in our criminal justice system, discretion is inherent.  But we should be prepared to defend our decisions by basing them on evidence that includes an assessment of the offenders' risks and needs.   
  • 4. We should have a preference for community-based sanctions, especially for nonviolent offenders, rather than a preference for incarceration.  Prison should be reserved for those we are afraid of, not those we are mad at.
  • 5. Everyone who works with an offender should know that person's risks and needs. All who work in the system - prosecutors, judges, probation officers, parole authorities and prison officials - should use the common language of risks and needs for managing the offender. 
  • 6. The goal of every sentence - whether in the community or in prison - is not only to punish but also to minimize the chances of recidivism. For any prison sentence less than life, from the day an offender enters prison, the system should be preparing for his or her release by developing a re-entry plan that will put that person back in the community with enough support to reduce the chances of re-offending. 
  • 7. All treatment programs, both in prison and in the community, should be evaluated on an ongoing basis, particularly with how well they meet the criminogenic needs of moderate and high-risk offenders.  The measurement of success is very simple:  Is the particular program effective in avoiding recidivism?
  • 8. We should evaluate sentencing outcomes.  For the most frequently committed crimes, by each category of risk, we should track the recidivism data for prison versus various forms of community sentences.  The important thing is that we need to inform those involved in sentencing - judges and prosecutors, especially - as to which sentences actually increase recidivism for particular categories of offenders. 
  • 9. Last, but not least: We should keep the public informed of what we are doing.  As the National Center for State Courts survey indicates, the public will understand.  The public wants to know that the sentencing that is done in their names - and by their authority - is promoting their safety.  We should make sure that - to the best extent humanly possible - we do that.