United States v. Michael Deen

706 F.3d 760, 2013 WL 452492, 2013 U.S. App. LEXIS 2613

• Court: Court of Appeals for the Sixth Circuit
• Decided: February 7, 2013
• Docket: 11-2271
• Status: Published

Opinion

OPINION

JANE B. STRANCH, Circuit Judge.

In the mid-1880s, Victor Hugo is said to have written, “He who opens a school door, closes a prison.” Our national debate about the relationship between education and crime, both in regard to prevention and rehabilitation, has long raged. A part of it — the part that asks whether rehabilitation is a viable penological goal — has been at the heart of Congress’s shaping of the American prison system. A unanimous Supreme Court recently held that the Sentencing Reform Act does not permit a court to “impose or lengthen a prison sentence to enable an offender to complete a treatment program or otherwise to promote rehabilitation.” Tapia v. United States, — U.S. -, 131 S.Ct. 2382, 2393, 180 L.Ed.2d 357 (2011). Tapia involved a defendant’s initial sentencing. This appeal gives us occasion to consider whether that prohibition applies when a court imposes or lengthens a prison term that follows the revocation of supervised release. We conclude that it does. As a result, we VACATE Defendant Michael Deen’s sentence and REMAND for resentencing.

I. BACKGROUND

In 2008, Michael Deen was convicted of distributing five grams or more of cocaine base and sentenced to prison for 66 months, to be followed by four years of supervised release. Deen was released from prison after the custodial portion of his sentence was reduced and began his supervised release term in March 2011.

Five months later, Deen’s probation officer recommended revocation of his supervised release. The district court conducted a revocation hearing in September at which Deen pleaded guilty to violations stemming from two domestic violence incidents, alcohol use, and his failure to report to the probation office and to attend behavioral therapy.

Although the applicable Guidelines’ policy statements recommended imprisonment for four to ten months, Deen’s probation officer suggested a two-year sentence. The government agreed and pushed for “a significant term of imprisonment where [Deen] hopefully can get some treatment for alcohol abuse, and perhaps counseling in terms of anger management [that] would be very helpful for [him].”

The court sentenced Deen to 24 months’ imprisonment, followed by 24 months of supervised release. The judge explained her decision: “[I]t is important to consider whether the goal of rehabilitation, which I think is the end game in terms of the criminal justice system, can be best achieved through incarceration, and it sounds as though maybe it can.” Deen’s violations of his supervised release conditions did not “bode well for simply continuing supervised release.” Instead, an above-range sentence was necessary to “give the Bureau of Prisons another chance to do some in-depth rehabilitation with Mr. Deen.”

In March 2012, Deen and the government jointly requested that this court vacate Deen’s sentence and remand his case to the district court for resentencing in light of the Supreme Court’s decision in Tapia. Although Tapia involved a defendant’s initial sentencing, the government and Deen argued that its bar on sentencing decisions based on a defendant’s rehabilitative needs applies equally to supervised-release sentences. Deen’s appeal presents an issue of first impression to which we now turn.

II. ANALYSIS

A. Standard of review

“We review a district court’s sentencing decision under a two-part test, ensuring first ‘that the district court committed no significant procedural error,’ and second that the sentence imposed was substantively reasonable.” United States v. Walker, 649 F.3d 511, 513 (6th Cir.2011) (quoting Gall v. United States, 552 U.S. 38, 51, 128 S.Ct. 586, 169 L.Ed.2d 445 (2007)). No procedural errors have been identified here, so we pass only on the substantive reasonableness of Deen’s sentence. Basing a sentence on impermissible factors is an example of substantive unreasonableness, United States v. Webb, 403 F.3d 373, 385 (6th Cir.2005), which we review for abuse of discretion, Gall, 552 U.S. at 41, 128 S.Ct. 586. The standard is the same regardless of whether the sentence was imposed following conviction or, as here, for supervised release violations. United States v. Peebles, 624 F.3d 344, 347 (6th Cir.2010).

B. Statutory framework

Tapia recounts the historical development of federal sentencing laws and the shifting sentiment toward the role of rehabilitation in sentencing that these laws reflect. See 131 S.Ct. at 2386-87. Before 1984, federal law granted district judges wide discretion to craft a defendant’s sentence. Id. at 2386. In doing so, courts could consider a defendant’s amenability to rehabilitation, which was viewed as a sound penological goal that the sentencing decision would further.

