Evidence-Based Strategies for Working with Offenders

Over the last 20 years, a critical mass of social science evidence has accumulated challenging what had previously been the prevailing notion that nothing works in the rehabilitation of offenders. Findings from academic and program evaluation literatures in the elds of psychology, criminal justice, sociology, and public policy suggest that evidence-based interventions, which effectively combine the core principles of rehabilitation (risk-need-responsivity), deterrence, ACKNOWLEDGEMENTS 2 procedural justice, and collaboration, can signicantly reduce recidivism. Additionally, emerging TABLE OF CONTENTS 3 treatments for previously undertreated and underpublicized criminogenic needs (i.e., cognitiveEXECUTIVE behavioral SUMMARY therapy for criminal thinking) are proving feasible and effective with offender popula- 4 tions in the United States and abroad. INTRODUCTION 5
FINDINGS This fact sheet seeks to distill a growing body of research about evidence-based strategies in RECOMMENDATIONS

ve areas for reducing recidivism among criminal offenders: (1) assessment, (2) treatment, (3) 13 deterrence, (4) procedural justice, and (5) collaboration.
14 17 21 26

A. IDENTIFYING AND DEFINING THE ISSUE B. CONNECTING ATTENDANCE TO FUTURE OUTCOMES

USE VALIDATED SCREENING AND ASSESSMENT TOOLS TO DETERMINE OFFENDER RISKS AND NEEDS. C. ADDRESSING STRUCTURAL BARRIERS TO SCHOOL ATTENDANCE

1. ASSESSMENT

D. PROVIDING INTERVENTIONS FOR YOUTHS AFFECTED BY, OR AT-RISK FOR, CHRONIC ABSENTEEISM

Offenders vary widely both in the future risk they pose to public safety and in their specic CONCLUSION 33 treatment needs. Evidence-based screening and assessment protocols can help criminal justice ofcials match each offender to an intervention of appropriate type and intensity. Screening WORKS CITED 34 refers to the use of one or more brief tools to identify possible risk and needs early in the justice YOUTH JUSTICE BOARD MEMBER STAFF BIOGRAPHIES 36 system process, such as at the AND booking or initial arraignment stage. Screening tools indicate the need for further assessment and typically do not exceed 10 or 15 minutes. Assessment refers 40 APPENDIX 1: RESEARCH DESIGN to a longer evaluation process that should occur before an offender is matched to a particular APPENDIX THE Both YOUTH JUSTICE BOARD PROGRAM tools should be validated and should focus on 42 treatment2: type. screening and assessment major criminogenic risk and need factors, or those factors that research has demonstrated to be statistically correlated with recidivism (see below).

This publication was written by Michael Rempel, director of research at the Center for Court Innovation. The publication was supported by Grant No. BJA 2011-DCBX-K002 awarded by the Bureau of Justice Assistance. The Bureau of Justice Assistance is a component of the U.S. Department of Justices Ofce of Justice Programs, which also includes the Bureau of Justice Statistics, the National Institute of Justice, the Ofce of Juvenile Justice and Delinquency Prevention, the Ofce for Victims of Crime, and the Ofce of Sex Offender Sentencing, Monitoring, Apprehending, Registering, and Tracking. Points of view or opinions in this document are those of the author and do not necessarily represent the ofcial positions or policies of the U.S. Department of Justice. Published April 2014.
1

What does it mean for a screening or assessment tool to be validated? It means that the tool actually measures what it purports to measure. If the tool is attempting to measure risk of recidivism, it is validated if research has demonstrated that individuals classied by the tool as high-risk are indeed much more likely than others to re-offend and those classied as low-risk are indeed much less likely than others to re-offend. Or if the tool is attempting to measure a clinical disorder, post-traumatic stress disorder (PTSD) for example, it is validated if research has demonstrated that, after further in-depth clinical assessment, those individuals that the tool ags as having PTSD symptoms are indeed much more likely than others to have a conrmed PTSD diagnosis. Evidence-based screening and assessment can help differentiate offenders both by the future recidivism risk they pose (low, medium, or high) and by their unique treatment needs. Offenders may have multiple needs (e.g., addiction, employment and skill decits, family dysfunction, or co-occurring mental health disorders). Only a thorough screening and assessment protocol can ensure that all offenders have their individualized needs understood and treated effectively. Criminal justice institutions often lack the time to conduct lengthy assessments on every offender. Fortunately, a great many short screening tools have been developed, involving anywhere from six to 30 questions. Texas Christian University is perhaps especially notable for the large number of non-proprietary (i.e., free of charge) screening tools it has developed, covering these topics and many others.1 Short screening tools can be used to weed out ineligible offenders. However, before assignment to specic interventions, partici-

