The Day Custody Program: Second Year Report

Home , Day

CJA NEW YORK CITY CRIMINAL JUSTICE AGENCY, INC.

Jerome E. McElroy Executive Director

THE DAY CUSTODY PROGRAM: SECOND YEAR REPORT

Project Director Freda F. Solomon, Ph.D. Senior Research Fellow

FINAL REPORT August 2008

52 Duane Street, New York, NY 10007 (646) 213-2500

THE DAY CUSTODY PROGRAM: SECOND YEAR REPORT

Project Director Freda F. Solomon, Ph.D. Senior Research Fellow

Project Co-Director Elyse J. Revere Research Analyst

Project Programmer: Geraldine Staehs-Goirn Information Systems’ Department Programmer/Analyst

Project Staff: Steve Mardenfeld Senior Research Assistant Raymond Caligiure Graphics and Production Specialist Annie Su Administrative Associate

FINAL REPORT August 2008

ACKNOWLEDGEMENTS

This report could not have been done without the assistance of many at CJA, and without the cooperation of the Center for Alternative Sentences and Employment Services (CASES), under the leadership of Joel Copperman its CEO and President. I am fortunate to have Elyse Revere as a collaborator on this project. Since early in the project she assumed day-to-day management of our program data collection. She was assisted in quality-control tasks by Ray Caligiure and Steve Mardenfeld. In addition, Elyse Revere provided invaluable assistance in the creation of the data sets used in this and the previous year’s study, performed much of the data analysis, and contributed to the writing of this report. Steve Mardenfeld provided additional assistance in the analysis of the recidivism data new to this year’s report. We are indebted to Geraldine Staehs-Goirn, the project’s programmer, for her skillful programming and understanding of the research process. Thanks also are due to Frank Sergi, CJA’s Director of Program Planning, and members of the Data Entry Unit under his supervision, for their help in resolving CJA database issues. My thanks also go to my colleague Mary Phillips for her suggestions regarding creating matched comparison groups, which were incorporated into the recidivism analysis, and to Jerry McElroy, CJA’s Executive Director, and my colleague Richard Peterson, for their review of an earlier draft of this report. We are greatly appreciative of the efforts of many at CASES who worked to make program information and data available to us. Special thanks are due to Elida Fahie, Director of Court Programs, Tonya C. Harvell, the program’s Executive Assistant, Quinn Cushing, Information Services Director, and Shana McMahan, Director of Strategic Planning and Analysis, as well as the court and program staff on who we ultimately rely for collecting data. In addition, Geraldine Ferrara, CJA’s counsel, along with CASES’ counsel Ellen Fried, successfully concluded the data-sharing agreement necessary to make this inter-organization collaboration possible. The seeds of the Day Custody Program developed from the work of the Discharge Planning Collaboration, an initiative begun by Martin Horn, NYC Commissioner of Correction, and Linda Gibbs, then Commissioner of NYC’s Department of Homeless Services. The Day Custody Program is made possible through the cooperation of the New York City Department of Correction, and by the commitment to data-driven policy, and openness to innovation, on the part of the City’s Office of the Criminal Justice Coordinator. While there were many who assisted in a myriad of ways, I alone am responsible for the design of this research project and the contents of this report.

Freda F. Solomon, Ph.D. Senior Research Fellow & Project Director

TABLE OF CONTENTS

INTRODUCTION ...... 1

PART ONE: THE DAY CUSTODY PROGRAM THE DAY CUSTODY PROGRAM ...... 4 Introduction ...... 4 Program Eligibility ...... 5 Data Sources, Data Collection and Information Sharing...... 8 PART TWO: THE PROGRAM’S SECOND YEAR SECOND YEAR DAY CUSTODY SENTENCED CASES...... 10 Characteristics of Defendants in the DCP-Sentenced Cases ...... 10 CJA Recommendation and Community Ties Characteristics...... 13 Arrest Crime Types and Severities for DCP-Sentenced Cases ...... 15 Charge Composition at Criminal Court Arraignment...... 17 Measuring Program Success: Retention...... 18 Time from Sentencing to Program Intake ...... 20 Setting Jail Lengths for Program Failure...... 22 Penalties for Program Failure ...... 26 SECOND YEAR DAY CUSTODY REJECTED CASES ...... 31 Rejection Source and Reason ...... 31 Characteristics of Defendants in DCP-Screened and Rejected Cases...... 33 CJA Recommendation and Community Ties Characteristics...... 35 Arrest Crime Types and Severities for Rejected Cases...... 37 Non-DCP Sentenced Case and Defendant Characteristics by Source of Rejection...... 38 Court Outcomes for Rejected Cases ...... 43 Arraignment Charge Composition for Rejected Cases Convicted at Arraignment ..45 Arraignment Decision Making for Non-DCP Sentenced Cases by Rejection Source ...... 46

RE-ARRESTS OF DEFENDANTS IN DCP-SENTENCED CASES ...... 51 RE-ARRESTS OF DEFENDANTS WITHIN FIVE MONTHS OF THE FIRST (OR ONLY) DCP-SENTENCED CASE ...... 53 Prevalence of Recidivism...... 53 Time to First Prosecuted Re-arrest...... 54 Type and Severity of Prosecuted Charge on the First Re-arrest ...... 55 Spotlight Targeting on First Prosecuted Re-arrests ...... 59 Criminal Court Decision Making in First Prosecuted Re-arrest Cases...... 60 Sentencing for Conviction in the First Re-arrest Case: Jail, Length of Jail...... 61 Frequency of Re-arrests ...... 63 RE-ARRESTS OF DEFENDANTS IN ALL SECOND-YEAR DCP-SENTENCED CASES...... 65 Prevalence of Recidivism...... 65 Time to a Prosecuted Re-arrest within Five Months of the Compliance Date for Each DCP-Sentenced Case ...... 67 Spotlight Targeting on First New Arrests ...... 68 PART THREE: COMPARING RECIDIVIM BETWEEN DAY CUSTODY SENTENCED CASES AND DEFENDANTS AND THOSE IN A PRE-PROGRAM PERIOD COMPARING RECIDIVISM BETWEEN DCP-SENTENCED CASES AND DEFENDANTS IN THE FIRST TWO PROGRAM YEARS WITH THOSE IN A PRE-PROGRAM PERIOD ...... 70 Creating Pre-Program Comparison Groups...... 71 Re-arrests Among Pre-Program and DCP-Sentenced Cases and Defendants ...... 73 Times to First Re-Arrests for Defendants in Pre-Program and DCP-Sentenced Cases...... 76 Frequency of Re-arrests ...... 78 COMPARING SUCCESSFUL PROGRAM CASES AND DEFENDANTS WITH MATCHED COMPARISON GROUPS ...... 81 Creating Matched Comparison Groups...... 83 Testing for Differences in Recidivism Between Categorical Groups of Successful Defendants in Each Program Year and Defendants in Matched Comparison Group Cases...... 86 Prevalence of Re-arrest Between Matched Comparison Groups of Pre-Program and Successful DCP-Sentenced Defendants ...... 87 Time to First Re-Arrest Between Successful DCP-Sentenced and Matched Comparison Group Defendants ...... 88 Frequency of Re-arrests Between Successful DCP-Sentenced and Matched Comparison Group Defendants ...... 89 PART FOUR: SUMMARY SUMMARY AND CONCLUSIONS ...... 92 APPENDIX CJA’S TYPOLOGY OF OFFENSES...... 103

INTRODUCTION

On September 22, 2005, the Center for Alternative Sentencing and Employment Services (CASES) began accepting defendants into its Day Custody Program (DCP). This program, created in partnership with the New York City Department of Correction (DOC), is designed as an alternative to traditional incarceration for defendants with at least three previous misdemeanor convictions arraigned in non-Spotlight-targeted misdemeanor cases in the downtown Manhattan Criminal Court. The DCP sentence is a three-day, eight-hour daily, program consisting of community service, comprehensive needs assessment and post-release referrals, and some on-site intervention programs. All sentence activities are conducted during daytime weekday hours with the defendant in the custodial setting of a DOC facility adjacent to the downtown Manhattan Criminal Court. However, defendants are released from DOC custody at the end of each program day and must return on their own volition for each successive program day. Defendants sentenced to the program, almost always upon a plea agreement at arraignment, have a ten-day window in which to successfully complete the three-day program requirement. At the time defendants receive the program sentence the court stipulates a jail sentence, usually of ten days, that is to be imposed as the penalty for failure to satisfactorily complete the program. As part of the program’s development and implementation, the New York City Criminal Justice Agency (CJA) was directed by the Office of the Criminal Justice Coordinator to provide a series of research-related items. CJA is a not- for-profit organization, working under a contract with the City of New York, whose principal operational mission is to provide pretrial services to defendants prosecuted for crimes in New York City’s adult criminal courts. As part of its pretrial services function the Agency maintains a computerized database consisting of information about defendants and the court processing and outcomes of their cases. This information is also used by CJA to perform research, and to provide criminal justice information and data on issues of 2

interest to City agencies, non-profit organizations, and academic-based researchers as appropriate. As part of its Day Custody Program research agenda, CJA previously prepared for the City an analysis of misdemeanor cases prosecuted in the downtown Manhattan Criminal Court in a several month period prior to the program’s implementation. The resulting report, THE DAY CUSTODY PROGRAM: A BASELINE CASES ANALYSIS (November 2005), examined the prospective volume and characteristics of cases and defendants most likely to be eligible for DCP. A second report, THE DAY CUSTODY PROGRAM: FY06 STATUS REPORT (June 2006), described research-related activities completed during the program’s start-up period, including the collection of program data and the building of data sets for use by CJA to report on DCP activities. This report, shared with our research partners at CASES and DOC, also provided an analysis of cases and defendants, both those DCP-sentenced and those in cases screened but not sentenced to the program, during the program’s early implementation. At the end of the first complete program year CJA prepared a first-year evaluation study (June 2007, Final Report December 2007), describing the characteristics of both DCP-sentenced and program-rejected cases and defendants, measuring successful completion rates and penalties in cases of defendants who failed to complete the program and were returned to court for re- sentencing, and a preliminary examination of recidivism rates among program- sentenced defendants.1 The current report builds on this work to assess DCP activities in its second full year of operation. Part One of this report provides a program overview, beginning with a description of the Day Custody Program and the data used in the analysis of program sentenced and rejected cases. Part Two is devoted to descriptions of various aspects of the program’s second year of operation. This begins with a section focused on cases with a DCP sentence, and a separate one focused on

1 Freda F. Solomon, THE DAY CUSTODY PROGRAM: FIRST YEAR REPORT: Final Report, (New York City Criminal Justice Agency, December 2007), available on the Agency web site, www.nycja.org . 3

cases screened and rejected, over the course of the program’s second full year of operation. Another section examines re-arrests among successful and unsuccessful second-year program-sentenced defendants and cases. Where appropriate, some comparisons are made with first program year cases. However, there almost always are changes that occur during the start up of a program, including procedures for program implementation and in data collection. As a result, the data for the two program years, both for sentenced cases, but more so for screened but not program-sentenced cases, are not entirely comparable. Part Three is new to the second year’s program report, and contains sections relating to recidivism of defendants in DCP-sentenced cases, from both the first and the second program years, in comparison with defendants and cases sharing similar characteristics from a pre-program period. These new sections provide comparisons in regard to the prevalence, time to, and frequency of new prosecuted arrests. The selected cases from the pre-program period create a baseline from which to analyze whether there were changes in recidivism patterns among program-sentenced cases. To create comparison groups for this recidivism analysis downtown Manhattan cases sharing similar defendant and charge characteristics, with arraignment convictions and jail sentences imposed greater than time served, were identified from calendar year 2004. These data were first used to compare aspects of recidivism with those in all DCP-sentenced cases in each of the program’s first two years. A separate section of the analysis refines the comparisons by using matched comparison groups to test differences in the prevalence, time to, and frequency of new arrests between successful program clients (the treatment group) and their pre-program counterparts (the control group), within different categories of eligibility for program participation,. The final part of the report contains a summary of the report’s findings. 4

THE DAY CUSTODY PROGRAM

Introduction

The Day Custody Program is designed as an alternative sentencing option in cases of recidivist misdemeanor defendants, held in custody after their arrest, and prosecuted in the downtown Manhattan Criminal Court for misdemeanor crimes. The DCP sentence is intended for use in cases of jail bound defendants convicted at arraignment in which both the case and defendant fit a series of eligibility criteria. The program sentence combines the elements of accountability such as would occur for defendants with a non-custodial community service sentence, early intervention programs, and discharge planning services that refer defendants to community-based and government services based on a needs assessment at intake. What makes the Day Custody Program unique is that all of these sentencing components are provided while the defendant is in custody only during daytime weekday hours in a Department of Correction facility adjacent to the downtown Manhattan courthouse, without using full-time incarceration. This avoids the need for DOC to transport defendants between the court and jail, conduct its normal lengthy screening processes at intake, and house defendants. In addition, it allows DCP activities to be provided without the normal interruptions that would occur in a traditional jail setting where prisoner movement and other institutional requirements circumscribe the amount of time each day in which defendants can participate in activities. Program activities are provided over the course of three days, which is about equivalent to, or only several days fewer than, the actual number of days defendants would serve post-arraignment for a ten-day jail sentence after accounting for pretrial detention time, and good-time credit which is equal to one day for each three sentence days. The program sentence is officially recorded as an intermittent jail sentence, with the stipulation of a jail sentence, usually of ten days, for failure to complete the DCP sentence. Defendants have a ten-day window in which to complete the three days’ of program activities. A court

compliance date is scheduled fifteen days after the sentencing at which time CASES’ staff report to the court whether or not the defendant successfully completed the program. Program failure results in the issuance of a bench warrant for the defendant in the case. In order to identify prospective program participants, program staff review the court papers of misdemeanor cases in advance of arraignment. Program court staff then seek to pursue further those defendants who appear to meet the program’s eligibility criteria. However, in some instances DCP court staff may be directed to screen for intake defendants in cases that may not meet the requisite criteria. Program Eligibility Under the program protocols, CASES court staff screen misdemeanor court papers of defendants awaiting arraignment at the downtown Manhattan Criminal Court. This first step is to determine if defendants are “paper eligible.” Defendants are considered paper eligible if they have at least three prior misdemeanor convictions, are not awaiting arraignment in a Spotlight-program designated case, have no history of violent crimes, and do not have an active hold for a warrant. Although neither homelessness nor being on parole or probation is an absolute exclusion to participation, these factors can affect the initial or subsequent screening for program eligibility. Defendants in each screened case must be able to be sentenced to the program immediately. Defendants who are subject to being held at the end of the arraignment court appearance, for out-of-county warrants or for violating parole conditions, cannot receive the DCP sentence in most instances. Consent by the Department of Parole is necessary before DCP can be offered to a defendant under its supervision, and an attempt to create a system for a quick response from parole officials has so far not been successful. In addition, getting a release from the Department of Parole is further handicapped when a prospective program client is screened during evening and weekend arraignments. 6

In the latter part of the program’s first year probation was no longer included as an exclusionary factor by DCP court staff, although there continue to be instances in which defense attorneys or defendants reject the program (or any arraignment plea) out of concern that a new conviction could jeopardize the probation. While homelessness was not originally envisioned as a bar to participation, early experience indicated that many such defendants, at least the street homeless or those without regular contact with a specific shelter facility, were poor risks for successful program completion. As a result, acceptance of otherwise eligible homeless defendants is considered on a case-by-case basis. The second screening stage requires review and consent by all parties— defense attorney, defendant, judge, assistant district attorney (ADA), and DCP court staff. For defendants in cases the program wishes to pursue further, the first step is to seek permission from the defense attorney to interview the defendant, so that defense attorneys serve as gate keepers to defendants. They are most likely to allow the CASES’ court staff to interview clients when they believe that the DCP sentence will be a better alternative than the anticipated or offered sentence by the prosecutor or judge. The next step before conducting the defendant interview is to check with the prosecuting attorney (ADA) to determine if there is an intent to seek jail time (greater than time served) for the defendant, as the program court staff will not conduct the interview for non-jail bound defendants. During the course of such an interview DCP court staff may find that the defendant is not likely to be a successful program participant. For example, because program activities are only during daytime weekday hours, defendants without any stable residential or other community ties, or who are found for other reasons to be poor program risks, may be rejected by DCP staff. Defendants also must consent to the program sentence. Among courtroom actors, judges influence the pool of the program-eligible population because they must be amenable to using alternative sentences. Further, judges have the power to seek a DCP sentence in a case otherwise 7 deemed ineligible by the program protocols, or to reject a proposed DCP sentence even when other participants are willing. Spotlight-targeted cases are a distinctive group, consisting of recidivist defendants who have been arrested for misdemeanor crimes during active offending cycles. Defendants in Spotlight-identified cases must have been arrested at least two previous times in the past twelve months, with one such arrest having a misdemeanor top arrest charge, and the defendant also must already have a minimum of two prior cases with convictions to misdemeanor crimes, one of which must have been within the previous twelve months. Defendants meeting these criteria are identified by the New York State Division of Criminal Justice Services (DCJS) upon review of the criminal record matching fingerprints sent by the New York City Police Department to DCJS at the time of a misdemeanor arrest. The rap sheets for defendants meeting the program criteria carry a printed banner with a “persistent misdemeanant” alert, which in turn is supposed to result in court papers for the case carrying a Spotlight stamp, although there are instances when the court papers are not appropriately marked. The goal of the Spotlight program is to encourage the courts to interrupt these offending cycles through jail sanctions for conviction, and for longer periods than would otherwise be used. As part of the conditions of funding the Day Custody Program, the City stipulated that defendants in Spotlight-targeted cases, even if otherwise DCP eligible, were not to be actively pursued. For these and other reasons, defendants in DCP-sentenced cases are a selective group from among all defendants with a minimum of three prior misdemeanor convictions arraigned in the downtown Manhattan Criminal Court. The characteristics of DCP-sentenced defendants and cases, and those screened but rejected from program participation, are examined in detail in the next sections of this report. It also is important to recognize that eligibility for the program attaches to the case and defendant characteristics at the time court papers are reviewed for cases pending arraignment for a misdemeanor charge. As a result, DCP court 8

staff may review different cases involving the same defendant over time, with program eligibility determined independently for each case. Because the same defendant may be screened multiple times but in different cases, the number of screened cases is greater than the number of screened defendants.

Data Sources, Data Collection and Information Sharing After the program’s implementation, CJA and CASES entered into a data– sharing agreement for the exchange of information about the cases and defendants sentenced to DCP, and about those screened but not sentenced to the program. As a result of that agreement CJA and CASES created a format for the regular transfer to CJA of identifiers of program-sentenced defendants and cases, and their program completion status, at the time of the scheduled court compliance review appearance. CJA staff review these cases to ensure that case and defendant identifiers match court records in the CJA database, and that the sentence information transmitted to CJA from the Office of Court Administration correctly records the DCP intermittent sentence. Problem cases are identified and sent back to CASES for correction. CASES also shares with CJA information about screened cases of defendants not sentenced to the program. Information about defendants, and the arrest and court activity for all DCP screened cases, both those sentenced and those rejected, are extracted from the CJA database. Defendant demographic and community ties information are based on information collected by CJA as part of its pre-arraignment interview process with defendants held for Criminal Court arraignment. As part of this process CJA staff record the number of prior cases (if any) in which the most serious conviction charge was for a felony crime, cases with the most serious conviction charge of misdemeanor severity, and open cases pending in the adult court system at the time of the interviewed arrest. In instances where CJA did not conduct an interview, or did not have access to the defendant’s criminal history (i.e. “rap sheet”), for a DCP-sentenced case, the CJA database, supplemented by rap sheet information supplied by the New York State Division of Criminal Justice Services (DCJS), have been used to create summary 9 conviction counts at the time of the DCP-sentenced case when possible. The same procedures are used to review instances in which it appears that the defendants did not have the requisite number of prior misdemeanor convictions to be program eligible. Lastly, Operation Spotlight designated cases are identified from data files routinely provided to CJA by DCJS, which are used by CJA as part of a separate research agenda. These files also are used to confirm cases identified but rejected by DCP because of their Spotlight designation. 10

SECOND YEAR DAY CUSTODY SENTENCED CASES

This section of the report describes DCP-sentenced cases from October 1, 2006 through September 30, 2007, the program’s second year of operation. During this time a total of 496 cases arraigned in the downtown Manhattan Criminal Court received a sentence to the Day Custody Program. The number of DCP sentenced cases was somewhat fewer than the first year’s total of 548 cases, even taking into account the additional ten cases that were sentenced to DCP in September 2005 which slightly extended the first year’s study period. There were a total of 485 defendants in the second year’s program- sentenced cases, with 11 defendants each receiving a DCP sentence in two separate cases. Among the 485 program-sentenced defendants were 93 who also had other screened but rejected cases during the second year. In addition, 54 of these 93 defendants were program-sentenced in Spotlight-targeted cases. The data that follow describe the characteristics of defendants in the DCP- sentenced cases, which in some respects may have been different at the time of each arrest for the 11 defendants with two separate DCP-sentenced cases.