Various factors — such as disparities imposed on similarly-situated defendants, changing social attitudes about the wisdom of rehabilitation, and budget constraints— ultimately pushed the concept of rehabilitation into disfavor. The perceived failures of the system led to a sentencing sea change culminating in Congress’s enactment of the Sentencing Reform Act of 1984 (the Act), Pub.L. No. 98-473, 98 Stat. 1987 (codified as amended at 18 U.S.C. §§ 3551-86), which effectively removed much of the discretion judges enjoyed in fashioning prison sentences. Instead, Congress created the Sentencing Commission to draft guidelines that “would provide courts with ‘a range of determinate sentences for categories of offenses and defendants.’ ” Tapia, 131 S.Ct. at 2387 (quoting Mistretta v. United States, 488 U.S. 361, 368, 109 S.Ct. 647, 102 L.Ed.2d 714 (1989)). And it vested control over where federal prisoners are incarcerated and the rehabilitation programs available to them in the Bureau of Prisons (BOP), the administrator of the national prison system. Though Congress sharply cabined the role of judicial discretion in sentencing, it did not wholly eliminate it.

As explained in Tapia, the Act establishes a “framework to govern [district courts’] consideration and imposition of sentences.” Id. It requires a court to order one or more criminal sanctions — which may include imprisonment (almost always followed by supervised release), probation, or fines — when sentencing a defendant. See 18 U.S.C. § 3551(b). The statute further compels a court to consider the extent to which the sentence imposed furthers the four purposes of sentencing — namely, retribution, deterrence, incapacitation, and rehabilitation. See id. § 3553(a)(2). 1 After such consideration, the court must fix a sentence that achieves these purposes “to the extent that they are applicable in light of all the circumstances of the case.” Id. § 3551(a) (emphasis added). “[A] particular purpose,” then, “may apply differently, or even not at all, depending on the kind of sentence under consideration.” Tapia, 131 S.Ct. at 2388.

District courts have authority to sentence defendants to a term of imprisonment pursuant to 18 U.S.C. § 3582. When deciding whether to impose an initial prison sentence — and, if so, its length — the statute requires the court to consider the “applicable” sentencing purposes. 18 U.S.C. § 3582(a). But it admonishes the court to “recogniz[e] that imprisonment is not an appropriate means of promoting correction and rehabilitation.” Id. In Tapia, the Supreme Court held that this statutory warning “precludes sentencing courts from imposing or lengthening a prison term to promote an offender’s rehabilitation,” at least when imprisonment follows a defendant’s initial conviction. 131 S.Ct. at 2391. Congress, Tapia observed, was “skeptical that rehabilitation can be induced reliably in a prison setting.” Id. (internal quotation marks omitted). Fueled by this skepticism, “Congress barred courts from considering rehabilitation in imposing prison terms, but not in ordering other kinds of sentences.” Id. (citations omitted).

. Distinct from the authority to imprison defendants under § 3582, the Act also empowers courts to order supervised release as part of a defendant’s sentence. See generally 18 U.S.C. § 3583. Supervised release permits a court to oversee to varying degrees a defendant’s social reintegration and personal rehabilitation following imprisonment. Section 3583 requires a court deciding whether to include a permissive supervised release term to first consider the extent to which it advances the purposes of deterrence, incapacitation, and rehabilitation (though not retribution). See id. § 3583(c). Since it occurs outside of a prison setting, supervised release is precisely one of the “other kinds of sentences” that a court may order to promote a defendant’s rehabilitative needs.

But § 3583 does more than allow a court to impose a term of supervised release. Based on a defendant’s later conduct, it also permits a court to terminate or extend a supervised release term, modify its conditions, or revoke it altogether. Id. § 3583(e). Critical to this case is the revocation power, which, if elected, will “require the defendant to serve in prison all or part of the term of supervised release.” Id. § 3583(e)(3). A court revoking a defendant’s supervised release term under § 3583(e)(3) and sending him back to prison must again consider the extent to which this action promotes deterrence, incapacitation, and rehabilitation — the same sentencing purposes it contemplated when it imposed the supervised release term in the first place.