pants should, ideally, receive a more thorough evidence-based assessment. What should such an assessment include? Research indicates that there are a group of eight criminogenic risk/need factors, known as the Central Eight, which are strongly associated with recidivism.2 The assessment protocol should cover most or all of these eight factors. The rst four (the Big Four) are the most predictive of recidivism. They are (1) a history of criminal behavior, (2) an anti-social personality, (3) criminal thinking patterns, and (4) frequent interaction with anti-social peers. Interventions that target substance abuse or mental health often make the mistake of solely assessing for those problems, without also assessing for the Big Four. Less important but also inuential are the next four factors (the Moderate Four): (5) unmarried or otherwise experiencing family instability, (6) unemployed/unemployable, (7) not involved in pro-social leisure activities (i.e., prone to hanging out or trouble), and (8) substance abuse. All eight of these factors are covered by many existing risk/need assessments, such as the proprietary COMPAS or LSI-R tools or the non-proprietary ORAS tools.3 Surprisingly, research is mixed as to whether or not mental health disorders are associated with an increased risk of recidivism. Despite the lack of a clearly proven link between mental illness and recidivism, mental health disorders can still deeply affect the quality of life of those in treatment. In addition, if left unaddressed, mental health problems can interfere with the effectiveness of treatments for the Central Eight factors.4 That is why it remains useful to supplement assessment for the Central Eight with a screening or assessment protocol focused on trauma/PTSD and other major mental disorders (i.e., bipolar disorder, major depression, schizophrenia).

2. TREATMENT
APPLY RISK-NEED-RESPONSIVITY (RNR) PRINCIPLES WHEN MATCHING OFFENDERS TO INTERVENTIONS.

The past 50 years of research has provided support to three principles of offender intervention that are generally identied with the Canadian researchers Donald Andrews and James Bonta, but have been embraced across the Western Hemisphere. They are Risk, Need, and Responsivity. Risk Principle: This principle recommends varying the intensity of treatment by risk level. Specically, intensive interventions are most suitable for medium-risk or high-risk offendersthose who are especially predisposed to re-offend in the rst place. However, interventions may have unintended deleterious effects if they are used with low-risk offenders. Specically, intervening with low-risk offenders may increase re-offending by creating counter-productive obstacles to their participation in prosocial work or school activities, exposing them to negative inuences from high-risk peers in group intervention settings, and unnecessarily labeling them as criminal.5 Need Principle: To be successful in reducing recidivism, this principle recommends placing offenders in interventions that target the Central Eight risk/need factors. (See Assessment above.) The principle implies that recidivism reductions will be maximized when offenders receive interventions that address multiple needs, rather than only one need, such as substance abuse treatment. To offer a common example, many high-risk offenders exhibit evidence of criminal thinking, which involves maladaptive thoughts and attitudes. These include denial of the legitimacy of laws and legal institutions (e.g., laws are wrong, laws are racist, or I live by a higher code); a perception that life outcomes fall outside ones own control (e.g., the system will always keep us down); and excuses for anti-social behavior (e.g., no one was hurt, if someone was hurt,

she/he deserved it, or blaming others for ones own actions). The concept of criminal thinking also encompasses poor judgment and decisionmaking skills (e.g., high levels of impulsivity and reactions based in anger), often leading to anti-social behavior. Fortunately, effective treatments exist for criminal thinking, such as Thinking for a Change (T4C), Moral Reconation Therapy (MRT), or Reasoning and Rehabilitation (R&R).6 Recidivism reductions will be maximized as criminal thinking treatments are paired with interventions targeting other needs (e.g., substance abuse or unemployment). Responsivity Principle: This principle holds that treatment is most effective if it (1) employs a cognitive-behavioral approach and (2) tailors the focus of the cognitive behavioral treatment to the specic learning style and attributes of the offender. Research demonstrates that when the goal is to reduce recidivism or related criminogenic needs, it is best to avoid educational groups, which are designed to educate participants on various symptoms and disorders (e.g., biochemistry of drug addiction). It is also best to avoid unstructured self-help groups that are not organized around an evidence-based written curriculum. These latter approaches have not been shown to work, despite what is oftentimes an intuitive appeal. By contrast, structured cognitive-behavioral approaches have been shown to reduce recidivism by unpacking and restructuring the beliefs, attitudes, and thinking patterns that lead to risky behaviors and by providing offenders with new decision-making skills and strategies for impulse control and behavior modication. In applying cognitive behavioral approaches, different curricula may be necessary for different populations, such as adolescents, young adults, women with children, or trauma victims; in short, the idea that cognitive-behavioral approaches are broadly preferred does not mean