Characteristics of Defendants in the DCP-Sentenced Cases

In the 496 second-year DCP-sentenced cases, most had defendants who were male and 35 years of age or older, with a mean (mathematical average) of about 40 years of age. These demographic characteristics are similar to those of the defendants sentenced in the first program year. DEFENDANT DEMOGRAPHIC AND ZIP CODE CHARACTERISTICS SEX N % ETHNICITY N % Male 436 87.9 Black 330 66.8 Female 60 12.1 Hispanic 123 24.9 Total 496 100.0 White 39 7.9 Other 2 0.4 Total 494 100.0 Unknown 2 AGE GROUP N % ZIP CODE N % 16-19 5 1.0 Manhattan 244 56.7 20-24 31 6.3 Bronx 77 17.9 25-29 50 10.1 Brooklyn 77 17.9

30-34 53 10.7 Queens 20 4.7 35-39 87 17.5 Staten Island 5 1.2 40-44 104 21.0 Other New York 5 1.2 45-49 99 20.0 New Jersey 2 0.5 50+ 67 13.5 Total 430 100.0 Total 496 100.0 Other/Unknown 66 Mean Age 39.69 Only a small percentage of program-sentenced cases had defendants identified as White (7.9%), with approximately two-thirds of the cases having non-Hispanic Black defendants, and in another quarter of the cases defendants identified themselves as Hispanic. In comparison with the first program year, the proportion of cases with defendants identified as Black was four percentage points higher, while the proportion self identifying as Hispanic was 2.7 percentage points lower. However, these differences may be accounted for by a smaller percentage of second-year cases for which ethnicity was not known. Of the cases in which defendants did not report being homeless, and for which home address was known, over half had addresses with Manhattan zip codes. There were some differences between the first two program years in the zip code locations of home addresses. The proportion of defendants in second- year DCP-sentenced cases reporting that they lived in Manhattan was five percentage points greater than in the first program year’s sentenced cases. In addition, in the second year of DCP the boroughs of Bronx and Brooklyn had the exact same proportion (17.9%) of cases of DCP-sentenced defendants, whereas in the first program year the proportion of defendants’ cases with Bronx addresses was 5.6 percentage points greater than those with Brooklyn addresses. Criminal conviction history, as shown in the next table, is characterized by the most severe conviction charge in prior adult-court cases. Most commonly, second-year DCP-sentenced cases had defendants with prior criminal convictions for both felony and misdemeanor crimes. There were no second year DCP-sentenced cases in which defendants had a prior criminal history consisting only of felony convictions, as compared to the first year when there were four such cases. 12

PRIOR CRIMINAL CONVICTION HISTORY TYPE OF CRIMINAL N % RECORD Misdemeanor Only 160 32.3 Felony Only 0 0 Both 335 67.7 Total 495 100.0 Missing 1

In all but the one DCP-sentenced case with missing data, defendants had some prior criminal conviction history, and almost all had a total of at least three such criminal convictions. However, there were seven cases in which it appeared that the defendants did not have the minimum program criteria of at least three prior cases with convictions to misdemeanor crimes. These seven cases represent a smaller proportion (1.4%) of cases lacking the minimum program conviction criteria compared to the first year of DCP sentenced cases in which almost five percent (4.9%) of defendants had fewer than three prior misdemeanor convictions.

COUNTS OF PRIOR MISDEMEANOR AND FELONY CONVICTIONS PRIOR PRIOR FELONY MISDEMEANOR N % N % CONVICTIONS CONVICTIONS Less than 3 7 1.4 None 160 32.3 3-5 109 22.0 1 110 22.2 6-9 105 21.2 2 101 20.4 10-14 107 21.6 3 65 13.1 15-19 52 10.5 4 38 7.7 20-24 43 8.7 5 15 3.0 25-29 25 5.1 6 6 1.2 30-34 18 3.6 7 0 0.0 35-39 11 2.2 8 0 0.0 40-44 6 1.2 10 0 0.0 45 or more 12 2.4 Total 495 100.0 Total 495 99.9 Missing 1 Missing 1 Overall, in the DCP-sentenced cases 89 percent of the defendants had more than the minimum criteria of three prior convictions for misdemeanor crimes, with a mean of 14.0 prior misdemeanor convictions for those with at least one prior misdemeanor conviction (data not shown). Over two-thirds of the 13

defendants in the DCP-sentenced cases also previously had been convicted of felony crimes (335 out of 495), with an average of 2.3 felony convictions per defendant in cases with felony convictions. These data are similar to the figures for prior convictions found for the DCP-sentenced cases in year one.

CJA Recommendation and Community Ties Characteristics

As a group, defendants in the DCP-sentenced cases consistently in both program years were rated by CJA to be poor risks for recognizance release if their cases were not disposed at arraignment. CJA’s recommendation classification is based on a point scale that assesses the risk of failure to appear (FTA), and includes factors such as warrant history. In the point scale system used by CJA since June 30, 2003, prior FTA for cases without an active bench warrant is included in the risk assessment, and it is the single most heavily weighted factor against recognizance release in the point scale’s indicators of risk of FTA. A recidivist defendant who has previously failed to appear can never score well enough to be given CJA’s unqualified, low-risk, recommendation, although such a defendant can receive a moderate-risk recommendation. Defendants with an outstanding active bench warrant are classified in a separate not recommended category. Because criminal record is an essential element of the recommendation system, no recommendation is made in cases of defendants for which a rap sheet is not available. These cases, along with cases with incomplete interviews, appear in the “no recommendation” category in the table that follows.

CJA RELEASE RECOMMENDATION RECOMMENDATION N % CATEGORY Recommended: Low Risk 7 1.5 Moderate Risk 21 4.5 Not Recommended: High Risk 387 82.3 Bench Warrant 22 4.7 Other Exclusionary 4 0.9 Categories 14

*No Recommendation 29 6.2 Total 470 100.0 Not Interviewed 26 n/a *The No Recommendation category includes cases in which no rap sheet was available, and circumstances in which the interview process could not be completed, often because the defendant declined or was unable to complete the interview.

In the course of its pre-arraignment interview CJA staff also collect information that serve as indicators of the strength of defendants’ ties to the community, including whether or not at the time of an interviewed arrest defendants are engaged on a full-time basis in work, school or other program activities, alone or in combination. Another item is whether or not defendants live alone or with others. CJA interviewers seek to verify the responses to these questions by telephone contact with third parties supplied by the defendant. For these items the table that follows shows all yes or no responses, with the verified and not verified responses combined. The final item, defendant responses to the question of whether they expected someone at arraignment cannot be verified, and therefore is based solely on whether the defendant answers this question with a yes or no. Percentage calculations are based only on yes and no answers (including verified responses as applicable) to each item. The numbers of cases for which there was an unresolved conflict between defendant and third-party verifier responses, for which no response to the item was available, or a “don’t know” response was given in regard to expecting someone at arraignment, are shown below the total. As seen in the table on the next page, three-quarters of the defendants in the DCP-sentenced cases reported that they were not engaged full time in work, school and/or program activities. One-sixth (16.9%) expected someone to attend their arraignment court appearance, although almost half (48.5%) reported living with others. There were differences in the proportion of yes and no answers to these items in DCP-sentenced cases between the two program years. The proportion of defendants in DCP-sentenced cases who reported that they were engaged full 15

time in work, school and/or program activities decreased to 25.1 percent in year two in comparison with 35.6 percent in year one. There also appears to be a change across the two years in whether or not defendants in DCP-sentenced cases live alone. In the second program year’s cases almost half (48.5%) of the defendants reported that they lived with others while in the first year defendants in 31.2 percent of sentenced cases reported that they lived with others. The most dramatic shift involved whether or not defendants in the DCP sentenced cases expected someone to appear at their arraignment. In year two the percentage of defendants in year two that expected someone at arraignment was 16.9 percent, while in year one 45.6 percent of defendants responded yes to this item. This is a surprising decline for which there is no easily explanation considering that a greater proportion of defendants in year two DCP-sentenced cases indicated that they lived with others.

CJA COMMUNITY TIES INTERVIEW ITEMS INTERVIEW ITEM N % Fulltime Employment/School: Yes 110 25.1 No 328 74.9 Total 438 100.0 Unresolved Conflict or N/A 58 Lives With Others: Yes 216 48.5 No 229 51.5 Total 445 100.0 Unresolved Conflict or N/A 51 Expects Someone at Arraignment: Yes 76 16.9 No 374 83.1 Total 450 100.0 Don’t Know or N/A 46

Arrest Crime Types and Severities for DCP-Sentenced Cases

Drugs, property crimes (mostly petit larceny charges), and fraud category crimes (mostly theft-of-services and forgery-related charges) were the most frequent types of offenses for which program-sentenced defendants in the second (and first) year’s cases were arrested. Because the program criteria only requires that the most serious prosecution charge at arraignment be of 16 misdemeanor severity, some cases had felony-severity crimes as the top arrest charge. However, overwhelmingly the top arrest charge was of misdemeanor severity. The table that follows displays arrest crime types using a CJA-created typology. A number of the categories contain charges defined in only one or a limited number of articles in the New York State Penal Law (PL), such as the weapon (PL 265) and drug (PL 220 and PL 221) categories. Other categories draw charges from an assortment of Penal Law articles. In addition, although all categories may contain charges with severities ranging from felony to misdemeanor, some categories such as “unknown/other” almost exclusively contain offenses of misdemeanor or lesser severity, while other categories such as harm-to-persons or persons-and-property contain a number of the most serious felony-severity crimes. A more detailed description of the charges in these crime categories, shown in the left-side column, can be found in the APPENDIX to this report.

MOST SEVERE ARREST CHARGE CHARACTERISTICS TOP ARREST TOP ARREST CHARGE N % CHARGE N % CJA CRIME SEVERITY CATEGORY Harm to 4 0.8 B-Felony 16 3.2 Persons Harm to Person & 1 0.2 C-Felony 5 1.0 Property Weapon 9 1.8 D-Felony 37 7.5 Property 167 33.7 E-Felony 20 4.0 Drugs 179 36.1 Subtotal Felony 78 15.7 Sex 3 0.6 A-Misdemeanor 373 75.2 Fraud 52 10.5 B-Misdemeanor 41 8.3 Misconduct 39 7.9 U-Misdemeanor 3 0.6 Obstructing Subtotal 10 2.0 417 84.1 Justice Misdemeanor Other 32 6.5 Other 1 0.2 Total 496 100.1 Total 496 100.0

Charge Composition at Criminal Court Arraignment

Because the DCP sentence only is available for cases pending arraignment for misdemeanor crimes, all of the cases with felony severity top arrest charges were prosecuted for less severe charges.2 Sometimes this resulted in cases being prosecuted for secondary arrest charges which were not necessarily in the same crime category as the top arrest charge. In other instances the arraignment charge was for a less severe offense within the same crime category. The table below shows the configuration of the prosecuted crime types by severity at Criminal Court arraignment for the DCP-sentenced cases. For the purposes of this display, only charges in the property, drug, fraud, and misconduct categories are shown separately, with all other crime types combined.

TOP PROSECUTED CRIME TYPE BY CHARGE SEVERITY AT CRIMINAL COURT ARRAIGNMENT FOR DCP-SENTENCED CASES ARRAIGNMENT CHARGE SEVERITY Lesser A- CRIME TYPE B-Misdemeanor Severity or Total* Misdemeanor Unknown N % N % N % N % Property 177 38.6 0 0.0 0 0 177 35.7 Drugs 124 27.1 21 72.4 1 11.1 146 29.4 Fraud 46 10.0 0 0.0 0 0 46 9.3 Misconduct 76 16.6 5 17.2 0 0 81 16.3 All other 9 2.0 3 10.3 0 0 12 2.4 Unknown 26 5.7 0 0.0 8 88.9 34 6.9 Total 458 100.0 29 99.9 9 100.0 496 100.0 (92.3%) (5.8%) (1.8%) (100.0%) *The total includes the amended charge severity for the one felony affidavit case (fraud type).

At arraignment the most substantial change from arrest was that the percentage of cases in the property crime category at arraignment, (almost 90% a petit larceny charge), increased and became over six percentage points greater than drug category cases overall at arraignment. At arrest the greatest percentage of cases was in the drug category, and this category’s cases were

2 The one case with a felony severity top arrest charge and felony affidavit charge was sentenced at arraignment on an amended charge of misdemeanor severity. 18

2.4 percentage points higher than the property crime category which was second largest. In addition, over 90 percent of all the DCP-sentenced cases were arraigned on a charge of A-misdemeanor severity, the most severe misdemeanor classification. Because almost all arraignment convictions were to the same charge severity, and almost always the same charge, this table also describes with very few exceptions the charge and severity composition of the cases for which the DCP sentence was imposed. Overall, the proportion of DCP sentenced cases that were arraigned on charges from the property crime type category was almost the same in year one and year two. There were some shifts from year one to year two in the other crime type categories at arraignment for DCP-sentenced cases. The proportion of sentenced cases arraigned on charges from the drug crime type category decreased from 34.7 percent in year one, to 29.4 percent in year two. Inversely, the proportion of sentenced cases arraigned on charges from the misconduct crime type category increased from 12.6 percent in year one, to 16.3 percent in year two.

Measuring Program Success: Retention

DCP offers an alternative to a traditional jail sentence. The program includes an accountability component through work activity at the program site, comparable to “community service,” a range of on-site services such as motivational videos and a drug-treatment readiness program, and referrals to additional services as part of a discharge plan. The ability to successfully provide this range of activities is dependent on program retention which requires defendants to complete the full three days of the schedule without being held in a custodial setting on a continuous basis. Depending of the day of the week, and time of day of the DCP-imposed sentence, defendants either are brought directly to the program to complete the first day, or are released from court and directed to report to the program at a later time. Therefore, the key measure of program success is the program’s completion rate. Of the 496 DCP-sentenced cases, a total of 394 (79.4%) were reported to have successfully completed the program sentence. (This is a slight decrease 19

from the first program year when 81.2 percent of cases had successful completions.) In the second program year this includes seven cases in which defendants successfully completed the program after being given a second chance. On the table that follows, success and failure rates are shown separately for the cases in which defendants appeared to meet the program’s eligibility criteria, and for two categories of cases in which it appears that the DCP sentence was imposed in instances in which defendants fell outside of the program’s parameters. The ‘Insufficient Priors’ category contains six of the seven cases of defendants without the minimum three prior misdemeanor convictions previously discussed. The ‘Spotlight’ category contains 183 cases that CJA identified as having been designated by DCJS as Spotlight program cases, including the one case of a defendant without the minimum three prior misdemeanor convictions, although this case had two prior misdemeanor convictions, making it Spotlight eligible. Some of these cases may have appeared program eligible because court papers were not properly identified with the Spotlight program stamp, while in other instances the program sentence most likely was solicited or imposed even though CASES court staff correctly identified the cases as within the Spotlight exclusion. The reasons for program failure—either because the defendant failed to successfully complete all three program days, or did not appear at all—are shown at the bottom of the table.

SUCCESSFUL AND UNSUCCESSFUL PROGRAM COMPLETION BY ELIGIBILITY CRITERIA Insufficient PROGRAM Eligible Spotlight Total Priors OUTCOME N % N % N % N % Success 259 84.4 5 83.3 130 71.0 394 79.4 Failure 48 15.6 1 16.7 53 29.0 102 20.6 Total Cases* 307 (61.9) 6 (1.2) 183 (36.9) 496 100.0 Did Not Complete 22 0 28 50 No Show 26 1 25 52 *Percentages in parentheses are the percent of all sentenced cases, not within category percentages.

Program eligible cases, which represented 61.9 percent of all DCP- sentenced cases, had an 84.4 percent successful completion rate. The successful completion rate also was high (83.3%) for the small number of cases in which defendants had fewer than the requisite three misdemeanor convictions. The successful completion rate was lowest for the Spotlight-targeted cases (71.0%). These are the same patterns found in the first program year, with only small differences in the successful completion rates within the three categories. As was noted for the first program year, it is not clear how much importance to attach to the success and failure rates for the two categories of ineligible cases because they represent such a small proportion of all such cases. For example, during the twelve months of the second program year there was in total 8,035 Spotlight cases arraigned in the downtown Manhattan Criminal Court, so that the 183 DCP-sentenced Spotlight cases represent only a tiny percentage (2.3%) of such cases. Similarly there would have been many times more non-Spotlight cases of defendants with fewer than three prior misdemeanor convictions than the seven program-sentenced cases.

Time from Sentencing to Program Intake

New in the second year DCP report is the addition of information on the number of days that it took from the time defendants were sentenced to the time that they appeared at the program for intake. This permits an examination of the extent to which immediately beginning the program might influence the likelihood of successful completion. There are several considerations that may affect the length of time to intake including the hour of the day of sentencing, the day of the week of sentencing, holidays, and other factors that would prevent a person from appearing at DCP on the same day as sentencing even if they would choose to begin the program immediately. The table below depicts the days to intake by eligibility category for successful completion cases. Overall, slightly less than one-third (30.7%) of successful completion cases involved a participant who appeared for intake on the same day as sentencing. The largest proportion overall was for cases where there was one day from sentencing to intake (44.2%). Successful cases from the 21

Spotlight category took the shortest amount of time from sentencing to intake, with a combined 81.3 percent appearing for intake on the day of sentencing, or one day later.

DAYS TO INTAKE BY ELIGIBILITY CRITERIA FOR SUCCESSFUL CASES* Insufficient Eligible Spotlight Total Days to intake Priors N % N % N % N % 0 days (sentence date) 78 30.7 38 29.7 3 60.0 119 30.7 1 day 104 40.9 66 51.6 1 20.0 171 44.2 2 days 11 4.3 9 7.0 1 20.0 21 5.4 3 days 37 14.6 12 9.4 0 0.0 49 12.7 4 days 11 4.3 2 1.6 0 0.0 13 3.4 5 or more days 13 5.1 1 0.8 0 0.0 14 3.6 Total 254 99.9 128 100.1 5 100.0 387 100.0 *Excludes the seven reinstated cases.

Unsuccessful cases involve defendants who failed either because they did not complete the three days of the program or because they never showed up at all to DCP. As was seen on a previous table, 52 of the 102 unsuccessful cases (or slightly over half), had defendants who never showed up and therefore had no length of time to intake because they never appeared at the program. Of the remaining 50 unsuccessful cases, there were none from the insufficient priors category, and the table below shows the distribution of days to intake for the eligible and the Spotlight cases.