And here the issue before us comes into view. On the one hand, Tapia instructs that § 3582(a) bars consideration of a defendant’s rehabilitative needs when sentencing him to prison following his conviction. 131 S.Ct. at 2391. On the other, § 3583(e)(3) appears to require a court to consider that same defendant’s rehabilitative needs when it revokes his supervised release term — which, of course, results in his return to prison for some or all of that term. So does § 3582(a)’s prohibition— grounded in part in Congress’s “skepticism] that rehabilitation can be induced reliably in a prison setting” — apply to § 3583(e)(3)?

C. Controlling precedent

In answering this question, we do not write on a blank slate (even though we ultimately have to wipe away some of what is written on it). In United States v. Jackson, 70 F.3d 874, 879-81 (6th Cir. 1995), we held that a court may properly consider a defendant’s rehabilitative needs in revocation sentencing. 2 Jackson rea soned that while § 3582(a) bars a court from imposing an initial sentence to promote rehabilitation, the court may consider rehabilitative goals in determining the sentence’s length. Id. at 879-80. Transposing this distinction — between the decision to imprison and the length of imprisonment — from the context of an initial sentence under § 3582(a) to a revocation sentence under § 3583, Jackson concluded that a court imposing the latter may similarly “consider[] the defendant’s rehabilitative needs in arriving at the length of his sentence.” Id. at 880-81. Query, though, whether this specific holding may survive after Tapia. We think it cannot.

As noted, Tapia held that § 3582(a) “precludes sentencing courts from imposing or lengthening a prison term to promote an offender’s rehabilitation.” 131 S.Ct. at 2391 (emphasis added). The Court’s analysis traversed the Act’s text, context, and delegation of authority to different actors in the sentencing scheme. First, Tapia observed, § 3582(a)’s language — which states that a court “shall consider” the § 3553(a) factors, including rehabilitation, but may only do so “to the extent that they are applicable, recognizing that imprisonment is not an appropriate means of promoting correction and rehabilitation” — expressly bars consideration of rehabilitation in initial sentencing. Id. at 2388 (quoting 18 U.S.C. § 3582(a)). And the prohibition applies “when determining both whether to imprison an offender and what length of term to give him,” since the court considers the relevant § 3553(a) factors at both turns. Id. at 2389.

Tapia also highlighted another provision in the Act that “restates § 3582(a)’s message, but to a different audience.” 131 S.Ct. at 2390. That provision, 28 U.S.C. § 994(k), tells the Sentencing Commission to ensure that the guidelines it authors reflect that rehabilitation is an inappropriate end to seek “when imposing a sentence to a term of imprisonment.” This, too, dispelled doubts as to the clarity of § 3582(a)’s prohibition.

Finally, the Court found support in the underlying practical problem — “the absence of any provision granting courts the power to ensure that offenders participate in prison rehabilitation programs.” Tapia, 131 S.Ct. at 2390. If Congress wanted “courts to base prison terms on offenders’ rehabilitative needs, it would have given courts the capacity to ensure” their participation. Id. But Congress didn’t. Instead, it gave the BOP “plenary control” over where a defendant serves his time and the treatment programs in which he participates. Id. While a court can make recommendations, “decisionmaking authority rests with the BOP.” Id. at 2390-91.

By contrast, Congress empowered sentencing courts to order a defendant to participate in certain treatment programs and facilities when imposing probation or supervised release. Id. at 2390. Tapia found this allocation of authority to be an “[ejqually illuminating” indicator of Congress’s intent to prohibit imposing or lengthening a prison sentence for rehabilitative purposes. Id. Taken together, these indicia all “point to the same bottom line,” id. at 2391: “when sentencing an offender to prison, the court shall consider all the purposes of punishment except rehabilitation — because imprisonment is not an appropriate means of pursuing that goal,” id. at 2389.

D. Supervised-release revocation sentences and Tapia

Though Congress explicitly required courts to consider rehabilitation when act ing pursuant to § 8583(e), it appears inescapable that Tapia applies to revocation sentencing under § 3583(e)(3), just as it does to initial sentencing after conviction under § 3582(a). First, Tapia teaches that the word “imprisonment,” as it is used in § 3582(a), “most naturally means ‘[t]he state of being confined’ or ‘a period of confinement.’” Id. (quoting Black’s Law Dictionary 825 (9th ed.2009)). We see no difference in the quality of a defendant’s confinement when he “serve[s] in prison all or part of’ his revoked supervised release term and his initial “imprisonment” after conviction. The prohibitory language of § 3582(a) interpreted in Tapia naturally encompasses a court order that requires a defendant to serve time in prison for violating a condition of supervised release.