that they should be utilized with a one-size-tsall model.7 Interventions that attempt to follow the Risk, Need, and Responsivity principles can produce disappointing results if they are not well implemented. Specically, interventions are more effective when a high percentage of partici-

pants actually complete the full program; when manualized (written) curricula are used; when counselors who lead group treatment sessions routinely debrief with and receive feedback from supervisors; when supervisors rigorously verify delity to the intended curriculum; and when ongoing staff training and protocols are followed.8

3. DETERRENCE
IMPOSE CERTAIN AND CONSISTENT CONSEQUENCES IN RESPONSE TO NONCOMPLIANCE.

Research indicates that supervision is inadequate by itself. Supervision typically involves frequent required check-ins with probation or parole ofcers or with a judge. However, when combined with proven strategies for therapeutic engagement and the consistent application of sanctions for noncompliance, supervision can then become an effective tool.9 In particular, sanctions have been shown to be most effective when they involve certainty (each infraction elicits a response), celerity (response is imposed soon after the infraction), and severity (responses are sufciently severe to deter misbehavior but not so severe as to preclude more serious sanctions in the future).10 Some research indicates that certainty is the most important of these principles. Thus, allowing repeated noncompliance to go unpunished and then suddenly imposing a severe sanction

(e.g., probation revocation) will be less effective than imposing milder sanctions in response to each and every noncompliant act.11 Research also suggests that offenders do not always appreciate what will happen if they are noncompliant, even if it is explained once or seems obvious to criminal justice ofcials. Engaging in frequent remindersconveyed in clear, non-technical languageon the positive consequences of compliance and the negative consequences of noncompliance increases program completion rates and reduces recidivism.12 While sanctions are important, research also indicates that positive incentives (verbal praise or tangible incentives like cash value certicates) can play an important role in changing behavior, especially if the incentives are frequently and predictably administered according to a schedule.13

4. PROCEDURAL JUSTICE
ESTABLISH FAIR AND CONSISTENT PROCEDURES AND TREAT OFFENDERS RESPECTFULLY.

Procedural justice refers to the fairness of justice procedures and interpersonal treatment of defendants or other litigants. Procedural justice is commonly contrasted with distributive justice, which concerns the fairness of the nal outcome (e.g., whether a litigant won or lost the case). Interestingly, some research indicates that litigant perceptions of procedural justice can actually play a greater role in their over-

all assessment of their court experience than whether or not they like the case outcome.14 In plain terms, litigants prefer to win their case, but they tend to accept losing if they consider court procedures and their interpersonal treatment to have been fair and respectful. Research also indicates that offenders who report a high level of procedural justice are more

likely to comply with court orders, to perceive laws and legal institutions as legitimate, and to engage in future law-abiding behavior.15 The basic dimensions of procedural justice include: Voice: Offenders have an opportunity to be heard, either directly or through their attorney. Respect: Offenders are treated with dignity and respect. Trust/neutrality: Offenders perceive decisionmakers as neutral and competent and their decisions as unbiased and accurate. Understanding: Offenders understand decisions, including the reasons for those decisions,

and understand any future responsibilities they have to comply with court orders. Helpfulness: Offenders perceive that decisionmakers have a genuine interest in their needs and their personal situation. It is unclear whether any one of these dimensions is more important than any other, although a recent study found that perceptions related to respect exerted the greatest inuence on defendants overall satisfaction with how their court case was handled.16 In addition, defendant perceptions of the judge are strongly associated with overall perceptions of the criminal justice systemand in turn with future law-abiding behavior.17

5. COLLABORATION
OBTAIN THE BUY-IN AND PARTICIPATION OF MULTIPLE CRIMINAL JUSTICE AGENCIES, INCLUDING BOTH TOP LEVEL OFFICIALS AND LINE STAFF.