DAYS TO INTAKE BY ELIGIBILITY CRITERIA FOR UNSUCCESSFUL CASES THAT DID NOT COMPLETE THE PROGRAM Eligible Spotlight Total Days to intake N % N % N % 0 days (sentence date) 10 45.5 11 39.3 21 42.0 1 day 5 22.7 8 28.6 13 26.0 2 days 3 13.6 6 21.4 9 18.0 3 days 3 13.6 0 0.0 3 6.0 4 days 1 4.5 1 3.6 2 4.0 5 or more days 0 0.0 2 7.1 2 4.0 Total 22 99.9 28 100.0 50 100.0

Compared to the successful cases, an overall higher proportion of the unsuccessful completion cases appeared for intake on the day of sentencing 22

(42.0% versus 30.7%). Over another one-quarter (26.0%) of cases that were unsuccessful involved defendants who had one day between sentencing and DCP intake. Almost the exact same proportion of unsuccessful eligible and Spotlight cases had defendants who appeared for intake on the day of sentencing, or one day later (68.2% and 67.9% respectively). When defendants in successful and unsuccessful cases are compared separately, these data suggest that for those who at least start the program, success does not appear to depend on whether or not program intake commences immediately after sentencing. However, it is not possible to determine whether there would have been greater success among the defendants in the ‘no show’ cases had they at least experienced an introduction to the program on the day of sentencing. A different perspective emerges if one combines all the cases of defendants who began the program on the day of sentencing with those that did not (including the no shows). As displayed on the table below, of the 140 cases in which defendants began the program on the sentencing date, 85 percent successfully completed the program. In comparison, of the 349 cases in which defendants were scheduled to begin the program at a later date there was a lower successful completion rate (76.8%).

SUCCESSFUL AND UNSUCCESSFUL CASES BY WHETHER PROGRAM COMMENCED ON DAY OF SENTENCING Successful* Unsuccessful Total Days to intake N % N % N % 0 days (sentence date) 119 85.0 21 15.0 140 100.0 1 or more days, or no show 268 76.8 81 23.2 349 100.0 Total 387 79.1 102 20.9 489 100.0 *Excludes the seven reinstated case with successful completion.

Setting Jail Lengths for Program Failure

In the course of program planning the expected jail sentence length to be imposed for program failure was 10 days, although judges could choose to set any statutory permissible alternative up to the maximum 365 days for conviction to a crime of A- misdemeanor severity. For general database quality-control CJA routinely receives a copy of courtroom calendars. For the purposes of DCP this 23

allowed us to compare the court system’s computerized records with original calendar entries when anomalies were suspected. In examining information about sentencing practices after the first program year, CJA identified some implementation issues in regard to the recording of the program sentence by various court personnel. First, not all courtroom personnel understood that the Day Custody Program required marking on the court calendar (docket) page the sentence as intermittent imprisonment; that the jail length to be recorded was an alternative to the three program days and not a sentence of three days; and, that attention needed to be paid to whether judges were imposing something other than the expected 10 jail alternative days. Second, that court personnel entering the courtroom data into the Office of Court Administration’s (OCA) computerized database (CRIMS) did not always correctly input the program’s sentencing information from the calendar entries. In occasional checking of court information for first-year sentenced program cases we found some instances in which courtroom staff recorded a conditional discharge sentence with no jail time instead of the intermittent jail sentence, or the intermittent jail sentence length was reported as the three program days rather than a longer jail alternative. We also found that almost all intermittent sentences showed a 10-day jail alternative even when courtroom observation and the court calendars indicated a different sentence length. In the course of further checking it appeared that some court clerks had simply written “DCP” on the court docket calendar which in turn was entered into the OCA computerized information system as an intermittent sentence with a 10-day jail sentence. After the first year’s findings CASES staff sought the court’s assistance in improving the recording of the DCP sentence. To assist CJA to better monitor these issues, in the second program year CJA requested that CASES’ courtroom staff attempt to record the judicially- imposed alternative jail sentence length. This also was designed as a mechanism for assisting CASES in determining the extent to which the court’s personnel were more frequently correctly recording the DCP program sentence. 24

We further requested that CASES courtroom staff attempt to record the District Attorney’s sentence length offer in the case to examine the extent to which DA offers might be influencing judicial decisions to impose jail sentence lengths other than 10 days. For the second year’s DCP-sentenced cases we first compared the consistency between the information in the OCA/CRIMS computerized database with the sentence lengths reported by CASES’ courtroom staff. Our research again found some discrepancies, as shown on the table below, when comparing the sentence length recorded by CASES courtroom personnel with the sentence reported in the court’s computerized database. According to OCA, a total of 450 of the 496 DCP-sentenced cases (90.7%) had a 10-day jail length, in comparison to the total of 314 cases (63.3%) recorded by CASES in the 10-day category. In a total of seven cases OCA showed either no intermittent program sentence or a sentence length of either three days or time served (the less-than-10-day category). However, when the both the CASES and court sentence length fell into categorical groups (e.g. 15-30 days), the sentence length reported was almost always identical, (except for three cases in the 15-30 day OCA reported category, but reported as ten days by CASES).

CASES AND COURT REPORTING OF DCP JAIL ALTERNATIVE LENGTHS Court (OCA-CRIMS) Reported Jail Alternative Sentence CASES’ Reported Jail Less than 10 15-30 45-60 90-180 365 Alternative Sentence Total 10 days days days days days days 10 days 5 306 3 0 0 0 314 15-30 days 2 67 17 0 0 1 87 45-60 days 0 30 0 10 0 0 40 90-180 days 0 32 0 0 5 0 37 365 days 0 15 0 0 0 3 18 Total 7 450 20 10 5 4 496

We also found that the longer than 10-day sentences reported by CASES’ more often were associated with cases in which the DA’s offices was seeking a longer jail sentence (data not shown). However, there remains an issue of the extent to which the sentence lengths of greater than 10 days in the CASES’ data 25

reflect the actual judicially imposed sentence or the DA offer, which would require comparison with the written notation on court papers. Although the CASES’ reported alternative jail lengths may not be completely correct, we believe that they are more accurate than the court’s computerized information. As a result, we chose the CASES’ reported jail lengths to test the extent to which the use of greater than 10-day jail alternatives was associated with program eligibility criteria, especially in regard to setting alternative jail sentence lengths in Spotlight-targeted cases. As the table below indicates, the percentage of Spotlight cases with greater than a 10-day jail alternative sentence (37.7%) is only slightly larger in comparison with all sentenced cases (36.7%) or with the eligible cases category (35.8%). There are too few cases in the insufficient priors category to analyze from this table.

CASES’ RECORDED ALTERNATIVE JAIL SENTENCE LENGTH BY ELIGIBILITY CRITERIA FOR PROGRAM-SENTENCED CASES CASES All Sentenced Insufficient Eligible Spotlight Alternative Jail Cases Priors Sentence N % N N N % N % 10 days 314 63.3 197 64.2 114 62.3 3 50.0 More than 10 182 36.7 110 35.8 69 37.7 3 50.0 days Total 496 100.0 307 100.0 183 100.0 6 100.0

The analysis of jail alternative lengths does show that in some majority of the DCP-sentenced cases a 10-day jail alternative for program failure was imposed by the court. However, the actual size of this majority, and the percentage and reasons for selecting longer jail alternative sentences remain unclear. There are a variety of consequences when the Day Custody Program sentence is not correctly recorded by the court system. For example, the incorrect computerized court information could affect the sanction imposed if defendants are returned on a warrant issue in program-failure cases. The OCA computerized data also is the information is reported to the New York State Division of Criminal Justice Services (DCJS), which maintains the statewide criminal history (rap sheet) information system. In addition, it also is the source 26

of the information that is electronically transferred and incorporated into CJA’s main computer database.

Penalties for Program Failure

There were a total of 102 cases in which defendants were unsuccessfully terminated from the program, and for most of these cases a warrant was issued when the failure to successfully complete the program was reported to the court on the scheduled compliance date. As of May 30, 2008, defendants in 87 (85.3%) of these cases had been brought back to court and received a new disposition on the DCP-sentenced case. This is a somewhat smaller percentage in comparison with the 91.3 percent of unsuccessful cases that were returned to court and re-sentenced, but the cut-off period in the first program year’s study was two weeks longer, so that the data are not entirely comparable. As shown in the next table, of the 15 unsuccessfully terminated cases without a court return for the program failure, nine continued to show the outstanding warrant as of May 30th, and there was no record in the CJA database of these defendants having been returned to court on the DCP case, either solely on the warrant, or as a result of a new arrest. However, in one of these cases the defendant had been re-arrested and sentenced to five days prior to the issuance of the warrant on the DCP case. In another case where there remains an outstanding bench warrant, the defendant was re-arrested on another case outside of Manhattan. The six remaining cases in the “other status” category are ones where the judge decided to satisfy the requirements of the DCP sentence by entering a “Community Service Completed” or “Execution of Sentence” disposition even though the program records showed that the defendants had not successfully completed the program. 27

STATUS OF UNSUCCESSFUL DCP-SENTENCED CASES AS OF MAY 30, 2008 Case Status N % Re-sentenced 87 85.3 Dismissed 0 0.0 Outstanding Bench Warrant 9 8.8 Other Status 6 5.9 Total 102 100.0

The DCP program was designed with an expectation of a ten-day jail sentence as the penalty for defendants who failed to complete the three days of program activity. However, as previously discussed, in some cases the court imposed a different alternative jail sentence length which complicates assessing the extent to which the penalty imposed for re-arrest matches the original plea agreement. It further appears that re-sentencing practices varied among judges, and also by whether the re-sentence in the DCP case occurred at the same time as a sentence in a re-arrest case was concurrently being imposed. This latter factor further complicates determining the consequences of program failure on defendants because the penalty imposed at re-sentencing in the DCP case may not fully reflect other actions being taken at the same time against the defendant. In total, of the 87 cases that were re-sentenced, 84 received an imprisonment sentence (96.6%). The remaining three cases were re-sentenced to a conditional discharge, two of which also included a condition of community service.

TYPE OF RE-SENTENCE IMPOSED IN UNSUCCESSFUL CASES Type of Re-sentence N % Imprisonment 84 96.6 Conditional Discharge 3 3.4 Total 87 100.0 Of the 87 re-sentenced cases, 67 had at least one re-arrest which occurred before (or on the day of) the re-sentence on the original DCP sentenced case. For the remaining 20 cases the defendant was returned on the DCP case warrant without a new prosecuted arrest. Ten of the 20 cases (50.0%) without a 28

re-arrest received a 10-day re-sentence. Among the other ten cases, one received a 5-day jail re-sentence, three a 7-day, three a 15-day, and one received a 30-day re-sentence; the other two cases received a conditional discharge sentence which included a community-service requirement. The re-sentencing in the DCP case appeared to have occurred independently from the re-arrest case in 32 of the 67 cases which had a re-arrest before the re-sentence on the original DCP case. In 20 of these cases (62.5%) the defendants received jail sentences of ten or more days on the re-sentence. Of these 20 re-sentenced cases: 17 received a 10-day sentence, one 20 days, and two were re-sentenced to 30 days. In the 12 remaining cases with a re- sentence independent of the sentence on the re-arrest case, three defendants received a 7-day sentence, eight defendants received a time-served sentence, and one was re-sentenced to a conditional discharge. For two of the 67 re-sentenced cases with a new arrest before the re- sentence on the DCP case, the re-arrests were dismissed. In both cases the defendant was arrested in a borough other than Manhattan and both re-arrests had two dockets. In both cases one of the dockets was disposed with a transfer on a criminal warrant, presumably because of the warrant issued in Manhattan on the DCP case. In one of the two cases the other docket was disposed with a dismissal, and in the second case the other docket was disposed with an ACD. The remaining 35 cases with a re-arrest before the re-sentence on the original DCP case were re-sentenced at the same time as the sentencing on the re-arrest case. In five cases the DCP re-sentence was for time served, but each received a jail sentence on the concurrent case. One of the five concurrent cases received a 10-day sentence, one received a 30-day sentence, one a 45- day sentence, one a 90-day sentence, and one a 120-day sentence on the concurrent case. In the remaining 30 cases re-sentenced concurrently with a re- arrest case: 17 received 10-day re-sentences, five received a 20-day re- sentence, and one each a re-sentence of 7 and 15 days, respectively. Three cases received 30-day re-sentences, one a 60-day sentence, one a 90-day 29 sentence, and there was one case in which a 180-day jail sentence appeared to have been imposed to satisfy both the DCP failure and the re-arrest. As illustrated in the table below, over half of the re-sentenced jail cases received a 10-day jail sentence, almost one-quarter (22.6%) more than ten days, and the remaining quarter of cases received a new sentence of fewer than ten days, including time-served sentences. Overall, seventy-five percent of all of the re-sentences with imprisonment received a jail sentence of ten or more days, when the 10-day and more than 10-day categories are combined.

LENGTH OF JAIL TIME (in days) FOR RE-SENTENCES OF IMPRISONMENT Length of Jail Sentence N % Time Served 13 15.5 5 days 1 1.2 7 7 8.3 Subtotal less than 10 days 21 25.0 10 days 44 52.4 15 days 4 4.8 20 6 7.1 30 6 7.1 60 1 1.2 90 1 1.2 180 1 1.2 Subtotal more than 10 days 19 22.6 Total 84 100.0

These data suggest that defendants were held accountable for program failure when returned to court, and in the majority of cases received a jail sentence of at least ten days. Eighteen of the 84 cases that were re-sentenced to incarceration had no intervening re-arrest, and in only four of these cases did the defendant receive a jail sentence of more than ten days. In 31 of the 84 cases re-sentenced to incarceration, the jail sentence imposed for the DCP failure appeared to have been independent of any new arrest, although defendants in these cases had a re-arrest which may have influenced the judge who was re-sentencing the defendant on the DCP case. In these 31 cases which had an incarcerative re-sentence independently from the DCP case, three received a re-sentence of more than ten days. 30

In the remaining 35 cases with a re-sentence of imprisonment, the re- sentence was imposed in conjunction with sentencing for the new arrest that returned the defendant to court for the DCP failure, and in one-third (34.3%) of these cases the defendant received jail sentence greater than ten days. However, in a number of these instances the jail sentence either was imposed on the re-arrest, and/or was imposed as a sentence to cover both the DCP-failure and the new case. 31

SECOND YEAR DAY CUSTODY REJECTED CASES This section examines cases that were screened but not sentenced to DCP. Unlike the previous section that examined DCP-sentenced cases through the program’s second year, this analysis describes the attributes and outcomes for screened but not DCP-sentenced cases and defendants. Information was available for 3,255 cases screened and rejected for placement in DCP over the course of the program’s second year of operation. After verifying Spotlight-identified cases, the screened pool of rejected cases consisted of 1,544 Spotlight-targeted cases of which 1,256 were rejected because of this exclusionary category and 288 were rejected for other reasons, and 1,711 rejected non-Spotlight cases. There were a total of 2,750 defendants found among the 3,255 cases, 2,341 (85.1%) defendants had only one rejected case, while 336 were screened for two rejected cases, 55 had three screened cases, 16 defendants had four screened cases, and one defendant each was screened and rejected in six and seven separate cases respectively, over the course of the second program year.

Rejection Source and Reason

The table that follows presents the rejection reasons from the computerized screening data received monthly from CASES. The first section of the table shows the various reasons for rejection by DCP program staff. In some instances the program rejection would have been based on paper eligibility screening, and other times only after interviewing the defendant. The DCP- program rejections are followed by rejections attributed to other courtroom participants. Shown separately at the bottom of the table are 28 rejected cases which were identified in the CASES data as being designated in the court papers as Spotlight targeted, but which were not targeted according to the information received by CJA from DCJS. For both the category of rejections by the DCP screeners, and for all rejections as a whole, Spotlight cases constituted the largest group of cases not pursued for program intake in the second program year: 61.5 percent of DCP 32

program rejections, and 38.6 percent of all rejections overall, respectively. The percentages attributed to DCP program and overall rejections due to the Spotlight exclusion were somewhat greater in the second program year, but in both program years the Spotlight designation had the largest percentage of rejected cases. Because Spotlight identification takes precedence over all other rejection reasons it is not possible to determine how many of the defendants in these cases would otherwise have been eligible for further consideration.

SOURCE AND REASON FOR PROGRAM REJECTION CASES’ DCP PROGRAM REJECTIONS: N % Overall % Operation Spotlight 1,256 61.5 38.6 Parole 87 4.3 2.7 Homeless/ Insufficient Community Ties 261 12.8 8.0 History of Violent Crimes 115 5.6 3.5 Other DCP Screener Rejection Reasons 117 5.7 3.6 Probation 10 0.5 0.3 Active Hold 148 7.2 4.5 Heavy Substance Abuser 23 1.1 0.7 Lives Outside NYC 13 0.6 0.4 Mental Illness (CPL 730) 13 0.6 0.4 DCP Program Rejections 2,043 100.0 62.8 Defense Attorney 698 59.0 21.4 Judge 233 19.7 7.2 Assistant District Attorney (ADA) 210 17.7 6.5 Defendant Refuses to Accept Program 43 3.6 1.3 Other Rejection Sources 1,184 100.0 36.4 Other Rejection Reason 28 100.0 0.9 Total Screened But Rejected 3,255 100.0 100.0

Among all rejections, the second largest category was the defense attorneys (21.4%), who act as the gatekeepers for their defendants and whose permission is almost always required before DCP screeners may interview their clients. This proportion increased from 15.8 percent of defense attorney rejections in the first year. Most often defense attorneys will refuse to have their clients interviewed by DCP staff because they believe defendants will be 33

receiving a time served or only a several day jail sentence, or a non-custodial sentence such as community service. On the other hand, there are situations where a defense attorney who has a client facing a short jail sentence may refer the client to DCP for screening. The third largest reason for program rejection was because the defendant was homeless or had insufficient community ties (8.0%). Although this is not an automatic exclusion from DCP, it may be more difficult for a defendant who does not have a stable home address to be able to complete the three day program. Judges were the fourth largest source of rejections (7.2%), although this was a smaller percentage than in the previous program year (11.6%). Some judges are unwilling to consider any programs as an alternative sentence and therefore will not utilize DCP, while in other instances the judge may want a different and most likely lengthier period of incarceration. Almost as many rejections (6.5%) were from the Assistant District Attorneys (ADA). This represented a combined category of different ADA-designated rejection reasons, but almost all of the ADA rejections (98.1% not shown) were because the ADA offer was for a non-jail sentence.

Characteristics of Defendants in DCP-Screened and Rejected Cases

The demographics of the defendants in the 3,255 rejected cases were similar to those in the DCP-sentenced cases in regard to age and sex. Most were male (88.5% rejected versus 87.9% sentenced) and 35 years of age or older (78.1% versus 72.0%). The ethnic identification of majorities of defendants in both the rejected (66.6%) and sentenced (66.8%) cases were categorized as non-Hispanic Black. Looking at the proportion of defendants for which home address was known, Brooklyn zip codes represented a smaller percentage rejected cases compared to sentenced cases (13.5% versus 17.9%). The distribution of cases with Manhattan and the Bronx zip codes showed greater differences between defendants in the rejected and sentenced cases (61.4% versus 56.7% and 16.5% versus 17.9%, respectively). However, the majority of defendants with known addresses in both groups’ cases gave a home address in Manhattan. In 34 addition, the exact same proportion of defendants in rejected cases in year one and year two reported a Manhattan zip code (61.4%).