There are, no doubt, textual contrasts between § 3582(a) and § 3583(e)(3). For example, Congress didn’t use the phrase “term of imprisonment” in § 3583(e)(3), as it did in § 3582(a), but spoke only in terms of requiring the defendant to “serve in prison all or part of the term” after his supervised release sentence was revoked. This objection — that the prohibition Tapia clarified only bears on a sentence that leads to a “term of imprisonment” — has little to recommend it. Congress did not meaningfully distinguish the imposition of a “term of imprisonment” from the act of sending a defendant to “serve in prison.” And the Act makes this clear enough. Elsewhere in § 3583, Congress used the phrase “term of imprisonment” to describe a prison sentence imposed on revocation of supervised release. 3 “Term of imprisonment” has no talismanic effect. Time spent in prison after one’s supervised release is revoked is plainly and unavoidably a “term of imprisonment.”

The more salient textual contrast — that § 3582(a) explicitly disallows consideration of rehabilitative goals, while § 3583(e)(3) states no equivalent limitation — is a better counterargument, but one that cannot carry the day. Our interpretation of § 3583(e)(3) works no harm on § 3583(e)’s more general requirement that a court consider the applicable § 3553(a) factors, including rehabilitation. Even though a court may not revoke supervised release and sentence a defendant as a means to rehabilitate him under § 3583(e)(3), the court may consider rehabilitation when deciding to end a supervised release term under § 3583(e)(1), or extend it or modify its conditions under § 3583(e)(2). Mindful that we must avoid statutory interpretations that render provisions superfluous, see Ellington v. City of E. Cleveland, 689 F.3d 549, 554 (6th Cir.2012), our conclusion executes the statute’s command and explains its language in light of the narrow prohibition of Tapia.

We also find support for our holding in Tapia’s logic. Recall that Tapia found one feature of the Act — its refusal to give a sentencing court authority to order prisoners to participate in rehabilitation programs — to be an “illuminating” indicator of Congress’s intent to prohibit sentencing for rehabilitative purposes. This same ra tionale applies in the context of revocation sentencing. A district court has no more power to order participation in prison-based rehabilitation programs when sending a defendant to prison after revoking his supervised release than it does when initially imprisoning him after conviction. (Note, though, that this rationale only constrains § 3583(e)(3) — where rehabilitation may not be effectuated because the BOP has authority — and has no impact on the other provisions of § 3583 — where the court retains discretionary authority and may effectuate a plan for rehabilitation.) Thus, the statute’s parallel limitation on the sentencing court’s authority strongly suggests that Tapia’s prohibition applies with equal force to revocation sentencing.

Finally, oxxr view comports with the conclusion reached by nearly every circuit that has considered this question. At the time briefing in this appeal was completed, two circuits had held as we do today. See United States v. Molignaro, 649 F.3d 1, 5 (1st Cir.2011) (Souter, J. (ret.), sitting by designation); United States v. Grant, 664 F.3d 276, 281 (9th Cir.2011). Since then, three others have weighed in and agreed with the conclusion we now reach. See United States v. Bennett, 698 F.3d 194, 198-99 (4th Cir.2012); United States v. Taylor, 679 F.3d 1005, 1006-07 (8th Cir. 2012); United States v. Mendiola, 696 F.3d 1033, 1040-42 (10th Cir.2012). Only one has taken a contrary view, see United States v. Breland, 647 F.3d 284, 290 (5th Cir.2011), but its position is in question as the Supreme Court summarily vacated and remanded that decision for reconsideration in light of the government’s position that Tapia applies to supexwised-release sentences. See United States v. Breland, 463 Fed.Appx. 376, 376-77 (5th Cir.2012) (remanding to the district court for resentencing in light of Supreme Court’s order); see also United States v. Garcia, 472 Fed. Appx. 316, 316-17 (5th Cir.2012) (same). In sum, the weight of circuit authority supports Tapia’s application to prison sentences upon revocation of supervised release.