Some research has found that criminal justice programs tend to be better implemented when strong interagency collaboration is present

and when that collaboration includes buy-in among the line-level staff who will be expected to implement the program on the ground.18 In

The Problem of Misdemeanor Offenders

Many of the evidence-based strategiesespecially those related to assessment, treatment, and deterrencepresume that it is legally feasible to order offenders to participate in lengthy interventions and to impose meaningful consequences in the event of noncompliance. However, courts and other criminal justice institutions are constrained by the principle of proportionality, which means that many crimes are simply too low-level to justify lengthy periods of treatment or community supervision. Many misdemeanor offenderswho comprise well over half the total number of offenders nationwideare at high-risk for future crimes and/or possess a multiplicity of serious treatment needs. However, the demands of legal proportionality mean that it is not possible to link these individuals to the high treatment dosage they might need from a strictly clinical perspective. Given these factors, there is currently an urgent need to develop evidence-based short-term interventions for misdemeanor offenders. Motivational interviewing may be a particularly appropriate practice that can be implemented in only a few sessions of offender-counselor interaction.* Evidence-based strategies that do not rely on lengthy interventionssuch as protocols designed to promote procedural justicemay also be especially critical when dealing with misdemeanants. The problem of misdemeanor offenders is a cutting-edge area where existing research has not yet provided denitive guidance, but where local innovation is to be strongly encouraged and promoted.
* Motivational interviewing is an evidence-based practice that takes a client-centered approach to elicit motivation for change. To learn more about motivational interviewing, visit: www motivationalinterview.org.

fact, a recent survey of criminal justice leaders nationwide cited lack of buy-in from line-level staff as the second most important obstacle (after lack of funding) explaining why some innovative programs fail to achieve their goals.19 Providing more tangible evidence for the role of collaboration, two recent evaluations of programs for drug-addicted offenders both found that recidivism reductions were correlated with broad inclusion of court staff, prosecutors, defense attorneys, treatment representatives, and law enforcement in policy planning meetings and court sessions where treatment participants had their progress monitored by the judge.20
FOR MORE INFORMATION:

D2)#E)'5'813#NDD%8-%)0"#N032&9#G2)&1BI#@4%#E15/(%33#E2331(2)1*'28"
c"#$%%#J'/0%B9#O"["#%*#13"#=6>>`?"#N/#E'*" f"#$%%#A%*%)0'3'19#<"#=!fff?"#a7#:%+1-%#2D#^K/%)'5%8*'8C#&'*4#R8*%)5%-'1*%# $18+*'280I#[41*#e1H%#[%#J%1)8%-Tb#A%)0/%+*'H%09#6LI#LfPXXQ#$4%)5189#J"["9# _2**D)%-0289#:"9#O1+W%8Z'%9#:"9#^+W9#<"9#S%,*%)9#A"9#18-#;,04&1B9#$"#=!ff`?"# A)%H%8*'8C#E)'5%I#[41*#[2)W09#[41*#:2%08 *9#[41* 0#A)25'0'8CT#[104'8C*289# :EI#G1*'2813#R80*'*,*%#2D#<,0*'+%Q#@1K5189#M"#=6>>6?"#a$,/%)H'0'28I#^K/32)'8C#*4%# :'5%80'280#2D#^DD%+*'H%8%00"b#M%-%)13#A)2(1*'289#ddI#!XP6`" !>"#$%%9#%"C"9#A1*%)820*%)9#S"#18-#A'g,%)29#7"#=!ffF?"#S%+28+%/*,13'Z'8C#:%*%)P )%8+%I#78#^5/')'+13#@%0*#2D#A%)02813#h'+1)'2,0#^K/%)'%8+%0"#<2,)813#2D#S%0%1)+4#'8# E)'5%#V#:%3'8g,%8+B#L6I#6F!P6cd"# !!"#$%%#E'008%)9#7";"9#S%5/%39#O"9#[13W%)#M)18W3'89#7"9#S25189#<"]"9#;'%3%)9#$"9# E24%89#S"9#18-#E1-2)%*9#E"#=6>!L?"#7#$*1*%&'-%#^H13,1*'28#2D#G%&#i2)W 0#7-,3*# :),C#E2,)*0I#@%0*'8C#[4'+4#A23'+'%0#[2)W#;%0*"#G%&#i2)W9#GiI#E%8*%)#D2)#E2,)*# R882H1*'28Q#18-#G1C'89#:"#18-#A2C1)0WB9#_"#=6>>!?"#aR8*%C)1*'8C#E%3%)'*B9# R5/,30'H'*B9#18-#^K*)13%C13#$18+*'28#@4)%1*0#'8*2#1#O2-%3#2D#_%8%)13#:%*%))%8+%I# @4%2)B#18-#^H'-%8+%"b#E)'5'8232CB#LfI#X>XPXL>"# !6"#i2,8C9#:"#18-#;%3%8W29#$"#=6>>6?"#aA)2C)15#S%*%8*'28#18-#A%)+%'H%-# E2%)+'28#'8#@4)%%#O2-%30#2D#O18-1*2)B#:),C#@)%1*5%8*"b#<2,)813#2D#:),C#R00,%0# L6I#6f`PL6c" !L"#O1)32&%9#:";"#18-#[28C9#E"<"#=6>>c?"#aE28*'8C%8+B#O181C%5%8*#'8#7-,3*# E)'5'813#:),C#E2,)*0b#'8#e'CC'809#$"@"9#$'3H%)5189#]"9#18-#e%'39#$"#e"#=%-0"?9# E28*'8C%8+B#O181C%5%8*#'8#$,(0*18+%#1(,0%#@)%1*5%8*#=//"#LLXPLFX?"#G%&# i2)W9#GiI#_,'3D2)-#A)%00" !X"#$%%#@B3%)9#@"S"#18-#e,29#i"<"#=6>>6?"#@),0*#'8#*4%#J1&I#^8+2,)1C'8C#A,(3'+# E22/%)1*'28#&'*4#*4%#A23'+%#18-#E2,)*0"#G%&#i2)W9#GiI#S,00%33P$1C%#M2,8-1*'28Q# ;,)W%9#]"#18-#J%(%89#$"#=6>>`?"#A)2+%-,)13#M1')8%00I#7#]%B#R8C)%-'%8*#'8#A,(3'+# $1*'0D1+*'28"#7#[4'*%#A1/%)#2D#*4%#75%)'+18#<,-C%0#7002+'1*'28"#4**/I..1j1"8+0+" -8'",0.4*-2+0.7<7[4'*%A1/%)fP6dP>`"/-D#" !F"#_2**D)%-0289#:"E"9#]%1)3%B9#;"["9#G1j1W19#$"$"9#18-#S2+419#E"O"#=6>>`?"#e2&# :),C#@)%1*5%8*#E2,)*0#[2)WI#78#7813B0'0#2D#O%-'1*2)0"b#<2,)813#2D#S%0%1)+4#'8# E)'5%#18-#:%3'8g,%8+B#XI#LI#LPLFQ#S2005189#$";"9#S25189#<"]"9#k&%'C9#<"O"9#S%5P /%39#O"9#18-#J'8-g,'0*9#E"e"#=%-0"?"#@4%#O,3*'P$'*%#7-,3*#:),C#E2,)*#^H13,1*'28"# [104'8C*289#:EI#@4%#Y)(18#R80*'*,*%Q#@B3%)#18-#e,2#=6>>6?9#N/#E'*" !d"#M1)3%B9#^"<"9#<%80%89#^"9#18-#S%5/%39#O"#=6>!L?"#R5/)2H'8C#E2,)*)225# E255,8'+1*'28I#7#M'%3-#@%0*#'8#O'3&1,W%%"#G%&#i2)W9#GiI#E%8*%)#D2)#E2,)*# R882H1*'28" !`"#M)1Z%)9#O"$"#=6>>d?"#@4%#R5/1+*#2D#*4%#E255,8'*B#E2,)*#O2-%3#28#:%D%8-18*# A%)+%/*'280#2D#M1')8%00I#7#E10%#$*,-B#1*#*4%#S%-#e22W#E255,8'*B#<,0*'+%#E%8*%)"# G%&#i2)W9#GiI#E%8*%)#D2)#E2,)*#R882H1*'28Q#J%%9#E"_"9#E4%%05189#M"9#S2**5189#:"9# $&18%)9#S"9#J15(0289#$"e"9#S%5/%39#O"9#18-#E,)*'09#S"#=6>!L?"#7#E25/)%4%80'H%#