CHARACTERISTICS OF DEFENDANTS IN DCP-REJECTED CASES SEX N % ETHNICITY N % Male 2,881 88.5 Black 2,158 66.6 Female 374 11.5 Hispanic 802 24.8 Total 3,255 100.0 White 240 7.4 Other 38 1.2 Total 3,238 100.0 Unknown 17 -- AGE GROUP N % ZIP CODE N % 16-19 28 0.9 Manhattan 1,413 61.4 20-24 149 4.6 Bronx 380 16.5 25-29 275 8.4 Brooklyn 310 13.5 30-34 265 8.1 Queens 113 4.9 35-39 526 16.2 Staten Island 18 0.8 40-44 745 22.9 Other New York 40 1.7 45-49 653 20.1 New Jersey 26 1.1 50+ 614 18.9 Total 2,300 100.0 Total 3,255 100.1 Other/Unknown 955 -- Mean Age 41.38

As shown in the table below, defendants in 71.6 percent of rejected cases had prior criminal convictions for misdemeanor and felony crimes, a higher percentage than was found among sentenced cases (67.7%). CJA found that defendants in a very small proportion (0.3%) of rejected cases had no prior criminal convictions at all, while all of the defendants in program-sentenced cases had at least one prior criminal conviction to either a misdemeanor or a felony, except for one case with missing data.

CRIMINAL HISTORY OF DEFENDANTS IN REJECTED CASES PRIOR CRIMINAL N % CONVICTIONS No Priors 10 0.3 Misdemeanor Only 914 28.1 Felony Only 2 0.1 Both 2,329 71.6 Total 3,255 100.0 35

For defendants in both the rejected and sentenced cases with more than the minimum criteria of three prior convictions for misdemeanor crimes, there was only a small difference between the two groups. Among the rejected cases, defendants in 90.7 percent of cases had more than the requisite three prior misdemeanor convictions, in comparison to 89.1 percent for defendants in the DCP-sentenced cases. (Data not shown)

NUMBERS OF PRIOR MISDEMEANOR AND PRIOR FELONY CONVICTIONS Prior Prior N % N % Misdemeanors Felonies Less than 3 48 1.5 None 924 28.4 3-5 675 20.7 1 733 22.5 6-9 629 19.3 2 559 17.2 10-14 577 17.7 3 488 15.0 15-19 394 12.1 4 297 9.1 20-24 238 7.3 5 153 4.7 25-29 179 5.5 6 59 1.8 30-34 139 4.3 7 27 0.8 35-39 105 3.2 8 11 0.3 40-44 62 1.9 9 3 0.1 45 or more 209 6.4 10 1 0.0 Total 3,255 99.9 Total 3,255 99.9

The mean number of misdemeanor convictions was higher among rejected cases of defendants with prior misdemeanor convictions (17.0) in comparison to those in DCP-sentenced cases (14.0), and at least some of this difference was due to a few outliers with extremely high counts. (Data not shown) The percentage of defendants in the rejected cases with prior felony convictions was higher than the percentage of defendants in the sentenced cases (71.6% and 67.7%). The mean number of felony convictions (2.6) was slightly higher for the rejected cases of defendants with prior felony convictions, compared to the sentenced cases of defendants with prior felony convictions (2.3), but the median (midpoint) was two for both groups. (Data not shown) CJA Recommendation and Community Ties Characteristics

Only small proportions of defendants in both the rejected and sentenced cases were rated as being of low or moderate risk of FTA if released in a case 36

continued at Criminal Court arraignment, and over three-quarters of defendants in both groups of cases were rated as being of high risk, (77.9% among rejected and 82.3% for sentenced cases).

CJA RELEASE RECOMMENDATION RECOMMENDATION N % CATEGORY Recommended: Low Risk 48 1.6 Moderate Risk 83 2.7 Not Recommended: High Risk 2,361 77.9 Bench Warrant 218 7.2 Other Exclusionary 22 0.7 Categories No Recommendation 300 9.9 Total 3,032 100.0 Not Interviewed 223 -- A smaller percentage of defendants in rejected cases reported being engaged full time in work, school and/or program activities in comparison to defendants in the DCP-sentenced cases (21.7% versus 25.1%). A much lower proportion of defendants in rejected cases reported living with someone (32.4%) when compared to defendants in DCP sentenced cases (48.5%). These differences most likely were due in part to homeless defendants being more likely among rejected than sentenced cases. A lower proportion of defendants in rejected cases expected someone to attend their arraignment court appearance, compared to defendants in sentenced cases (10.2% and 16.9%).

CJA COMMUNITY TIES INTERVIEW ITEMS INTERVIEW ITEM N % Fulltime Employment/School: Yes 593 21.7 No 2,145 78.3 Total 2,738 100.0 Unresolved Conflict or N/A 517 Lives With Others: Yes 890 32.4 No 1,854 67.6 37

Total 2,744 100.0 Unresolved Conflict or N/A 511 Expects Someone at Arraignment: Yes 284 10.2 No 2,490 89.8 Total 2,774 100.0 Don’t Know or N/A 481

Based on the individual interview items relating to being engaged full time in work, school and/or program activities, and whether or not the defendant lives alone, it appears that defendants in the sentenced cases had stronger community ties, compared to those in the rejected cases, according to these measures.

Arrest Crime Types and Severities for Rejected Cases

Drug charges represented the largest CJA crime category of top arrest charges for cases in both the rejected and sentenced groups, with a higher percentage for the rejected cases (40.1% versus 36.1%). The second largest category for both groups was property crime, mostly petit larceny charges, with the defendants in rejected cases having a much smaller proportion of top arrest charges in this crime category, (25.0% versus 33.7%). The third largest crime type category for both rejected and sentenced cases was fraud, almost all fare beating or forgery, with a higher percentage for rejected cases (14.0% versus 10.5%). A smaller percentage of top arrest charges for rejected cases were of felony severity (11.8%) than was found among the DCP-sentenced cases (15.7%). To at least some extent this was influenced by the large presence of Spotlight cases in the rejected group. Because Spotlight targeting only occurs when the top arrest charge is of misdemeanor severity, this reduces the proportion of cases in the rejected group, prosecuted for misdemeanors crimes, which might originally have had felony top arrest charges. Conversely, the overall percentage of rejected cases with misdemeanor severity charges was greater than for sentenced cases (87.8% versus 84.1%).

MOST SEVERE ARREST CHARGE CHARACTERISTICS FOR REJECTED CASES TOP ARREST TOP ARREST CHARGE N % CHARGE N % CJA CRIME SEVERITY CATEGORY Harm to 15 0.5 A-Felony 0 0.0 Persons Harm to Person & 8 0.2 B-Felony 66 2.0 Property Weapon 40 1.2 C-Felony 17 0.5 Property 810 25.0 D-Felony 189 5.8 Drugs 1,298 40.1 E-Felony 111 3.4 Sex 6 0.2 Subtotal Felony 383 11.8 Fraud 452 14.0 A-Misdemeanor 2,456 75.9 Misconduct 377 11.6 B-Misdemeanor 364 11.3 Obstructing 85 2.6 U-Misdemeanor 21 0.6 Justice Subtotal VTL 3 0.1 2,841 87.8 Misdemeanor Other 145 4.5 Other 11 0.3 Total 3,239 100.0 Total 3,235 99.9 Missing 16 Missing 20

Non-DCP Sentenced Case and Defendant Characteristics by Source of Rejection On the pages that follow the case and defendant characteristics are shown for the non-DCP-sentenced cases by the source of the rejection. In some instances percentage comparisons are not meaningful because of the small number of cases in the rejection category as a whole, or in individual cells for an item. Although the “other” rejection reason category is the smallest in numbers, it has a higher percentage of rejected females than in any other rejection category. In comparison with first program year rejected cases, the ADA rejection category had the largest increase in rejections of female defendants, 10.0 percent in program year two in comparison with 5.6 percent in year one, while the defendant rejection category had the largest decrease between the two program years, 7.0 percent in year one in comparison with 15.7 in year one. Among cases rejected by DCP court staff, Spotlight cases are shown separately from the other cases rejected by program staff. Defendants in 39

Spotlight-targeted cases do appear to have some distinguishing characteristics, some of which are due to the nature of that program’s requirements, such as having only misdemeanor severity top arrest charges. However, as noted previously, because Spotlight trumps all other rejection reasons, the extent to which some defendants in these cases would otherwise have been rejected by DCP court staff or by other court participants, or would have been defendants in cases of interest to the program, cannot be determined.

CHARACTERISTICS OF DEFENDANTS IN REJECTED CASES BY SOURCE OF PROGRAM REJECTION Defendant Other DCP Defense Spotlight Judge ADA Defendant Other Character- Reason Attorney istics Sex N % N % N % N % N % N % N % Male 1,098 87.4 701 89.1623 89.3 209 89.7 189 90.0 40 93.0 21 75.0 Female 158 12.6 86 10.9 75 10.7 24 10.3 21 10.0 3 7.0 7 25.0 Total 1,256 100.0 787 100.0698 100.0 233 100.0 210 100.0 43 100.0 28 100.0 Age Group N % N % N % N % N % N % N % 16-19 17 1.4 3 0.4 3 0.4 2 0.9 2 1.0 1 2.3 0 0.0 20-24 76 6.1 30 3.8 22 3.2 10 4.3 6 2.9 3 7.0 2 7.1 25-29 99 7.9 79 10.0 46 6.6 21 9.0 19 9.0 9 20.9 2 7.1 30-34 101 8.0 63 8.0 56 8.0 18 7.7 21 10.0 4 9.3 2 7.1 35-39 210 16.7 134 17.0 117 16.8 29 12.4 26 12.4 5 11.6 5 17.9 40-44 290 23.1 190 24.1 157 22.5 49 21.0 45 21.4 10 23.3 4 14.3 45-49 253 20.1 159 20.2 135 19.3 55 23.6 41 19.5 6 14.0 4 14.3 50+ 210 16.7 129 16.4 162 23.2 49 21.0 50 23.8 5 11.6 9 32.1 Total 1,256 100.0 787 100.0698 100.0 233 100.0 210 100.0 43 100.0 28 100.0

The percentage of rejected cases of defendants having only prior misdemeanor criminal conviction records was greatest for ADA rejections (33.8%), followed closely by Spotlight cases (32.0%). For all rejection categories for which the reason was known, the proportion of rejected cases without prior convictions was smaller in each category in comparison with the first program year, or in the case of Spotlight rejections, stayed the same at zero, from year one to year two. 40

CRIMINAL HISTORY BY REJECTION SOURCE Other DCP Defense CRIMINAL Spotlight Judge ADA Defendant Other CONVICTION Reason Attorney HISTORY N % N % N % N % N % N % N % No Priors 0 0.0 5 0.6 0 0.0 1 0.4 0 0.0 0 0.0 4 14.3 Prior Misd. 402 32.0 172 21.9 187 26.8 65 27.9 71 33.8 13 30.2 4 14.3 Only Prior Felony 0 0.0 1 0.1 0 0.0 0 0.0 0 0.0 0 0.0 1 3.6 Only Both 854 68.0 609 77.4 511 73.2 167 71.7 139 66.2 30 69.8 19 67.9 Total 1,256 100.0 787 100.0 698 100.0 233 100.0 210 100.0 43 100.0 28 100.0

Across all second-year rejection categories defendants overwhelmingly were rated by CJA as being of high risk for FTA if released in cases continued at arraignment. Judge rejected cases had the largest percentage of defendants rated by CJA to be of high risk of FTA (87.6%), followed by defense attorneys, other DCP reasons, and Spotlight rejected cases (86.7%, 86.6%, and 86.5% respectively). With only one exception, less than 30 percent of the cases in each rejection category had defendants who reported being engaged in full-time activities. The defendant refused DCP rejection category had the largest proportion (34.1%) of defendants engaged in full-time activities, which may be why they were unwilling to accept a program sentence if they believed that they could receive a non-incarcerative sentence without conditions. Spotlight rejected cases had the smallest percentage of defendants who reported being engaged on a full-time basis in work, school or program activities (15.4%). Defendants in the other DCP reason rejection category had the smallest proportion of defendants who reported living with others (28.1%).

CJA RECOMMENDATION AND INTERVIEW ITEMS BY REJECTION SOURCE Other DCP Defense Spotlight Judge ADA Defendant Other CJA Reason Attorney RECOMMEND- ATION N % N % N % N % N % N % N %

Recommended: Low or Moderate 32 3.2 23 3.6 47 7.6 10 4.8 10 5.5 6 14.6 3 12.5 Risk

Not Recommended: High Risk 876 86.5 560 86.6 533 86.7 183 87.6 156 85.2 33 80.5 20 83.3

Other Not 105 10.4 64 9.9 35 5.7 16 7.7 17 9.3 2 4.9 1 4.2 Recommended *Total 1,013 100.0 647 100.0615 100.0 209 100.0 183 100.0 41 100.0 24 100.0 Other DCP Defense INTERVIEW Spotlight Judge ADA Defendant Other Reason Attorney ITEMS N % N % N % N % N % N % N % Fulltime Employment/School: Yes 157 15.4 152 23.5 153 24.9 62 29.8 48 26.2 14 34.1 7 29.2 No 864 84.6 494 76.5 462 75.1 146 70.2 135 73.8 27 65.9 17 70.8 Total 1,021 100.0 646 100.0615 100.0 208 100.0 183 100.0 41 100.0 24 100.0 Unresolved Conflict or 235 141 83 25 27 2 4 N/A Lives With Others: Yes 314 30.7 182 28.1 216 35.0 83 39.7 71 38.8 15 36.6 9 37.5 No 708 69.3 466 71.9 401 65.0 126 60.3 112 61.2 26 63.4 15 62.5 Total 1,022 100.0 648 100.0617 100.0 209 100.0 183 100.0 41 100.0 24 100.0 Unresolved Conflict or 234 139 81 24 27 2 4 N/A Expects Someone at Arraignment: Yes 88 8.5 59 9.0 76 12.2 32 15.2 23 12.5 4 9.8 2 8.3 No 949 91.5 594 91.0 548 87.8 179 84.8 161 87.5 37 90.2 22 91.7 Total 1,037 100.0 653 100.0624 100.0 211 100.0 184 100.0 41 100.0 24 100.0 Don’t Know 219 134 74 22 26 2 4 or N/A *Excludes cases without a CJA recommendation.

As shown below, the drug category had the largest percentage of cases in every rejection category although the percentages varied, from 43.8 percent of the cases in the judge rejection category, to 35.7 percent of the cases in the defendant rejection category. The property crime category had the second largest percentage of cases in all of the rejection categories except ADA and defendant rejections, ranging from 29.9 percent among Spotlight rejection cases to 13.3 percent in the ADA reject category. The second largest proportion of rejections in the ADA category was for cases from the fraud crime category (23.3%), and for defendant rejections the second largest percentage of rejected 42 cases were in the misconduct crime category (19.0%). The percentage of cases in the fraud category ranged from 23.3 percent of ADA rejections to 8.6 percent of rejections in the judge category. As previously discussed, no Spotlight-targeted case can have a felony severity top arrest charge so it is not surprising to find the greatest percentage of cases with A-misdemeanor in the Spotlight rejection category. The percentage of cases with B-misdemeanor severity top arrest charges was largest among the defense attorney-rejected cases (15.5%) compared to the other rejection categories.

ARREST CHARGE CHARACTERISTICS BY REJECTION SOURCE Other DCP Defense TOP Spotlight Judge ADA Defendant Other Reason Attorney ARREST CHARGE TYPE N % N % N % N % N % N % N %

Drugs 505 40.3 301 38.4 283 41.0 102 43.8 80 38.1 15 35.7 12 42.9 Property 375 29.9 198 25.3 128 18.6 67 28.8 28 13.3 7 16.7 7 25.0 Fraud 150 12.0 112 14.3 110 15.9 20 8.6 49 23.3 7 16.7 4 14.3 Misconduct 153 12.2 76 9.7 84 12.2 19 8.2 34 16.2 8 19.0 3 10.7 Other 70 5.6 96 12.3 85 12.3 25 10.7 19 9.0 5 11.9 2 7.1 Total* 1,253 100.0 783 100.0 690 100.0 233 100.0 210 100.0 42 100.0 28 100.0 TOP ARREST CHARGE N % N % N % N % N % N % N % SEVERITY Felony 0 0.0 158 20.3 135 19.7 61 26.5 19 9.1 9 21.4 1 3.6 A 1,129 90.2 537 68.8 440 64.3 139 60.4 158 76.0 30 71.4 23 82.1 Misdemeanor B 123 9.8 73 9.4 106 15.5 27 11.7 28 13.5 3 7.1 4 14.3 Misdemeanor U 0 0.0 12 1.5 3 0.4 3 1.3 3 1.4 0 0.0 0 0.0 Misdemeanor Total** 1,252 100.0 780 100.0 684 100.0 230 100.0 208 100.0 42 100.0 28 100.0 *Total excludes cases with unidentified top arrest charges. **Total excludes cases with charges of lesser or unknown severity.

Court Outcomes for Rejected Cases The next table shows the arraignment outcome for all rejected cases in the left-side column. The right-side column shows the disposition, or last known status as of February 4, 2008, for cases not disposed at Criminal Court arraignment. (Among the 83 cases shown as not disposed are one consolidated case and two cases covered by a prosecution in a different case.) As can be seen, slightly more than two-thirds of all cases screened but not DCP-sentenced were disposed at arraignment, virtually all with a conviction.

COURT OUTCOMES FOR REJECTED CASES AT AND POST ARRAIGNMENT ARRAIGNMENT POST N % N % OUTCOME ARRAIGNMENT ACD* 18 0.6 ACD* 6 0.6 Dismissed 8 0.2 Dismissed 236 22.7 Convicted 2,189 67.3 Convicted 682 65.6 Transferred to Continued 1,040 32.0 8 0.8 Supreme Court Warrant Total 3,255 100.1 25 2.4 Ordered *An ACD is an Adjournment in Contemplation Not Disposed 83 8.0 of Dismissal. Total 1,040 100.1 Among the cases with an arraignment conviction most (84.2%) received a jail sentence as the most severe penalty imposed. Jail sentences also dominated the sentencing patterns for cases convicted at a post-arraignment court appearance. Almost all other sentences were a conditional discharge (CD) or community service. These proportions are an increase from the first program year when 78.7 percent of cases convicted at arraignment had a jail sentence as the most severe penalty, and 88.7 percent of cases convicted post arraignment had a jail sentence imposed.

MOST SEVERE SENTENCE IMPOSED AT AND POST ARRAIGNMENT AT POST N % N % ARRAIGNMENT ARRAIGNMENT Imprisonment 1,842 84.2 Imprisonment 628 93.9 Probation 1 0.0 Probation 0 0.0 CD 326 14.9 CD 35 5.2 Other Sentence 19 0.9 Other Sentence 6 0.9 Total* 2,188 100.0 Total** 669 100.0 *Total excludes one case convicted at arraignment in a treatment part and still active. 44

**Total excludes 13 cases convicted post-arraignment and not yet sentenced, including four Dismiss-Do Not Seal cases which had been in a treatment part, and nine cases convicted post- arraignment in a treatment part and still active as of the cutoff date. Over three-fifths of the cases with a jail sentence imposed for an arraignment conviction were for time less than ten days, including the time- served sentences, with the remainder receiving sentences of either ten or greater than ten days. (As will be discussed later, almost two-thirds of the cases with jail time greater than ten days were Spotlight-targeted cases.) In addition, jail time of three or fewer days, and even for some in the 4-5 day category depending on the day of the week, would be tantamount to a time-served jail sentence after crediting defendants for one day toward early release for every three jail days (including credit for pre-conviction detention time), and early scheduled weekend releases. Although a far greater proportion of the jail sentences imposed for a post-arraignment conviction were longer than those imposed at arraignment, this does not take into account how much pre-trial detention time would have been credited in these cases toward post-conviction incarceration time.