Returning to Jackson, we can only conclude that the permission it gave courts to consider a defendant’s rehabilitative needs when imprisoning him subsequent to revoking his supervised release cannot stand after Tapia. The intervening rule of Tapia — that courts may not “impos[e] or lengthen[J a prison term to promote an offender’s rehabilitation,” 131 S.Ct. at 2391 (emphasis added) — collapses the analytical distinction between the decision to imprison and the length of imprisonment that Jackson rested on, 70 F.3d at 880-81. This distinction can no longer be maintained. Although a panel of this court cannot overrule the decision of another panel, we are to revise our law when “an inconsistent decision of the United States Supreme Court requires modification of the [earlier] decision.” Phillip v. United States, 229 F.3d 550, 552 (6th Cir.2000). We must do so here.

E. Application

With the legal framework clarified, all that remains is to apply it to Deen’s sentence. The district court departed from the Guidelines’ policy statement’s recommendation to send Deen to prison for four to ten months, and sentenced him instead to two years’ imprisonment, followed by two years of supervised release. The court determined that “the goal of rehabilitation ... can be best achieved through incarceration” and ordered a sentence that would “give the Bxireau of Prisons another chance to do some in-depth rehabilitation with Mr. Deen.” This is plainly inconsistent with Tapia’s instruction that a sentencing court “shall consider all the purposes of punishment except rehabilitation — -because imprisonment is not an appropriate means of pursuing that goal.” 131 S.Ct. at 2389. And it leads to the ineluctable result that the district court abused its discretion when it expressly rested Deen’s revocation sentence on his rehabilitative needs.

We recognize that some may assume Tapia will thwart the more hopeful instincts of courts that seek to direct defendants to rehabilitative opportunities that can improve their lives. It should not. Our conclusion applies only to prison sentences imposed on revocation of supervised release under § 3583(e)(3). It leaves untouched a district court’s considerable authority to impose conditions specifically geared towards a defendant’s rehabilitative needs when the court initially orders supervised release under § 3583(a) or modifies its terms under § 3583(e)(2). 4

Moreover, “[a] court commits no error by discussing the opportunities for rehabilitation within prison or the benefits of specific treatment or training programs.” Id. at 2392. Nor will it run afoul of Tapia by “urg[ing] the BOP to place an offender in a prison treatment program.” Id. And, for that matter, a court may always “make a recommendation concerning the type of prison facility appropriate for the defendant,” 18 U.S.C. § 3582(a), considering his needs for “educational or vocational training, medical care, or other correctional treatment,” id. § 3553(a)(2)(D). All of these options remain on the table when a court determines how best to encourage the rehabilitation of a defendant before it.

Trouble only comes when a court imposes or lengthens a sentence “to enable an offender to complete a treatment program or otherwise to promote rehabilitation” inside a prison’s walls. Tapia, 131 S.Ct. at 2393. Although a court may “wisely believe” a prison sentence will have rehabilitative benefits, “those benefits cannot be the reason for imposing it.” Grant, 664 F.3d at 282. A court would do well, then, to articulate the reasons for a prison sentence imposed upon revocation of a defendant’s supervised release in terms of other legitimate criteria.

III. CONCLUSION

The issue in this case is whether it is proper for a court to use rehabilitative goals as the basis of its decision to impose or lengthen a defendant’s prison sentence following the revocation of his supervised release. We recognize that a long public debate on the role of rehabilitation in the American prison system lies beneath this issue. As that debate has expanded to incorporate more empirical study of the viability of rehabilitation, recent developments^ — including the proliferation of judge-involved supervision programs in the federal courts, 5 as well as the growth of scholarship demonstrating that rehabilita tive programs cut recidivism rates 6 — suggest that the “nothing works” 7 consensus reached in the 1970s and 1980s may be in need of reexamination. That full panoply of issues remains for another day. Today, we resolve the propriety of Michael Deen’s sentence. The record in this case permits no conclusion but that the length of his prison sentence was fixed to promote his rehabilitation. This is inappropriate in light of the Sentencing Reform Act and governing Supreme Court authority. As a result, we VACATE Deen’s sentence and REMAND to the district court for resentencing.

1 . Section 3553(a)(2) requires that the court,

in determining the particular sentence to be imposed, shall consider ... the need for the sentence imposed ... (A) to reflect the seriousness of the offense, to promote respect for the law, and to provide just punishment for the offense; (B) to afford adequate deterrence to criminal conduct; (C) to protect the public from further crimes of the defendant; and (D) to provide the defendant with needed educational or vocational training, medical care, or other correctional treatment in the most effective manner[.]