Center for Court Innovation 520 Eighth Avenue, 18th Floor New York, New York 10018 646 386 4462 info@courtinnovation.org www.courtinnovation.org
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`?"#a@4%#G%K,0#(%*&%%8#:),C#18-#73+2423#@)%1*5%8*# A)2C)15#R8*%C)'*B#18-#:),C#E2,)*#^DD%+*'H%8%00I#A23'+B#S%+255%8-1*'280#D2)# A,)0,'8C#$,++%00"b#E)'5'813#<,0*'+%#A23'+B#S%H'%&#!cI#LI#66dP6XF" F"#J2&%8W15/9#E"@"#18-#J1*%0019#^"<"#=6>>X?"#aY8-%)0*18-'8C#*4%#S'0W#A)'8+'/3%I# e2&#18-#[4B#E2))%+*'2813#R8*%)H%8*'280#E18#e1)5#J2&PS'0W#NDD%8-%)0"b#@2/'+0# '8#E255,8'*B#E2))%+*'280"#[104'8C*289#:EI#G1*'2813#R80*'*,*%#2D#E2))%+*'280Q# J2&%8W15/9#E"@"9#J1*%0019#^"<"9#18-#e230'8C%)9#7"O"#=6>>d?"#a@4%#S'0W#A)'8+'/3%# '8#7+*'28I#[41*#e1H%#[%#J%1)8%-#D)25#!L9d`d#NDD%8-%)0#18-#f`#E2))%+*'2813# A)2C)150Tb#E)'5%#V#:%3'8g,%8+B#F6I#``Pf6" d"#$%%#G1*'2813#R80*'*,*%#2D#E2))%+*'2809#4**/I..8'+'+"C2H.*X+Q#ND! #+%#2D#<,0*'+%# A)2C)150 #E)'5%$23,*'280"C2H9#&&&"+)'5%023,*'280"C2H.A)2C)15:%*1'30" 10/KTR:U6X6Q#18-#$7O$e7 0#G1*'2813#S%C'0*)B#2D#^H'-%8+%P(10%-#A)2C)150#18-# A)1+*'+%09#&&&"8)%//"015401"C2H.h'%&R8*%)H%8*'28"10/KT'-ULX"# `"#$%%#78-)%&09#:"7"9#18-#;28*19#<"#=6>!>?"#N/#E'*Q#18-#J'/0%B9#O"["9#J18-%8P (%)C%)9#G"7"9#18-#['30289#$"<"#=6>>`?"#^DD%+*0#2D#E2C8'*'H%P;%41H'2)13#A)2C)150#

^H13,1*'28#2D#*4%#S%-#e22W#E255,8'*B#<,0*'+%#E%8*%)I#7#E255,8'*B#E2,)*#_)2&0# '8#;)22W3B8"#['33'150(,)C9#h7I#G1*'2813#E%8*%)#D2)#$*1*%#E2,)*0Q#S2005189#%*#13"#

=6>!!?"#N/#E'*" !c"#E'008%)9#7";"#18-#M1)23%9#:"<"#=6>>f?"#7H2'-'8C#M1'3,)%0#2D#R5/3%5%8*1*'28I# J%00280#D)25#A)2+%00#^H13,1*'280"#[104'8C*289#:EI#;,)%1,#2D#<,0*'+%#700'0*18+%Q# J'8-g,'0*9#E"9#e1)-'0289#<"9#18-#J1**'52)%9#A"#=6>>X?"#a@4%#S%%8*)B#E2,)*#R8'*'1P *'H%I#E2,)*P(10%-#$*)1*%C'%0#D2)#O181C'8C#S%3%10%-#A)'028%)0"b#<,0*'+%#S%0%1)+4# 18-#A23'+B#dI#!I##f`P!!c"#$&18%)9#S"#18-#]2489#<"9#&'*4#S%5/%39#O"9#E15/(%339#O"9# <1DD%9#A"9#18-#[23D%9#:"#=6>!!?"#@4%#Y"$"#7**2)8%B#_%8%)13 0#:%D%8-'8C#E4'3-422-# R8'*'1*'H%I#M2)51*'H%#^H13,1*'28#2D#*4%#A410%#R#:%5280*)1*'28#A)2C)15"#G%&# i2)W9#GiI#E%8*%)#D2)#E2,)*#R882H1*'28" !f"#J1()'2319#O"#&'*4#_23-9#^"#18-#]2489#<"#=6>!L?"#R882H1*'28#'8#*4%#E)'5'813# <,0*'+%#$B0*%5I#7#G1*'2813#$,)H%B#2D#E)'5'813#<,0*'+%#J%1-%)0"#G%&#i2)W9#GiI# E%8*%)#D2)#E2,)*#R882H1*'28" 6>"#E1)%B9#$"O"9#O1+W3'89#<"S"9#18-#M'8'C189#O"["#=6>!6?"#[41*#[2)W0T#@4%#@%8# ]%B#E25/28%8*0#2D#:),C#E2,)*I#S%0%1)+4P;10%-#;%0*#A)1+*'+%0"#:),C#E2,)*#S%H'%&# hRRRI#!I#dPX6Q#E'008%)#%*#13"#=6>!L?9#N/#E'*"