JAIL TIME (in days) FOR CASES WITH A MOST SEVERE SENTENCE OF IMPRISONMENT AT POST N % N % ARRAIGNMENT ARRAIGNMENT Time Served 728 39.5 Time Served 147 23.4 1-3 days 78 4.2 1-3 days 6 1.0 4-5 229 12.4 4-5 21 3.3 6-9 126 6.8 6-9 30 4.8 Subtotal less Subtotal less than 1,161 62.9 204 32.5 than 10 days 10 days 10 days 252 13.7 10 days 83 13.2 11-15 days 120 6.5 11-15 days 60 9.6 16-30 205 11.1 16-30 136 21.7 31-45 41 2.2 31-45 34 5.4 46-60 38 2.1 46-60 33 5.3 61-75 0 0.0 61-75 3 0.5 76-90 11 0.6 76-90 34 5.4 91-120 10 0.5 91-120 18 2.9 121-180 2 0.1 121-180 17 2.7 181-240 0 0.0 181-240 0 0.0 Over 240 days 2 0.1 Over 240 days 6 1.0 Subtotal greater Subtotal greater 429 23.3 341 54.3 than 10 days than 10 days Total 1,842 100.0 Total 628 100.0 45

Arraignment Charge Composition for Rejected Cases Convicted at Arraignment

The table below shows the configuration of the prosecuted crime types by their severities at Criminal Court arraignment for convicted cases. For the purposes of this display, only charges in the property, drug, fraud, and misconduct categories are shown separately, with all other crime types combined. With some exceptions, convictions in these cases were to charges of the same severity, and usually the same charge type, as the prosecuted charge. Because virtually all DCP sentences were imposed for conviction at Criminal Court arraignment, this permits us to compare charge characteristics for the most comparable group of rejected cases, those convicted and sentenced at arraignment. Among rejected cases there was a smaller percentage of cases arraigned on A-misdemeanor severity charges (88.6%) than was found among sentenced cases (92.3%). There were a greater number of rejected cases with unspecified charges and most of these would have been charges outside of the Penal or Vehicle and Traffic Laws which usually are of misdemeanor-equivalent severity. The largest percentage of rejected cases was found in the drug category (35.5%), smaller than the 40.4 percent of rejected drug cases in the first year of DCP. As in the first year of the program, the second largest percentage of rejected cases was in the property crime category (23.0%). This is unlike DCP- sentenced cases where the property crime category has a larger percentage of cases than the drug category. In comparison to DCP-sentenced cases, rejected cases had a larger percentage in the CJA fraud category (9.3% versus 15.9% respectively). In addition, while almost all cases in the misconduct category had criminal trespass charges, a larger percentage of such cases were for the B- misdemeanor offense among rejected cases. (Data not shown) 46

TOP PROSECUTED CRIME TYPE BY CHARGE SEVERITY AT CRIMINAL COURT ARRAIGNMENT FOR DCP-REJECTED CASES CONVICTED AT CRIMINAL COURT ARRAIGNMENT ARRAIGNMENT CHARGE SEVERITY Lesser A B CRIME TYPE Severity or Total Misdemeanor Misdemeanor Unknown* N % N % N % N % Property 503 25.9 0 0.0 0 0 503 23.0 Drugs 636 32.8 142 71.0 0 0.0 778 35.5 Fraud 345 17.8 3 1.5 0 0.0 348 15.9 Misconduct 330 17.0 52 26.0 0 0.0 382 17.5 All other 42 2.2 3 1.5 0 0.0 45 2.1 Unknown 83 4.3 0 0.0 50 100.0 133 6.1 Total 1,939 100.0 200 100.0 50 100.0 2,006 100.0 (88.6%) (9.1%) (2.3%) (100.0%) *Includes six cases with a violation charge severity. Overall, the charge composition of rejected cases convicted at arraignment is consistent with the no or very short jail sentences regularly imposed on recidivist misdemeanor defendants in these types of cases in the downtown Manhattan Criminal Court. Further, the differences in charge composition between the rejected cases and the DCP-sentenced cases strongly suggest that the DCP-sentenced cases as a group were more likely to receive comparatively longer jail sentences because of the larger percentage of charges of A- misdemeanor severity and the greater percentage of property-crime cases which have a higher likelihood of jail, and more of it.

Arraignment Decision Making for Non-DCP Sentenced Cases by Rejection Source Because the Day Custody Program is designed to provide an alternative sentence for cases with convictions at arraignment in the downtown Manhattan Criminal Court, this section exclusively focuses on court outcomes and case characteristics for DCP-rejected cases at this first court appearance by the source of program rejection. Over half of the cases in each rejection category examined, except for the defendant refusal category, were disposed at arraignment, almost always by 47

conviction, although the percentages varied with the rejection source. Overall, defendant rejection cases, (followed by judge rejections) had the comparatively smallest percentages of cases disposed at arraignment, with non-Spotlight cases rejected by the defendant being much more likely to be continued than the cases in the Spotlight rejection group (55.8% versus 29.5% for Spotlight cases). Cases refused by the ADAs had the highest rates of arraignment dispositions (91.4%), followed by the other reason category, Spotlight rejections, and cases rejected by defense attorneys.

CASE OUTCOMES AT ARRAIGNMENT IN NON-DCP-SENTENCED CASES BY REJECTION SOURCE Other DCP Defense ARRAIGN- Spotlight Judge ADA Defendant Other MENT Reason Attorney DISPOSITION N % N % N % N % N % N % N % ACD or 7 0.6 3 0.4 14 2.0 0 0.0 2 1.0 0 0.0 0 0.0 Dismissed Convicted 879 70.0 487 61.9 457 65.5 135 57.9 192 91.4 19 44.2 20 71.4 Continued 370 29.5 297 37.7 227 32.5 98 42.1 16 7.6 24 55.8 8 28.6 Total 1,256 100.1 787 100.0698 100.0 233 100.0 210 100.0 43 100.0 28 100.0

There were only a few items of note that distinguished the prosecuted charge characteristics across rejection sources for cases convicted and sentenced at arraignment, displayed in the next table. In some instances, percentage differences may not be as meaningful as they might appear because of the small number of cases in some of the groups. Most (88.6%, not shown) of the DCP-rejected cases that were convicted and sentenced at arraignment had an A-misdemeanor prosecuted charge, and in all rejection categories the percentage was over eighty percent. These are cases for which the penalty can be as great as one year’s jail time. And, although not shown in the table that follows, the conviction charge severity was almost always the same as the prosecuted charge severity. The smallest proportion of A- misdemeanor severity charges was in the defense attorney category (84.5%), followed by the ADA rejection category (84.9%). Among all rejection groups the largest percentage of cases was in the drug crime category, but this varied by rejection group. The judge rejection category, 48 followed by the defendant refusal group, had the largest percentages of cases in the drug crime category. In comparison with the composition of drug charges among the rejected categories (excluding the other rejection reason category), the ADA rejection group had a larger percentage of drug cases with marijuana crimes of B-misdemeanor severity (not shown). This largely explains the greater percentage of B-misdemeanor, and smaller percentage of A-misdemeanor, cases for the ADA rejection group in comparison to these distributions among rejection sources. The percentage of cases that fell into the property crime category, most with petit larceny charges, was largest for the judge rejection group.

ARRAIGNMENT CHARGE CHARACTERISTICS FOR CASES CONVICTED AND SENTENCED AT ARRAIGNMENT BY REJECTION SOURCE TOP Other DCP Defense Spotlight Judge ADA Defendant Other ARRAIGN- Reason Attorney MENT CHARGE N % N % N % N % N % N % N % SEVERITY A 806 91.7 432 88.9 386 84.5 116 85.9 163 84.9 18 94.7 17 85.0 Misdemeanor B 70 8.0 31 6.4 54 11.8 15 11.1 27 14.1 0 0.0 3 15.0 Misdemeanor Other/ 3 0.3 23 4.7 17 3.7 4 3.0 2 1.0 1 5.3 0 0.0 Unknown Total 879 100.0 486 100.0 457 100.0 135 100.0 192 100.0 19 100.0 20 100.0 TOP ARRAIGN N % N % N % N % N % N % N % CHARGE TYPE Drugs 301 34.2 157 32.3 176 38.5 58 43.0 69 35.9 8 42.1 8 40.0 Property 242 27.5 115 23.7 77 16.8 39 28.9 22 11.5 3 15.8 5 25.0 Fraud 135 15.4 73 15.0 76 16.6 9 6.7 48 25.0 3 15.8 4 20.0 Misconduct 156 17.7 89 18.3 78 17.1 17 12.6 39 20.3 1 5.3 2 10.0 Other 45 5.1 52 10.7 50 10.9 12 8.9 14 7.3 4 21.1 1 5.0 Total 879 99.9 486* 100.0 457 99.9 135 100.1 192 100.0 19 100.1 20 100.0 *The totals shown in the other DCP reason-rejected category excludes one case convicted at arraignment but still active as of the cutoff date. 49

A majority of rejected cases received a jail sentence for conviction at arraignment, although the percentages varied by the source or reason for the rejection. Cases rejected by defendants or by ADAs had the lowest percentages of jail-time sentences imposed for conviction. Conversely, defendants rejected for Spotlight reasons, judge-rejected cases, and the other reason category had the largest percentage of jail sentences. To some extent these differences are related to the differences in the arraignment crime type and severity among rejection groups. For example, misdemeanor cases with B-marijuana charges, which were more frequent among the ADA rejection group, are also the type of case least likely to receive jail time and the shortest sentences when jail is imposed.

SENTENCING AT ARRAIGNMENT IN NON-DCP SENTENCED CASES BY REJECTION SOURCE MOST Other DCP Defense Spotlight Judge ADA Defendant Other SEVERE Reason Attorney SENTENCE TYPE N % N % N % N % N % N % N % Imprison- 787 89.5 401 82.5 365 79.9 116 85.9 144 75.0 12 63.2 17 85.0 ment Non- Imprison- 92 10.5 85 17.5 92 20.1 19 14.1 48 25.0 7 36.8 3 15.0 ment Total 879 100.0 486* 100.0 457 100.0 135 100.0 192 100.0 19 100.0 20 100.0 *Excludes one other DCP reason rejection case convicted at arraignment but still active as of the cutoff date.

When a jail sentence was imposed at arraignment the length of the sentence varied among the rejection categories. In all categories except for judge rejections, over a majority of the cases received a jail sentence of less than ten days (including time-served sentences), ranging from 47.4 percent of the judge rejection cases to 90.3 percent among the sentences in the ADA rejection category. Cases in the defense attorney rejection group had the second largest percentage of cases with jail lengths less than ten days (80.1%). Not only did a greater percentage of cases in the Spotlight category receive jail sentences, but they also had the second largest percentage of sentences imposed at arraignment that were greater than ten days. The largest percentage 50 of sentences imposed at arraignment greater than ten days was from the judge rejection category. The very small percentages of jail sentences of ten, or greater than ten days, among the ADA and defense rejection groups are consistent with the gate-keeping function these court participants provide in determining whether DCP court staff pursue an interview with defendants who appear paper eligible for the program sentence.

JAIL TIME (in days) FOR CASES WITH A MOST SEVERE SENTENCE OF IMPRISONMENT IMPOSED FOR AN ARRAIGNMENT CONVICTION Other DCP Defense SENTENCE Spotlight Judge ADA Defendant Unknown LENGTH IN Reason Attorney DAYS N % N % N % N % N % N % N % Time Served 222 28.2 146 36.4 190 52.1 33 28.4 120 83.3 6 50.0 11 64.7 1-3 days 38 4.8 14 3.5 21 5.8 1 0.9 3 2.1 0 0.0 1 5.9 4-5 102 13.0 51 12.7 54 14.8 14 12.1 6 4.2 1 8.3 1 5.9 6-9 67 8.5 24 6.0 27 7.4 7 6.0 1 0.7 0 0.0 0 0.0 Subtotal less than 429 54.5 235 58.6 292 80.1 55 47.4 130 90.3 7 58.3 13 76.5 10 days 10 days 131 16.6 70 17.5 25 6.8 17 14.7 6 4.2 2 16.7 1 5.9 11-15 days 65 8.3 25 6.2 21 5.8 7 6.0 1 0.7 1 8.3 0 0.0 16-30 100 12.7 48 12.0 21 5.8 25 21.6 6 4.2 2 16.7 3 17.6 31-45 23 2.9 10 2.5 3 0.8 4 3.4 1 0.7 0 0.0 0 0.0 46-60 23 2.9 8 2.0 2 0.5 5 4.3 0 0.0 0 0.0 0 0.0 61-75 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 76-90 7 0.9 2 0.5 0 0.0 2 1.7 0 0.0 0 0.0 0 0.0 91-120 6 0.8 2 0.5 1 0.3 1 0.9 0 0.0 0 0.0 0 0.0 121-180 2 0.3 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 Over 180 1 0.1 1 0.2 0 0.0 0 0.0 0 0.0 0 0.0 0 0.0 days Subtotal greater 227 28.8 96 23.9 48 13.2 44 37.9 8 5.6 3 25.0 3 17.6 than 10 days Total 787 100.0 401 100.0 365 100.0 116 100.0 144 100.0 12 100.0 17 100.0

RE-ARRESTS OF DEFENDANTS IN DCP-SENTENCED CASES This is the first of several report sections designed to examine a variety of issues related to re-arrests among DCP-sentenced defendants. Although the Day Custody Program makes no representation that it can more effectively reduce recidivism among defendants who otherwise would receive a traditional short jail sentence, this is a topic of interest among policy makers. In this section we begin to address the issue of re-arrests by examining whether successful DCP-program defendants had lower rates of new arrests, and longer times to a new arrest, than unsuccessful program defendants. In addition, it describes the characteristics and court outcomes of new arrests. However, because the characteristics of successful and unsuccessful program cases and defendants are so dissimilar, these comparisons cannot be used to determine whether there is a program effect on recidivism. This is an issue addressed in the sections in the next part of this report, in which the DCP- sentenced population is compared to similar cases in a pre-program period. For the purpose of this report recidivism is defined as new prosecuted (docketed) arrests within five months of the scheduled compliance date for defendants in DCP-sentenced cases. The five month interval for each defendant in a DCP-sentenced case is used because this is the longest period possible to track new arrests and court activity following the latest DCP-sentenced cases in order to create a recidivism data set within the time frame for completing the DCP study reports. Every DCP-sentenced case is scheduled for a compliance hearing approximately fifteen days after sentencing, at which time CASES’ program staff report to the court whether or not the defendant has successfully completed the program. If defendants were arrested during the interval between the DCP- sentencing and compliance dates these arrests were not included in our definition of recidivism, but new arrests that occurred on the compliance date were included. In a few instances DCP-sentenced defendants were given a 52

second chance to complete the program in the original case. When this occurred the compliance date of the reinstatement was used in the analysis. In most instances recidivism is discussed and illustrated not only between successful and unsuccessful defendants, but also by defendants’ program- eligibility status in the DCP-sentenced case. Among the successful and unsuccessful program clients were defendants who met the program’s eligibility requirements when sentenced to the Day Custody Program. That is, the defendants had a minimum of three prior misdemeanor convictions, were prosecuted in a non-Spotlight targeted misdemeanor case, and the District Attorney’s plea offer at arraignment was for jail time greater than time served. There were a very small number of DCP-sentenced defendants in non-Spotlight targeted cases with fewer than the required three prior misdemeanor convictions. Another group contains normally ineligible defendants sentenced to the program in Spotlight-targeted cases, either at the initiative of court participants other than CASES’ court staff, or who appeared eligible to CASES’ court staff because the court papers were not properly marked with the Spotlight identifier. For this and a variety of other reasons defendants in the DCP-sentenced Spotlight cases are a particularly problematic group. For example, while in general defendants in Spotlight-targeted cases are more likely to receive jail sentences and longer sentence lengths than recidivist misdemeanor defendants in comparable but not Spotlight-targeted cases, other research suggests that at least some of the DCP-sentenced Spotlight cases might not have been jail sentenced. The analysis about recidivism in this section is presented only in descriptive terms using comparisons of percentages, or mean (mathematical average) and median (midpoint) computations. However, when defendants or cases are dividing among program-status groups, or other categorical breakdown, numbers can become so small that comparisons are not meaningful. Finally, re-arrests are examined in two different ways. In the second program year there were a total of 485 defendants sentenced to the Day Custody Program in a total of 496 cases. This includes eleven defendants each with two 53

separate DCP-sentenced cases in the second program year’s time. Most of the analysis of recidivism focuses only on re-arrests within the five months of the first or only DCP-sentenced case for each of the 485 defendants. Later there is some additional discussion of re-arrests within five months of all DCP-sentenced cases.

RE-ARRESTS OF DEFENDANTS WITHIN FIVE MONTHS OF THE FIRST (OR ONLY) DCP-SENTENCED CASE Prevalence of Recidivism

The first table shows the prevalence of at least one docketed re-arrest for 484 defendants3 within five months of the scheduled compliance date of the first (or only) DCP-sentenced case. The “number of defendants” column shows how the 484 defendants were distributed among the program eligibility categories. The next column shows the number of defendants in each category and then the total number of defendants with one or more docketed re-arrests following the compliance date of the first DCP-sentenced case, and the final column shows the percent re-arrested. For example, among all 484 defendants, a total of 255 were re-arrested within five months of the compliance date of the first (or only) DCP- sentenced case, for a 52.7% overall re-arrest rate. Among the 302 defendants who met the program’s eligibility criteria in the original DCP-sentenced case, the re-arrest rate was 47.7% in comparison with a 61.4% re-arrest rate for DCP- sentenced defendants from the Spotlight cases.

RECIDIVISM AMONG DCP-SENTENCED DEFENDANTS PROGRAM- Number of Number Percent ELIGIBILITY STATUS Defendants Re-Arrested Re-arrested Eligible 302 144 47.7 Insufficient Priors 6 3 50.0 Spotlight 176 108 61.4 Total 484 255 52.7

The next table compares the prevalence of recidivism within five months of the first DCP-sentenced case separately for successful and for unsuccessful

3 One case was excluded from the re-arrest analysis because it was reinstated late in the study period and did not have five months at risk for re-arrest. 54

defendants—those who either failed to attend or who began but did not successfully complete their DCP-program sentence—overall and by their program eligibility status at the time they were given their first DCP-sentence. It shows that there was a lower re-arrest rate among the defendants in the eligible and Spotlight categories who successfully completed the program in comparison with unsuccessful defendants, although the percentages differed. However, as previously discussed, the smaller re-arrest rates of successful program participants cannot necessarily be attributed to the Day Custody Program, but may be due other differences in defendant and case characteristics.

RECIDIVISM AMONG SUCCESSFUL AND UNSUCCESFUL DCP-SENTENCED DEFENDANTS SUCCESSFUL UNSUCCESSFUL PROGRAM Number Percent Number Percent STATUS AT Number of Number of Re- Re- Re- Re- SENTENCING Defendants Defendants arrested arrested arrested arrested Eligible 255 115 45.1 47 29 61.7 Insufficient 5 2 40.0 1 1 100.0 Priors Spotlight 125 72 57.6 51 36 70.6 Total 385 189 49.1 99 66 66.7

Time to First Prosecuted Re-arrest The next table shows the mean and median number of days to the first prosecuted re-arrest on or after the court compliance date for both the successful and unsuccessful DCP-sentenced defendants who were re-arrested within five months of the compliance date of the first DCP-sentenced case. Although the times to the first prosecuted re-arrest for the two recidivist successful defendants and the one recidivist defendant with insufficient priors are included in the total, the data for them are not shown because there are too few cases for meaningful comparisons. For the 189 recidivist successful program clients the average time to their first re-arrest was 57.17 days, with a median of 48.0 days. In comparison, for the 66 recidivist defendants among the unsuccessful program clients the average time to re-arrest was 57.45 days with a median of 52.0 days. The mean and 55

median times to first re-arrests for the successful and unsuccessful cases in the second program year differ even less than they did in the first program year.