18 U.S.C. § 3553(a)(2).

2 . Although Jadcson interpreted the statutory provision dealing with mandatory revocation of supervised release, see 18 U.S.C. § 3583(g), and the clauses at issue here relate to permissive revocation, see id. § 3583(e), this does not affect our analysis for a simple reason. The permissive revocation provisions explicitly instruct the sentencing court to consider rehabilitative goals, but the mandatory revocation provisions do not. If, as Jackson concluded, a court may consider rehabilitation in the face of the mandatory revocation statute’s silence on this point, it must follow that a court can do so when the Act specifically instructs that it must. See Jackson, 70 F.3d at 880.

3 . See, e.g., 18 U.S.C. § 3583(g) (stating in the mandatory revocation provision that "the court shall revoke the term of supervised release and require the defendant to serve a term of imprisonment”); id. § 3583(h) (empowering the court to impose a subsequent term of supervised release after an earlier “term of supervised release is revoked and the defendant is required to serve a term of imprisonment”); id. § 3583(i) (granting the court power "to revoke a term of supervised release ... and to order the defendant to serve a term of imprisonment” beyond the term's expiration in certain cases); id. § 3583(k) (requiring the court to "revoke the term of supervised release and require the defendant to serve a term of imprisonment” with respect to specified offenses committed by certain sex offenders).

4 . For example, a court may order as a condition of supervised release "any condition set forth as a discretionary condition of probation in section 3563(b) and any other condition it considers to be appropriate.” See 18 U.S.C. § 3583(d). This may include directing the defendant to pursue a course of study or vocational training, id. § 3563(b)(4), refrain from excessive use of alcohol or specified drugs, id. § 3563(b)(7), undergo medical and mental health treatment, id. § 3563(b)(9), and participate in community service, id. § 3563(b)(12).

5 . U.S. Dep't of Justice, Reentry Toolkit for United. States Attorneys’ Offices 8 (2011), http://www.nationalreentryresourcecenter. org/documents/0000/1163/Reentry_Council_ Reentry_Toolkit.pdf (last visited Feb. 5, 2013) (reporting that as of June 2011 approximately half of all federal judicial districts had some form of post-conviction reentry court to reduce recidivism).

6 . See, e.g., David Wilson et al., A Systematic Review of Drug Court Effects on Recidivism, 2 J. of Experimental Criminology 459-87 (2008) (using meta-analytic methods to review 50 studies on the effectiveness of drug courts in reducing future criminal offending and suggesting, with some reservations, that drug offenders participating in a drug court are less likely to reoffend than similar offenders sentenced to traditional correctional options); Ojmarrh Mitchell et al.. Does Incarceration-Based Drug Treatment Reduce Recidivism? A Meta-Analytic Synthesis of the Research, 3 J. of Experimental Criminology 353-75 (2007) (synthesizing results from 66 evaluations of incarceration-based drug treatment programs and finding that therapeutic community, residential substance abuse treatment, and group counseling programs are effective in reducing post-release offending and drug use); David Wilson et al., A Meta-Analysis of Corrections-Based Education, Vocation, and Work Programs for Adult Offenders, 37 J. of Res. in Crime & Delinq. 347-68 (2000) (analyzing 33 evaluations of prison-based education, vocation, and work programs and finding that participants recidivate at a lower rate than nonparticipants); Nana Landenberger & Mark Lipsey, The Positive Effects of Cognitive-Behavioral Programs for Offenders: A Meta-Analysis of Factors Associated with Effective Treatment, 1 J. of Experimental Criminology 451-76 (2005) (reviewing 58 studies on the effects of cognitive behavioral therapy on offender recidivism rates and confirming prior positive findings); David Wilson et al., A Quantitative Review of Structured, Group-Oriented, Cognitive-Behavioral Programs for Offenders, 32 Crim. Just. & Behav. 172-204 (2005) (finding that extant empirical evidence on structured cognitive-behavioral programs that address offenders’ dysfunctional thought processes shows such programs effectively reduce recidivism).

7 . Don Andrews et al., Does Correctional Treatment Work? A Clinically Relevant and Psychologically Informed Meta-Analysis, 28 Criminology 369, 370 (1990) (showing that the then-influential "nothing works” perspective popularized by Robert Martinson, What Works? Questions and Answers About Prison Reform, 35 Pub. Int. 22-54 (1974), is an inaccurate description of rehabilitation and treatment).