TIME TO FIRST RE-ARREST FOR SUCCESSFUL AND UNSUCCESSFUL DCP- SENTENCED DEFENDANTS Recidivist Unsuccessful Recidivist Successful Completers PROGRAM- Completers ELIGIBILITY Median Median Mean Number Mean Number STATUS Number of Number of of Days of Days Days Days Eligible 61.83 53.0 60.03 49.0 Spotlight 48.76 35.0 52.86 52.0 *Total 57.17 48.0 57.45 52.0 *The total includes the mean and median times for the two successful and one unsuccessful recidivist defendants with insufficient priors but separate data for this program-eligibility category are not shown.

For recidivist defendants in the successful program completion group the shortest time to re-arrest occurred for three defendants who had a first re-arrest on the compliance date and the longest was for one defendant first re-arrested 149 days from the compliance date. For recidivist defendants in the unsuccessful program group there were two defendants arrested the day after the compliance date and the longest time was for one defendant re-arrested after 148 days. (Data not shown)

Type and Severity of Prosecuted Charge on the First Re-arrest

There were differences in both the types and severity distributions of charges at Criminal Court arraignment on the first prosecuted re-arrest between successful and unsuccessful defendants. Crime types are categorized using the CJA-created typology previously discussed in an earlier section, and as provided in the APPENDIX to this report. The types of prosecuted (arraignment) charges on the first re-arrest case among successful DCP clients were dispersed among a variety of crime categories, and had different distributions depending on the program-eligibility status of the defendants in the DCP-sentenced cases. (The Criminal Court arraignment charge in the first re-arrest case for each of the two recidivist 56

successful defendants with insufficient priors fell into a different crime category so the percentage allocation, 50.0%, is the same for each.) As in the first program year, overall, the property (30.2%) and drug (25.9%) crime categories have the largest percentages of Criminal Court arraignment charges among the first re-arrest cases for recidivist successful defendants, although the proportions have shifted in the two categories. These are followed in descending percentage order by charges in the fraud, misconduct and other/unknown crime categories. Whereas in the first program year the drug crime category had the largest proportion of cases (28.3% in year one and 25.9% in year two), in the second program year the property crime category had the largest proportion of cases (20.8% in year one and 30.2% in year two). For recidivist successful program defendants in the eligible group, property crimes, mostly the petit larceny charge, made up the largest percentage of Criminal Court arraignment charges for first re-arrest cases (27.8%). For the defendants in this group charges in the drug category had the second largest percentage (24.3%) of arraignment charges on the first re-arrest. The next largest percentage of charges was found in the misconduct category, most of which involved criminal trespass. The proportion of cases in the fraud category, a majority of which were fare-evasion and secondarily possession of forged instrument crimes, was smaller in the second program year (12.2%) in comparison with the first program year (17.7%) percent. For recidivist successful defendants from the Spotlight group, the property crime category had the largest percentage of arraignment charges on the first re- arrest cases (34.7%), an increase over the proportion of these cases in the first year (29.3%). The next largest category of cases was those with charges in the drug crime category (27.8%). Charges in the fraud category had the third largest percentages of cases (13.9%), followed by first re-arrest cases with arraignment charges in the misconduct category (9.7%). Among the recidivist successful program defendants in both the eligible and Spotlight groups there were a small and similar percentage of first re-arrest 57

cases with arraignment charges that fell into the harm-to-persons and persons- and-property crime categories.

CRIME TYPE OF CRIMINAL COURT ARRAIGNMENT CHARGE ON THE FIRST RE- ARREST CASE CJA CRIME Recidivist Successful Defendants CATEGORY OF Insufficient Eligible Spotlight Total ARRAIGNMENT Priors CHARGE ON 1ST RE-ARREST N % N % N % N % Harm to Persons 5 4.3 0 0.0 3 4.2 8 4.2 Persons & 1 0.9 0 0.0 0 0.0 1 0.5 Property Weapon 1 0.9 0 0.0 1 1.4 2 1.1 Property 32 27.8 0 0.0 25 34.7 57 30.2 Drug 28 24.3 1 50.0 20 27.8 49 25.9 Sex Crime 1 0.9 0 0.0 0 0.0 1 0.5 Fraud 14 12.2 0 0.0 10 13.9 24 12.7 Misconduct 15 13.0 1 50.0 7 9.7 23 12.2 Obstruct Justice 4 3.5 0 0.0 2 2.8 6 3.2 VTL 1 0.9 0 0.0 0 0.0 1 0.5 Unknown Other 13 11.3 0 0.0 4 5.6 17 9.0 Total 115 100.0 2 100.0 72 100.1 189 100.0

There were differences in the distribution among crime-type categories of the charges at Criminal Court arraignment in first re-arrest cases for recidivist unsuccessful DCP-sentenced defendants. Overall these defendants had first re- arrest cases with Criminal Court arraignment charges far more heavily concentrated in the drug (39.4%) and property (21.2%) crime categories, with about twelve percent of the remainder prosecuted for charges in the fraud category. As with the cases of recidivist successful defendants, the petit larceny charge constituted a majority of the cases in the property crime category. (The one defendant in the insufficient priors group had an arraignment charge in the fraud category.) For the unsuccessful program-eligible defendants, drug crime had a larger percentage of arraignment charges (24.1%) than charges in the property crime category (20.7%), with the second largest percentage of arraignment charges. For recidivist unsuccessful Spotlight defendants, drug crime charges (52.8%) by far had the largest percentage of arraignment charges in the first re-arrest case, 58

an increase of over 27 percentage points from this same category in year one (25.6%). The next largest category of cases was those with property crime charges (22.2%) in the first re-arrest case.

CRIME TYPE OF CRIMINAL COURT ARRAIGNMENT CHARGE ON THE FIRST RE- ARREST CASE CJA CRIME Recidivist Unsuccessful Defendants CATEGORY OF Eligible Insufficient Priors Spotlight Total ARRAIGNMENT CHARGE ON 1ST RE- N % N % N % N % ARREST Harm to 2 6.9 0 0.0 0 0.0 2 3.0 Persons Persons & 0 0.0 0 0.0 1 2.8 1 1.5 Property Weapon 2 6.9 0 0.0 0 0.0 2 3.0 Property 6 20.7 0 0.0 8 22.2 14 21.2 Drug 7 24.1 0 0.0 19 52.8 26 39.4 Sex Crime 0 0.0 0 0.0 0 0.0 0 0.0 Fraud 2 6.9 1 100.0 5 13.9 8 12.1 Misconduct 5 17.2 0 0.0 2 5.6 7 10.6 Obstruct Justice 2 6.9 0 0.0 1 2.8 3 4.5 VTL 0 0.0 0 0.0 0 0.0 0 0.0 Unknown Other 3 10.3 0 0.0 0 0.0 3 4.5 Total 29 99.9 1 100.0 36 100.1 66 99.8

Differences also were found in the distributions among the severities of the charges at Criminal Court arraignment for first re-arrest cases between successful and unsuccessful DCP defendants re-arrested with five months of the program compliance date. As shown in the table below, similar percentages of first re-arrest cases had a felony-severity arraignment charge between the successful (11.1%) and unsuccessful (9.1%) recidivist defendants, although this proportion dropped in both categories from year one to year two (15.8% successful and 15.3% unsuccessful in year one). As in the first year, there was a larger concentration of first re-arrest cases in the A-misdemeanor category for the unsuccessful (78.8%) than for the successful (74.1%) defendants. 59

SEVERITY OF CRIMINAL COURT ARRAIGNMENT CHARGE ON FIRST RE-ARREST CASE Recidivist Successful Defendants SEVERITY OF Insufficient ARRAIGNMENT Eligible Spotlight Total CHARGE Priors N % N % N % N % Felony 13 11.3 0 0.0 8 11.1 21 11.1 A- Misdemeanor 81 70.4 1 50.0 58 80.6 140 74.1 B- Misdemeanor 11 9.6 1 50.0 2 2.8 14 7.4 Unclassified 1 0.9 0 0.0 0 0.0 1 0.5 Misdemeanor Lesser/ Unknown 9 7.8 0 0.0 4 5.6 13 6.9 Total 115 100.0 2 100.0 72 100.1 189 100.0

SEVERITY OF Recidivist Unsuccessful Defendants Insufficient ARRAIGNMENT Eligible Spotlight Total CHARGE Priors N % N % N % N % Felony 3 10.3 0 0.0 3 8.3 6 9.1 A- Misdemeanor 23 79.3 1 100.0 28 77.8 52 78.8 B- Misdemeanor 1 3.4 0 0.0 4 11.1 5 7.6 Unclassified 0 0.0 0 0.0 0 0.0 0 0.0 Misdemeanor Lesser/ Unknown 2 6.9 0 0.0 1 2.8 3 4.5 Total 29 100.0 1 100.0 36 100.0 66 100.0

Spotlight Targeting on First Prosecuted Re-arrests

The percentage of first re-arrests that were Spotlight-targeted was slightly larger among recidivist unsuccessful program defendants. To some extent these differences reflect the fact that a greater percentage of defendants originally sentenced in a Spotlight-targeted case were unsuccessful in completing the DCP sentence, compared to the other two categories. Overall, among successful program completers half of the first re-arrests were Spotlight targeted (50.3%) compared to just over half (53.0%) of the first re-arrests for unsuccessful program clients. This means that the first re-arrest had a misdemeanor as the most severe arrest charge, and the defendant met both the number and timing of prior arrests and convictions required by the Spotlight targeting protocols to be designated as an active repeat offender. 60

For defendants whose original DCP sentence came from the Spotlight category, there was about a seven percentage point difference between the recidivist successful and unsuccessful completers with a Spotlight-targeted first re-arrest (70.8% versus 63.9%). This is a reversal in percentage categories from year one where the percentage of Spotlight-targeted re-arrests was smaller for unsuccessful completers in comparison with successful completers (72.8% versus 79.1%). In the second program year there was almost no difference between DCP-sentenced eligible successful and unsuccessful defendants having a Spotlight targeted re-arrest (37.4% versus 37.9%). This is in contrast to the first program year when there was a slightly greater difference (7.5 percentage points) between DCP-sentenced eligible successful and unsuccessful defendants having a Spotlight targeted re-arrest (36.9% versus 44.4%).

SPOTLIGHT PROGRAM TARGETING OF FIRST RE-ARREST CASES FOR SUCCESSFUL AND UNSUCCESFUL DEFENDANTS BY PROGRAM STATUS IN DCP-SENTENCED CASE WAS 1ST RE- Program Status for Recidivist Defendants in DCP-Sentenced Case ARREST Insufficient SPOTLIGHT Eligible Spotlight Total TARGETED? Priors Successful N % N % N % N % No 72 62.6 1 50.0 21 29.2 94 49.7 Yes 43 37.4 1 50.0 51 70.8 95 50.3 Total 115 (60.8) 2 (1.1) 72 (38.1) 189 100.0 Unsuccessful N % N % N % N % No 18 62.1 0 0.0 13 36.1 31 47.0 Yes 11 37.9 1 100.0 23 63.9 35 53.0 Total 29 (43.9) 1 (1.5) 36 (54.5) 66 100.0

Criminal Court Decision Making in First Prosecuted Re-arrest Cases

All Criminal Court dispositions, or the outcome at the latest appearance, are as reported to CJA by OCA as of April 22, 2008. First re-arrests occurred in Manhattan for over three-fourths of the successful recidivist defendants, and in over four-fifths of the first re-arrest cases for unsuccessful defendants. (Data not shown) When the first re-arrest case was prosecuted in the Bronx, a Criminal Court equivalent outcome was created for non-felony cases in order to present comparable data for all prosecuted re-arrests. This is necessary because of the 61

consolidation of the Bronx criminal court system for all post-arraignment court appearances on or after November 4, 2005. For almost all cases not disposed at the Criminal Court arraignment as of this date, the continuation of the prosecution occurs in the Criminal Division of the Bronx Supreme Court. As the table below shows, convictions were the dominant outcomes for all re-arrest cases, which is consistent with expected outcomes in non-felony cases in a recidivist misdemeanor population. Only a small percentage ended with a dismissal, and when this occurred is most often was in the eligible category. This occurrence also may be related to the nature of the charge.

CRIMINAL COURT (or equivalent) OUTCOME ON THE FIRST RE-ARREST CASE AS OF April 22, 2008 CRIMINAL Program Status for Recidivist Defendants in DCP-Sentenced Case COURT OUTCOME ON Insufficient ST Eligible Spotlight Total 1 RE- Priors ARREST Successful N % N % N % N % Conviction 94 81.7 2 100.0 67 93.1 163 86.2 Dismissal 10 8.7 0 0.0 0 0.0 10 5.3 ACD 0 0.0 0 0.0 0 0.0 0 0.0 Pending 0 0.0 0 0.0 1 1.4 1 0.5 Warrant 3 2.6 0 0.0 1 1.4 4 2.1 Transfer to 8 7.0 0 0.0 3 4.2 11 5.8 Supreme Court Total 115 100.0 2 100.0 72 100.1 189 99.9 Unsuccessful N % N % N % N % Conviction 25 86.2 1 100.0 34 94.4 60 90.9 Dismissal 2 6.9 0 0.0 0 0.0 2 3.0 ACD 0 0.0 0 0.0 0 0.0 0 0.0 Pending 0 0.0 0 0.0 0 0.0 0 0.0 Warrant 0 0.0 0 0.0 0 0.0 0 0.0 Transfer to 2 6.9 0 0.0 2 5.6 4 6.1 Supreme Court Total 29 100.0 1 100.0 36 100.0 66 100.0

Sentencing for Conviction in the First Re-arrest Case: Jail, Length of Jail

The next illustration shows the use of jail as the most severe sentence imposed upon recidivist DCP-sentenced defendants upon conviction in the first re-arrest cases for those cases that were sentenced. For re-arrest cases that go to a dedicated treatment court the original court outcome may be a conviction, 62

but the ultimate disposition of those cases may be a dismissal, as a result of compliance with the treatment court requirements. Cases with these circumstances will also affect the number of cases with no imprisonment for conviction on the next table. Among successful program defendants over four-fifths received a sentence of jail, in comparison to all but three defendants among the unsuccessful recidivist defendants. In comparison with successful recidivist defendants, there was a greater likelihood (or equal likelihood) of the imposition of a jail sentence for unsuccessful defendants in all program-status groups.

THE USE OF JAIL AS THE MOST SEVERE SENTENCE FOR CONVICTION IN THE FIRST RE-ARREST CASE JAIL FOR Program Status for Recidivist Defendants in DCP-Sentenced Case CONVICTION ST Insufficient ON 1 RE- Eligible Spotlight Total ARREST Priors Successful N % N % N % N % No 16 17.2 0 0.0 10 15.4 26 16.3 Yes 77 82.8 2 100.0 55 84.6 134 83.8 Total 93 100.0 2 100.0 65 100.0 160 100.1 Sentenced Unsuccessful N % N % N % N % No 2 8.0 0 0.0 1 2.9 3 5.0 Yes 23 92.0 1 100.0 33 97.1 57 95.0 Total 25 100.0 1 100.0 34 100.0 60 100.0 Sentenced

When a jail sentence was imposed, successful defendants were more likely to receive a time-served sentence (25.4%) than unsuccessful defendants with a jail sentence for conviction in the first re-arrest case (12.3%). However, some of the differences in sentence lengths may be accounted for by the differences in the types of charges in the first re-arrest case between successful and unsuccessful program clients. 63

LENGTH OF JAIL SENTENCES (in days) IMPOSED FOR CONVICTION IN THE FIRST RE-ARREST CASE LENGTH OF Program Status for Recidivist Defendants in DCP-Sentenced Case JAIL SENTENCE FOR 1ST RE- Eligible Insufficient Priors Spotlight Total ARREST CONVICTION Successful N % N % N % N % 3-9 days 16 20.8 1 50.0 3 5.5 20 14.9 10 13 16.9 0 0.0 10 18.2 23 17.2 12-15 7 9.1 1 50.0 9 16.4 17 12.7 20-30 7 9.1 0 0.0 10 18.2 17 12.7 35-65 5 6.5 0 0.0 8 14.5 13 9.7 90-365 4 5.2 0 0.0 6 10.9 10 7.5 Subtotal Non- 52 67.6 2 100.0 46 83.7 100 74.7 Time-Served Time Served 25 32.5 0 0.0 9 16.4 34 25.4 Total 77 100.1 2 100.0 55 100.1 134 100.0 Unsuccessful N % N % N % N % 3-9 days 1 4.3 0 0.0 4 12.1 5 8.8 10 3 13.0 1 100.0 9 27.3 13 22.8 12-15 4 17.4 0 0.0 1 3.0 5 8.8 20-30 6 26.1 0 0.0 10 30.3 16 28.1 35-65 3 13.0 0 0.0 2 6.1 5 8.8 90-365 2 8.7 0 0.0 4 12.1 6 10.5 Subtotal Non- 19 82.5 1 100.0 30 90.9 50 87.8 Time-Served Time Served 4 17.4 0 0.0 3 9.1 7 12.3 Total 23 99.9 1 100.0 33 100.0 57 100.1 Overall, these data do not suggest that greater penalties than would be expected were imposed upon re-arrest for either successful or unsuccessful defendants who previously received a DCP sentence.

Frequency of Re-arrests In addition to the prevalence of re-arrests, a separate issue is the frequency (numbers) of re-arrests between successful and unsuccessful defendants within five months of the compliance date of the first DCP-sentenced case. The next table shows the frequency of re-arrests for the successful and unsuccessful defendants with re-arrests. Overall, there were only small differences in the frequency of new prosecuted arrests between the recidivist successful and unsuccessful DCP-program sentenced defendants. 64

The 189 recidivist successful program defendants had a combined total of 355 new prosecuted arrests within five months of their original program compliance date, with a mean of 1.88 arrests and a median of two arrests. In comparison, the 66 recidivist unsuccessful DCP-sentenced defendants had a combined total of 118 re-arrests, with both a mean of 1.79 and median of 1.5 arrests. Among recidivist defendants in the program-eligible group there was virtually no difference in the frequency of re-arrests between the successful and unsuccessful program completers. The mean number of re-arrests for the 115 recidivist successful program defendants was 1.69 in comparison with 1.62 for the 29 unsuccessful defendants in the program-eligible group, and both groups had a median of 1.0. The two recidivist successful defendants in the insufficient priors groups had a combined total of four re-arrests, while the one recidivist unsuccessful defendant with insufficient priors in the DCP-sentenced case had only one re- arrest, and these numbers are simply too small for meaningful comparison. The frequency of re-arrests among the recidivist Spotlight defendants was higher for the 72 defendants who successfully completed the DCP program, with a mean of 2.18 re-arrests, in comparison with a mean of 1.94 re-arrests for the 36 unsuccessful recidivist defendants originally sentenced to DCP in a Spotlight case in year two. However, these differences still are quite small and, in addition, for both groups the median was two. This is in contrast to year one in which the frequency of re-arrests was higher for unsuccessful, in comparison to successful, re-arrested Spotlight defendants, and the median for the unsuccessful was two in comparison with one for successful Spotlight defendants. 65

FREQUENCY OF RE-ARRESTS BETWEEN SUCCESSFUL AND UNSUCCESSFUL DEFENDANTS BY THEIR PROGRAM-ELGIBILITY STATUS IN THE DCP- SENTENCED CASE PROGRAM SUCCESSFUL UNSUCCESSFUL STATUS OF Total Mean Median Total Mean Median ST 1 DCP- Number Number Number Number Number Number SENTENCED of Re- of Re- of Re- of Re- of Re- of Re- CASE arrests arrests arrests arrests arrests arrests Eligible 194 1.69 1.0 47 1.62 1.0 Insufficient 4 2.0 2.0 1 1.00 1.0 Priors Spotlight 157 2.18 2.0 70 1.94 2.0 Total 355 1.88 2.0 118 1.79 1.5

RE-ARREST OF DEFENDANTS IN ALL SECOND-YEAR DCP-SENTENCED CASES In addition to examining the prosecuted re-arrests after the first (or only) DCP-sentenced case for each of the 484 program-sentenced clients, we also examined re-arrests within five months of the compliance data for all of the 4954 DCP-sentenced cases during the second program year. The total number of cases includes 22 belonging to 11 defendants each with two DCP-sentenced cases within the first program year. Because there was very little difference in the findings about recidivism after accounting for the additional 11 cases, only selected aspects of this analysis are presented. Among all 495 cases, 393 had defendants who successfully completed the program and 102 cases had defendants reported to the court as unsuccessful. Prevalence of Recidivism

The first table shows the prevalence of at least one docketed re-arrest within five months of the scheduled compliance date for defendants in the 495 DCP-sentenced cases, shown by the program-eligibility status of the defendants when the program sentence was imposed. The second column shows how the DCP-sentenced cases were distributed among the program-eligibility categories.

4 One case was not included in the re-arrest analysis because time at risk was less than five months. 66

The next column shows the number of cases of defendants in each category with one or more docketed re-arrests, and the final column the percent re-arrested. Among all the 495 DCP-sentenced cases, a total of 262 had defendants who were re-arrested within five months of the compliance date, for a 52.9% overall re-arrest rate. Among the 307 DCP-sentenced cases in which defendants met the program’s eligibility criteria, the re-arrest rate was 47.9% in comparison with a 61.5% re-arrest rate for DCP-sentenced cases of defendants in the Spotlight group. In all instances these percentages are only fractionally different than was found when only the first re-arrest for each defendant was examined, and the number and percentage of defendants in the insufficient priors group is identical.

RECIDIVISM AMONG DEFENDANTS IN ALL SECOND-YEAR DCP-SENTENCED CASES PROGRAM- Number of Cases Percent ELIGIBILITY Number of Cases with Re-arrested Re-arrested STATUS Defendants Eligible 307 147 47.9 Insufficient Priors 6 3 50.0 Spotlight 182 112 61.5 Total 495 262 52.9

The next table compares the prevalence of recidivism in the cases of successful and unsuccessful defendants in DCP-sentenced cases, overall and by the eligibility status of the defendants in the DCP program cases. It shows that there was a lower re-arrest rate (49.4%) among cases with successful program completers in comparison with the cases of unsuccessful defendants (66.7%). This was true in every program-status category although the percentages differed. These are comparable to the findings about the first re-arrest cases for each defendant, but again this does not necessarily reflect a program impact on recidivism. 67

RE-ARRESTS OF DEFENDANTS IN SUCCESSFUL AND UNSUCCESSFUL DCP- SENTENCED CASES SUCCESSFUL UNSUCCESSFUL PROGRAM Number Number Number Number Percent Percent STATUS AT of DCP- with Re- of DCP- with Re- Re- Re- SENTENCING sentenced arrested sentenced arrested arrested arrested Cases Defendants Cases Defendants Eligible 259 118 45.6 48 29 60.4 Insufficient 5 2 40.0 1 1 100.0 Priors Spotlight 129 74 57.4 53 38 71.7 Total 393 194 49.4 102 68 66.7

Time to a Prosecuted Re-arrest within Five Months of the Compliance Date for Each DCP-Sentenced Case The next table shows the mean and median number of days to a new prosecuted arrest within five months on or after the court compliance date for each DCP-sentenced case. The data are shown for defendants in the successful and unsuccessful DCP-sentenced cases. This table shows that there was a longer average time to the first prosecuted re-arrest among the cases of successful program defendants. For the 194 cases of recidivist successful program clients the average time to a re-arrest within five months of the compliance date was 56.81 days, with a median of 46.5 days. In comparison, for the 68 cases with recidivist unsuccessful program clients the average time to a re-arrest was 56.35 days with a median of 51.0 days.

TIME TO FIRST RE-ARREST BY DEFENDANTS IN SUCCESSFUL AND UNSUCCESSFUL DCP-SENTENCED CASES PROGRAM- Successful Unsuccessful ELIGIBILITY Mean Number Mean Number Median Median STATUS of Days of Days Eligible 61.26 53.0 60.03 49.0 Spotlight 48.76 37.0 51.13 51.0 *Total 56.81 46.5 56.35 51.0 * Total includes cases of defendants with insufficient priors, but separate data for the few defendants in this program-eligibility group are not shown.

Among the cases of recidivist defendants in the successful group the shortest time to re-arrest was for three defendants with a re-arrest on the compliance date and the longest was one defendant re-arrested 149 days after the compliance date. For the cases of recidivist defendants in the unsuccessful group there were two defendants arrested one day after the compliance date and the longest time was for one defendant re-arrested after 148 days. (Data not shown)

Spotlight Targeting on First New Arrests

The percentage of first prosecuted re-arrests on or after the DCP compliance date that were Spotlight-targeted was greater among the cases of recidivist unsuccessful program completers. Among cases of re-arrested successful program completers 51.0% of the first new arrests after the DCP- sentenced compliance dates were Spotlight targeted in comparison with 54.4% of the first new arrests for re-arrested defendants in the unsuccessful category of cases. For defendants with DCP-sentenced Spotlight-targeted cases the difference between the recidivist successful and unsuccessful completers with a Spotlight- targeted first re-arrest was small (70.3% versus 65.8%). In contrast to the year one cases, in the second year the cases of recidivist unsuccessful completers in the program-eligible category had a smaller percentage of unsuccessful defendants with a Spotlight-targeted first re-arrest (37.9%) in comparison to cases of recidivist successful completers (39.0%).

SPOTLIGHT TARGETING OF FIRST PROSECUTED RE-ARREST BY PROGRAM STATUS FOR THE DCP-SENTENCED CASES OF SUCCESSFUL AND UNSUCCESSFUL DEFENDANTS WAS 1st NEW Program Status for Recidivist Defendants in DCP-Sentenced Cases ARREST Insufficient SPOTLIGHT Eligible Spotlight Total TARGETED? Priors Successful N % N % N % N % No 72 61.0 1 50.0 22 29.7 95 49.0 Yes 46 39.0 1 50.0 52 70.3 99 51.0 Total 118 (60.8) 2 (1.0) 74 (38.1) 194 100.0 69

Unsuccessful N % N % N % N % No 18 62.1 0 0.0 13 34.2 31 45.6 Yes 11 37.9 1 100.0 25 65.8 37 54.4 Total 29 (42.6) 1 (1.5) 38 (55.9) 68 100.0 70

COMPARING RECIDIVISM BETWEEN DCP-SENTENCED CASES AND DEFENDANTS IN THE FIRST TWO PROGRAM YEARS WITH THOSE IN A PRE-PROGRAM PERIOD This is the first of two report sections designed to examine how the patterns of recidivism among defendants in DCP-sentenced cases compare with cases of defendants similar to those in the program’s population. As was seen in the previous section, defendants in DCP-sentenced cases who successfully completed the program’s requirements had a lower prevalence of re-arrests than did unsuccessful defendants in DCP-sentenced cases. That analysis also indicated that among DCP-sentenced defendants with re-arrests, there was little if any difference in the frequency of re-arrests between the successful and unsuccessful program-sentenced defendants. However, defendant and case characteristics between successful and unsuccessful program cases are too dissimilar for the purposes of examining whether the program has had any affect on recidivism. This only can be done by finding cases and defendants with similar characteristics. From a methodological perspective the optimum way to create comparison groups would be by random assignment. That is, to have the court randomly assign cases to either the Day Custody Program (the treatment group), or to impose the otherwise “normal” sentence (the control group). However, this is neither a practical nor realistic approach for many reasons. In addition, experience shows that DCP, by the inclusion of Spotlight-targeted arrests and occasionally other non-eligible cases, has not been limited only to its target population. Another alternative approach might have been to use eligible cases for which the program was rejected by court participants or declined by defendants. However, CASES’ is quite successful in recruiting into DCP eligible cases so that there are too few rejected or declined jail-sentenced program- eligible cases with similar defendant, case, and court outcome characteristics to compare with program-sentenced cases. After exploring various other alternatives we chose to create comparison groups with cases from a proximate pre-program period. Although the case and defendant composition can vary from one year to another within and across

boroughs, at least in regard to Manhattan there is no evidence to suggest that DA plea policies or court outcomes for cases and defendants with similar characteristics experienced any substantial changes during the years under examination. Creating Pre-Program Comparison Groups For the purposes of creating a pool of cases from a pre-program period with characteristics most similar to those likely to be targeted by, or receive, a DCP sentence, we began by selecting from the CJA database cases that fit the following criteria: y Downtown Manhattan cases in calendar year 2004; y in which defendants were held for Criminal Court arraignment; y arraigned on an A- or B-misdemeanor severity charge, or a charge outside the Penal or Vehicle and Traffic Laws of unknown severity; y were disposed by a guilty plea at the Criminal Court arraignment; y had a jail sentence imposed at the time of the arraignment conviction of between 3 and 365 days; y in non-Spotlight cases in which defendants interviewed by CJA had a minimum of three prior misdemeanor convictions; y in non-Spotlight cases in which defendants had prior convictions but not necessarily three of misdemeanor severity; or, y were Spotlight-targeted cases with misdemeanor charges at Criminal Court arraignment. Calendar year 2004 was chosen as the latest pre-program period that would permit the selection of cases that also would have had at least a five month post-jail-release interval for a new arrest prior to the start of the Day Custody Program. In addition to misdemeanors, charges outside the Penal or Vehicle and Traffic Laws were included because they almost always are of misdemeanor-equivalent severity. The selection criteria excluded all non-jail sentenced cases. In addition, the minimum jail sentence length of at least three days was used to eliminate not only actual time-served sentences but also any sentence that would automatically be the equivalent of time served when imposed for a conviction at Criminal Court arraignment. 72

The program protocol requires non-Spotlight cases in which defendants have a minimum of three prior misdemeanor convictions. This is the ‘eligible’ program status category, and a majority of DCP-sentenced cases met these criteria in the first two program years. Spotlight-targeted cases were included for comparison purposes because, although CASES does not actively recruit defendants from this population, about a third of all DCP-sentenced cases have come either by solicitation to CASES from other courtroom participants, or because court papers have not been not properly marked as Spotlight- designated cases. There was no prior conviction criterion applied to the inclusion of Spotlight cases from 2004 that met the other selection requirements. CJA independently receives information about arrests targeted by the Spotlight program, which otherwise is not indicated either in the CJA database or in the court system’s computerized information system. In some Spotlight-targeted arrests CJA may not have recorded criminal history including the numbers of prior misdemeanor and prior felony convictions. However, in every Spotlight-targeted arrest the defendant must have at minimum at least two prior misdemeanor convictions, and an overwhelming majority of defendants in these cases have more than two. Also included in the pool of selected cases from calendar year 2004 were a small number of misdemeanor-prosecuted non-Spotlight cases in which defendants had prior criminal convictions but not the minimum three convictions to misdemeanor crimes. Especially during the first year’s start-up period there were occasional miscounts that resulted in the inclusion of a small number of cases of defendants that did not meet the program’s prior conviction criterion among DCP-sentenced cases. For each of the pre-program cases in calendar year 2004 meeting the above criteria we computed an estimated jail release date. We took the length of the jail sentence imposed at Criminal Court arraignment, uniformly converted to into days, subtracted one day’s time to account for arrest-to-arraignment detention, and then multiplied the remaining time by .6666 to account for one day’s good-time credit for each three days of the remaining jail sentence, 73

rounded to the nearest full number of days. We added the resulting computed time to the arraignment date to determine the start date from which to search for new prosecuted arrests five months forward for each case.

Re-arrests Among Pre-Program and DCP-Sentenced Cases and Defendants To determine pre-program re-arrest rates we divided the selected cases from calendar year 2004 into the three categories comparable to those of program-sentenced cases—program eligible, insufficient priors, and Spotlight. Within each group we searched for new prosecuted arrests based on the reappearance of defendant NYSID numbers in the CJA database within five months of the projected jail release date for each case within the eligibility status groups. In a handful of pre-program cases defendant NYSID numbers appeared in more than one category. As a result, in a very small number of instances the same defendant may appear with cases in more than a single category. We found a combined total of 4,879 cases arraigned in the downtown Manhattan Criminal Court in calendar year 2004 that met our selection criteria, distributed among the program eligibility categories as shown on the table below. The Spotlight category cases had the highest rate of re-arrests, 73 percent, with over half (55.0%) of the cases in the program-eligible category having re-arrested defendants. Overall, there were a total of 3,076 cases in which defendants had re-arrests within five months of each jail-sentenced cases, for a 63.0 percent overall re-arrest rate within five months of the projected jail release date.

RE-ARRESTS AMONG PRE-PROGRAM CASES PROGRAM- Number of EQUIVALENT Number of Cases with Re- Percent ELIGIBILITY Cases Arrested Re-arrested STATUS: All jail Defendants sentenced cases Eligible 2,231 1,226 55.0 Insufficient Priors 256 103 40.2 Spotlight 2,392 1,747 73.0 Total Cases 4,879 3,076 63.0 The table below shows the comparable information for all first and second year DCP-sentenced cases, which includes cases with both successful and 74 unsuccessful program outcomes. Among the first-year DCP-sentenced cases the percentages of cases with re-arrested defendants in the eligible (54.9%) and Spotlight (72.5%) categories are almost identical with those in these categories in the pre-program cases. There is a several percentage point difference between the first-year cases and the 2004 cases in the insufficient priors category, which is most likely due to the very small number of cases being compared. In comparison with the pre-program period and first-year sentenced cases, there are lower re-arrest rates for eligible (47.9%) and Spotlight cases (61.2%) among the second-year DCP-sentenced cases, with the insufficient priors category having too few cases for any meaningful comparison.

RE-ARRESTS AMONG ALL FIRST-YEAR DCP-SENTENCED CASES Number of PROGRAM Number of Cases with Re- Percent ELIGIBILITY STATUS Cases Arrested Re-arrested Defendants Eligible 319 175 54.9 Insufficient Priors 25 9 36.0 Spotlight 204 148 72.5 Total Cases 548 332 60.1 RE-ARRESTS AMONG ALL SECOND-YEAR DCP-SENTENCED CASES Number of PROGRAM Number of Cases with Re- Percent ELIGIBILITY STATUS Cases Arrested Re-arrested Defendants Eligible 307 147 47.9 Insufficient Priors 6 3 50.0 Spotlight 183 112 61.2 Total Cases 496 262 52.8

However, these displays may exaggerate re-arrest rates because they are based on cases with re-arrested defendants, not defendants from one starting case forward. This is especially relevant in regard to Spotlight cases which have active repeat offenders, and also in considering the second program year in which fewer defendants were program sentenced more than once in comparison with the first program year. To ameliorate this potential bias we recalculated re- arrest rates by selecting the first case for each defendant that met the selection 75 criteria within each category, and examined re-arrest rates five months forward from the jail-release date of the first case for pre-program defendants and five months forward of the compliance date for each DCP-sentenced defendant. The top portion of the next table shows the prevalence of re-arrests within five months of the first projected jail release date for each category of pre- program defendants, and overall. This is followed by table sections that show the re-arrest rates within the program-eligibility categories and overall for defendants in the first or only DPC-sentenced case in each of the two program years.

DEFENDANTS IN PRE-PROGRAM CASES PROGRAM- EQUIVALENT Number of Number of Percent ELIGIBILITY Re-Arrested Defendants STATUS: 3-365 Day Defendants Re-arrested Jail Sentences Eligible 2,002 1,071 53.2 Insufficient Priors 252 101 40.1 Spotlight 1,596 1,106 69.3 Total Defendants 3,850 2,278 59.2 DEFENDANTS IN FIRST PROGRAM YEAR SENTENCED CASES REARREST FROM FIRST DCP- Number of Number of Percent SENTENCED CASE: Re-Arrested Defendants Re-arrested First-year DCP Defendants sentenced cases Eligible 310 168 54.2 Insufficient Priors 25 9 36.0 Spotlight 190 135 71.1 Total Defendants 525 312 59.4 DEFENDANTS IN SECOND PROGRAM YEAR SENTENCED CASES REARREST FROM FIRST DCP- Number of Number of Percent SENTENCED CASE: Re-Arrested Defendants Re-arrested Second-year DCP Defendants sentenced cases Eligible 302 144 47.7 Insufficient Priors 6 3 50.0 Spotlight 177 108 61.0 Total Defendants 485 255 52.6 76

The overall re-arrest rates between pre-program and first-year program- sentenced defendants were almost identical, and there also was very little difference within program status categories. The overall re-arrest rate for second-year program-sentenced defendants (52.6%) was smaller than either the comparable data for the pre-program (59.2%) or first-year sentenced defendants (59.4%). Further, the re-arrest rates in the second program year in both the eligible (47.7%) and Spotlight (61.0%) categories also were lower than for defendants in these categories among either the 2004 or first-year program sentenced cases. Because of the small numbers of cases in the insufficient priors category in the second program year, no meaningful comparison can be made in this category.

Times to First Re-Arrests for Defendants in Pre-Program and DCP-Sentenced Cases The next series of tables displays the mean (mathematical average) and median (midpoint) times to first re-arrests, measured in days from the date of jail release for pre-program cases and defendants and from the compliance date for DCP-sentenced cases and defendants. The left side of the table shows the average times to first re-arrest within five months for all cases in which defendants had new prosecuted arrests within this interval, and the right side the times to the first re-arrests from the first (or only case) for each re-arrested defendant. For the pre-program period there was very little difference whether all cases or only the first case was used, and there was very little difference in the results by eligibility status. In the pre-program period, the shortest mean time to re-arrest is found among Spotlight-targeted cases. For all pre-program Spotlight cases the mean time to new arrests is 42.28 days with a median of 29.0 days, and for re-arrested Spotlight defendants the mean time to the first re-arrest is 44.33 days with a median of 32.0 days. 77

TIME TO FIRST RE-ARREST FOR PRE-PROGRAM CASES AND DEFENDANTS From First Case for Each Re- All Cases PROGRAM- arrested Defendant ELIGIBILITY Mean Median Median Mean Number STATUS Number of Number of Number of of Days Days Days Days Eligible 50.82 42.0 50.82 42.0 Insufficient Priors 49.37 31.0 49.91 32.0 Spotlight 42.28 29.0 44.33 32.0 All Categories 45.9 34.0 47.6 36.0 Combined Total Number 3,076 Cases 2,278 Defendants The tables below show the comparable data—mean and median times to first re-arrests—for DCP-sentenced cases and defendants in each of the two program years. As was found among the pre-program cases, there was almost no difference in the mean and median number of days to first re-arrests, overall and within program eligibility status categories, when average times are computed for all cases with re-arrested defendants versus just the first case forward for each re-arrested defendant. In comparison with the cases from 2004, program-sentenced cases and defendants appear to have had longer times to first re-arrests, overall and for cases and defendants in the eligible and Spotlight categories. The second program year cases and defendants had the longest mean and median times to first re-arrests among the three time periods.

TIME TO FIRST RE-ARREST FOR FIRST-YEAR DCP-SENTENCED CASES AND DEFENDANTS From First Case for Each Re- All Cases PROGRAM- arrested Defendant ELIGIBILITY Mean Median Median Mean Number STATUS Number of Number of Number of of Days Days Days Days Eligible 54.98 45.0 55.40 45.0 Insufficient Priors 68.44 81.0 68.44 81.0 Spotlight 43.18 29.5 43.14 29.0 All Categories 50.09 37.0 50.48 37.0 Combined Total Number 332 Cases 312 Defendants 78

TIME TO FIRST RE-ARREST FOR SECOND-YEAR DCP-SENTENCED CASES AND DEFENDANTS From First Case for Each Re- All Cases PROGRAM- arrested Defendant ELIGIBILITY Mean Median Median Mean Number STATUS Number of Number of Number of of Days Days Days Days Eligible 61.02 53.0 61.47 53.0 Insufficient Priors 110.67 135.0 110.67 135.0 Spotlight 49.56 41.50 50.13 42.50 All Categories 56.69 48.5 57.24 49.0 Combined Total Number 262 Cases 255 Defendants Frequency of Re-arrests Another area of comparison is the frequency of re-arrests within the five- month interval studied when defendants were re-arrested in this time frame. When re-arrests occurred, the mean and median number of re-arrests was computed within eligibility status categories for all cases with re-arrested defendants shown in the left side columns for each table, and separately from the first case for each re-arrested defendant, displayed in the right side columns. Again, the data on the tables below show little difference when mean and medians are computed for cases within eligibility categories or just from the first case for each defendant. In comparison with pre-program cases, the mean number of re-arrests among first year sentenced cases is fractionally smaller in the eligible, and Spotlight categories, and overall (2.16 re-arrests among the pre-program cases and 1.84 among the first-year program-sentenced cases). Similarly, the differences in the frequency of re-arrests from the first case for each defendant in the eligible, Spotlight, and total categories, is fractionally smaller among first program year cases in comparison with their pre-program counterparts. The differences found in the insufficient category are undoubtedly due to the small number of cases.

For second program year cases we again find that the mean number of re- arrests is fractionally smaller among the eligible, Spotlight and total categories in 79

comparison with the pre-program cases and defendants. However, of note is that the overall mean number of re-arrests for the two program years for all cases (1.84) is identical, and the difference in overall mean number of re-arrests for defendants, 1.80 and 1.85 for first and second program year defendants respectively, also are almost the same.

FREQUENCY OF RE-ARRESTS FOR PRE-PROGRAM CASES AND DEFENDANTS From First Case for Each PROGRAM- All Cases Defendant ELIGIBILITY STATUS Mean Median Mean Median Eligible 1.92 1.0 1.87 1.0 Insufficient Priors 1.51 1.0 1.48 1.0 Spotlight 2.36 2.0 2.21 2.0 Total 2.16 2.0 2.02 2.0 FREQUENCY OF RE-ARRESTS FOR FIRST-YEAR DCP-SENTENCED CASES AND DEFENDANTS From First Case for Each PROGRAM- All Cases Defendant ELIGIBILITY STATUS Mean Median Mean Median Eligible 1.73 1.0 1.66 1.0 Insufficient Priors 1.44 1.0 1.44 1.0 Spotlight 1.99 2.0 2.01 2.0 Total 1.84 1.0 1.80 1.0 FREQUENCY OF RE-ARRESTS FOR SECOND-YEAR DCP-SENTENCED CASES AND DEFENDANTS From First Case for Each PROGRAM- All Cases Defendant ELIGIBILITY STATUS Mean Median Mean Median Eligible 1.66 1.0 1.67 1.0 Insufficient Priors 1.67 2.0 1.67 2.0 Spotlight 2.09 2.0 2.10 2.0 Total 1.84 1.0 1.85 2.0

These data suggest that there were virtually no differences in the prevalence of re-arrests within a five-month interval when comparing first program-year sentenced cases with those in the pre-program period. That is, from the pre-program period we would anticipate that over half of the defendants in the cases in the eligible category, approximately 73 percent in the Spotlight category, and over 60 percent overall, would be re-arrested within five months. 80

The data for the first-year program-sentenced cases in these categories and overall were almost identical. In addition, the findings are relatively unchanged when the prevalence of re-arrests is examined among defendants from the first (or only) case. The prevalence of re-arrests in the eligible, Spotlight, and total categories were somewhat lower among both second year program-sentenced cases and defendants, in comparison with both their pre-program and first-year counterparts. Because there were a very small number of first-year program cases in which defendants did not meet the minimum three prior misdemeanor conviction criteria (the insufficient priors category) and even fewer such cases in the second program year, no meaningful comparisons could be made for this category, and they had little affect on the computation of overall re-arrest outcomes. In comparing the mean times to first re-arrests among cases and defendants in the eligible category we found slightly longer mean times to the first re-arrests among program-sentenced cases, with the longest time for cases and defendants in second-year program-sentenced cases, in comparison with cases and defendants in the pre-program period in this category. In the Spotlight category there were essentially no differences between the pre- program and first program-year’s cases and defendants, with a somewhat longer time to first re-arrests among second program-year sentenced cases and defendants in this category. There also appeared to be only small differences in the mean numbers of re-arrests, although once again the differences were somewhat greater for second-year DCP-sentenced cases and defendants. Whether any of the differences found reflect a program impact on defendants in successful program completion cases is explored in the next section. 81

COMPARING SUCCESSFUL PROGRAM CASES AND DEFENDANTS WITH MATCHED COMPARISON GROUPS The previous section of the report examined the prevalence, time to, and frequency of re-arrests of DCP-sentenced cases and defendants in comparison with cases and defendants prosecuted in the downtown Manhattan Criminal Court with similar characteristics to the DCP-sentenced population during a pre- program period. In general we found little or no differences between first program-year sentenced cases and defendants within the eligible and Spotlight categories, or overall, in comparison with the pre-program period. Only somewhat greater differences were found among second program-year sentenced cases and defendants, in comparison with the pre-program period, for the eligible and Spotlight categories, and overall. However, the comparisons in the previous section only provided a contrast to all DCP-sentenced cases. They did not isolate the cases with defendants that successfully completed the program sentence, as shown in the table below, along with their re-arrest rates.

RE-ARRESTS AMONG DEFENDANTS IN SUCCESSFUL PROGRAM COMPLETION CASES 1st Year Successful Program 2nd Year Successful Program Cases Cases PROGRAM Number of Number of ELIGIBILITY Number Cases with Percent Number Cases with Percent STATUS of Re- Re- of Re- Re- Cases Arrested arrested Cases Arrested arrested Defendants Defendants Eligible 277 146 52.7 259 118 45.6 Insufficient 22 7 31.8 5 2 40.0 Priors Spotlight 146 102 69.9 130 74 56.9 Total Cases 445 255 57.3 394 194 49.2 In this section of the report we focus exclusively on the prevalence, time to, and frequency of, re-arrests between successful program defendants in each of the two program years, in comparison with pre-program cases and defendants with similar characteristics, within program eligibility categories. The goal is to gauge whether, on the selected measures of recidivism, there were differences 82

between successful program defendants (the treatment group) in comparison with defendants and cases sharing similar characteristics but without the program’s intervention (the control group). As was done in the previous section of this report, we use the same calendar year 2004 cases to create pre-program comparison groups. However, we refined our analysis to create matched comparison groups, within program eligibility status categories, using selected case, defendant and criminal history characteristics in each program year. This was done because the case and defendant characteristics in calendar year 2004 that met our selection criteria, described in the previous section of the report, were not identical to the characteristics of successful DCP-sentenced cases and defendants in each of the first two program years. As an example, and solely for the purposes of illustration, the table below displays the crime type and charge severity categories of the prosecuted charge at the downtown Manhattan Criminal Court for all program equivalent pre-program cases, and for the cases of successful program completers in each of the two program years.

ARRAIGNMENT CHARGE CHARACTERISTICS 1st Program Year 2nd Program Year TOP Pre-Program Cases ARRAIGNMENT Successful Cases Successful Cases CRIME CATEGORY N % N % N % Harm to Persons 40 0.8 2 0.5 2 0.5 Weapon 22 0.5 1 0.2 2 0.5 Property 1197 24.5 147 34.4 150 38.1 Drugs 1659 34.0 147 34.4 110 27.9 Sex 46 0.9 2 0.5 2 0.5 Fraud 833 17.1 52 12.2 34 8.6 Misconduct 705 14.4 56 13.1 59 15.0 Obstructing Justice 93 1.9 5 1.2 5 1.3 Other/Unknown 284 5.8 15 3.5 30 7.6 Total 4879 100.0 427 100.0 394 100.0

1st Program Year 2nd Program Year Top Arraignment Pre-Program Cases Successful Cases Successful Cases Charge Severity N % N % N % A-Misdemeanor 4244 87.0 370 86.7 364 92.4 B-Misdemeanor 351 7.2 40 9.4 23 5.8 Subtotal 4595 94.2 410 96.0 387 98.2 Misdemeanor Unknown Severity (Non-Penal Law, 284 5.8 15 3.5 6 1.5 Non-VTL Charges) Felony or Violation n/a n/a 2 0.5 1 0.3 Severity TOTAL 4879 100.0 427 100.0 394 100.0

Creating Matched Comparison Groups In each program year the characteristics of successful DCP defendants and cases in each eligibility status category were matched from among those in calendar year 2004 that fit our original selection criteria—downtown Manhattan misdemeanor prosecutions in cases with recidivist defendants with an arraignment conviction and a jail sentence imposed between 3-365 days, with prior conviction records according to program eligibility statuses. For successful program defendants in the Spotlight category our matching was limited to those Spotlight category cases in 2004 in which the CJA database included counts of prior misdemeanor convictions. Using these matched comparison groups permits us to test whether there were any differences in the various facets of recidivism examined between successful DCP defendants in each program year with matched groups of pre- program defendants and cases with comparable characteristics. In addition, to further eliminate bias we matched only on the first case, by earliest arrest date, for pre-program and successful DCP-sentenced defendants. Using first-year successful DCP-sentenced cases and defendants as a guide, we selectively compared a series of demographic, criminal history, and prosecuted charge characteristics to determine which might prove to be 84

important elements for matching. We repeated this analysis for second-year cases as well, which resulted in some minor variations in the matching criteria. The number of factors on which we could match groups of cases in each program year was limited by the number of cases in each eligibility category in calendar year 2004 available for matching. Once we determined the factors on which to match cases, a composite score was created for each case. When there were more cases from calendar 2004 in a program eligibility category with the same score, the case with the earliest arrest date was selected for the matched comparison group. For each first-year successful program completion our final matching criteria with pre-program cases were based on the following: • Whether the defendant’s age was 40 or younger, or 41 or older. • Whether the defendant was non-Hispanic Black versus all other ethnic designations. • Whether the criminal record consisted of only misdemeanor convictions, both misdemeanor and felony, or only felony convictions to account for the few first-year cases in which defendants did not fit the program criteria. • Whether the severity of the top prosecuted charge at Criminal Court arraignment was of A-misdemeanor, B-misdemeanor, or was a charge outside of the Penal or Vehicle and Traffic Laws of unknown severity classification. • The categorical type of the arraignment charge using CJA’s typology described elsewhere in this report and also found in the report’s APPENDIX. For the first program year we were able to match the calendar year 2004 case and defendant characteristics with all but five of the first DCP-sentenced cases for successful program defendants, all of which were in the program- eligible category. Two of these non-matched cases had a felony charge entering arraignment that was reduced for plea purposes. In two other cases the prosecuted charges were attempted property-crime charges which reduced their severity to a B- rather than an A-misdemeanor, and we were unable to match the charge characteristics of these two cases with ones in calendar year 2004 that also had the same defendant criminal record and ethnicity characteristics. Lastly, we were unable to create a match for one of two cases of successful first- year DCP-sentenced defendants prosecuted for prostitution-related crimes. At 85 the downtown Manhattan Court location CJA does not routinely interview defendants held for Criminal Court arraignment on prostitution-related charges which greatly limited the ability to match even the very few program defendants sentenced in prostitution-related cases. For successful defendants in the second program year’s cases our final matching criteria with the cases selected for comparison from 2004 in each category were as follows: • Whether the defendant’s age was 39 or younger, or 40 or older. • Whether the defendant was non-Hispanic Black versus all other ethnic designations. • Whether the criminal record consisted of only prior misdemeanor convictions, or of both misdemeanor and felony convictions. • Whether the severity of the top prosecuted charge at Criminal Court arraignment was of A-misdemeanor, B-misdemeanor, or was a charge outside of the Penal or Vehicle and Traffic Laws of unknown severity classification. • The categorical type of the arraignment charge using CJA’s typology described elsewhere in this report and also found as APPENDIX A. The population of second-year successful defendants was slightly younger than in the first program year, with a mean age just under 40. As a result, for the age criteria in creating matched groups of second-program year cases and their 2004 counterparts we used age 39 (instead of 40) as a dividing point. We had no need to include cases with only prior felony convictions because there were no such cases in the second program year. In addition, because there were so few cases in the insufficient priors category in the second program year we excluded this group in its entirety for matched comparison group analysis. For matching of cases in the Spotlight category we again only used pre-program Spotlight-targeted cases when the CJA database included counts of prior convictions. In total we were able to match year 2004 cases with the characteristics of all but one defendant’s case in the eligible category. In this instance the court reported charge at arraignment was a low-level marijuana possession case of violation severity. For all of the analyses that follows, re-arrest for defendants in the pre- program matched comparison groups was measured within five months of the 86 projected jail release date, and for successful DCP defendants within five months of the court compliance date for the first program-sentenced case. The goal was to measure re-arrest after successful program completion so that for DCP defendants this therefore excluded any new arrest that might have occurred between the time of sentencing and the compliance date, which could have been before, during or immediately after, successful program completion, but prior to the compliance date. Although it would have been preferable to use the actual program completion date this was not an item that was monitored by CJA and therefore was not part of the data requested from CASES. However, because there is such a small window from sentencing to the compliance date, and the interval between program completion and the compliance date would be extremely short, this would at most have only a very slight effect for a very few cases.

Testing for Differences in Recidivism Between Categorical Groups of Successful Defendants in Each Program Year and Defendants in Matched Comparison Group Cases For each program year data were run to measure the different aspects of recidivism for successful program defendants in the eligible and Spotlight categories. The same information was computed for the matched comparison groups of defendants from calendar year 2004 within these categories. In addition to descriptive comparisons in text and tables, different types of tests of statistical significance were performed when differences were found, as appropriate to the nature of the measurement. For determining if differences in re-arrest rates were statistically significant we used a Chi Square test. In this analysis the Chi Square test measured whether there was a statistically significant difference between the actual re- arrest rate (percent re-arrested) found among successful program defendants in each category, in comparison with the expected re-arrest rate based on defendants in the matched pre-program comparison group cases that were without the benefit of the program intervention. When comparing differences in the mean (mathematical average) number of re-arrests, or mean times to first re- arrests, between successful program defendants and defendants in the matched 87

comparison group cases, a T-Test was applied. This statistic measures whether the mean of one distribution differs significantly from the mean of the other distribution. We used a two-tailed test which examines mean differences regardless of the direction (positive or negative) of the difference. In our tests of statistical significance for each matched comparison group the .05 level uniformly was applied as the criterion. This means that we can be at least 95 percent confident that differences found did not occur by chance.

Prevalence of Re-arrest Between Matched Comparison Groups of Pre-Program and Successful DCP-Sentenced Defendants

As can be seen from the table below, in the first program year there was an almost identical re-arrest rate between the matched groups of pre-program and successful first-year defendants and cases in the eligible category, 52.5 versus 53.2 percent. There also was very little difference in the Spotlight category, 68.1 percent for successful defendants and 72.6 percent for defendants in the matched pre-program cases. For neither of these categories did we find any statistical significance between the re-arrest rates. Because of the very small number of cases in the insufficient priors category, no meaningful interpretation can be made about this category (and the difference is not statistically significant), and this group is excluded from further analysis.

RE-ARRESTS AMONG MATCHED COMPARISON GROUPS: FIRST YEAR Defendants in First-Year Defendants in Pre-Program PROGRAM Successful Program Cases Matched Cases ELIGIBILITY Number Percent Number Percent Number of Number of STATUS Re- Re- Re- Re- Defendants Defendants Arrested arrested Arrested arrested Eligible 265 139 52.5 265 141 53.2 Insufficient 22 7 31.8 22 4 18.2 Priors Spotlight 135 92 68.1 135 98 72.6

Our findings in regard to successful defendants in second program year cases in comparison with the pre-program matched comparison groups are slightly different as shown on the table below. Among defendants in the eligible status category we do find that the more than nine percentage point lower rate of 88

re-arrests within five months of the compliance date among successful program- sentenced defendants (44.9%), in comparison with the re-arrest rate for pre- program defendants in this category (54.3%), is statistically significant. However, there is no statistical significance to the more than ten percentage point lower re- arrest rate for the successful program-sentenced defendants (57.1%) in comparison with the matched group of pre-program defendants (68.3%) in the Spotlight category. No calculation was performed for the insufficient priors group for second-year successful program defendants because there were too few in this category for analysis.

RE-ARRESTS AMONG MATCHED COMPARISON GROUPS: SECOND YEAR Defendants in Second-Year Defendants in Pre-Program PROGRAM Successful Program Cases Matched Cases ELIGIBILITY Number Percent Number Percent Number of Number of STATUS Re- Re- Re- Re- Defendants Defendants Arrested arrested Arrested arrested Eligible* 254 114 44.9 254 138 54.3 Insufficient n/a n/a n/a n/a n/a n/a Priors Spotlight 126 72 57.1 126 86 68.3 *Statistically significant (p=.03)

Time to First Re-Arrest Between Successful DCP-Sentenced and Matched Comparison Group Defendants In addition to measuring the prevalence of re-arrests between successful recidivist program-sentenced defendants and their matched comparison group counterparts within eligibility categories, the time to the first re-arrest within the five-month study interval also was measured. Time was computed in days from the compliance date, or from the projected jail release date for defendants in the matched comparison groups, and then the mean and medians of these times was compared. The T-Test was then used to determine if the differences in the mean time to first re-arrests was statistically significant within categories. Within each program status category in each program year, defendants in successful program cases had longer average (mean) times to the first re-arrest than did their counterparts in the pre-program period. However, only for second program year re-arrested defendants in the eligible category was the longer 89 mean time to first re-arrests (62.11 days) statistically significant in comparison with their pre-program counterparts (46.25 days).

TIME (In Days) TO FIRST RE-ARREST FOR FIRST-YEAR DCP- SUCCESSFUL DEFENDANTS AND MATCHED COMPARISON GROUP DEFENDANTS Defendants in First-Year Defendants in Pre-Program PROGRAM- Successful Program Cases Matched Comparison Cases ELIGIBILITY N of Re- N of Re- Mean Median Mean Median STATUS arrested arrested Time Time Time Time Defendants Defendants Eligible 139 58.60 48.0 141 46.12 35.0 Spotlight 92 48.21 35.0 98 38.31 27.0 TIME (In Days) TO FIRST RE-ARREST FOR SECOND-YEAR DCP- SUCCESSFUL DEFENDANTS AND MATCHED COMPARISON GROUP DEFENDANTS Defendants in Second-Year Defendants in Pre-Program PROGRAM- Successful Program Cases Matched Comparison Cases ELIGIBILITY N of Re- N of Re- Mean Median Mean Median STATUS arrested arrested Time Time Time Time Defendants Defendants Eligible* 114 62.11 54.0 138 46.25 35.0 Spotlight 72 48.76 35.0 86 37.33 22.5 * Difference in mean time to the first re-arrest in the eligible category is statistically significant (p> .003)

Frequency of Re-arrests Between Successful DCP-Sentenced and Matched Comparison Group Defendants The next series of tables show the results of comparing the average frequencies of re-arrests, within the five-month interval studied, among re- arrested defendants in the successful and matched pre-program comparison groups within the eligible and Spotlight categories in each of the two program years. The differences in the mean number of re-arrests between re-arrested successful program defendants and those in the matched comparison groups for eligible and Spotlight cases were not statistically significant in either program year. In the first program year there was no finding of statistical significance even though in the Spotlight category the difference in the mean number of re- arrest for successful program defendants and those in the matched comparison 90

group was larger than the difference between mean numbers of re-arrests in the eligible category, and the median number of re-arrests in the Spotlight category for successful defendants (1.0) was smaller than for those in the matched comparison group (2.0). The bottom half of the table compares the matched comparison groups for the second program year. There was only a tiny difference in the mean number of re-arrests in the eligible category (1.69 versus 1.73), and a very slight difference in the mean number of re-arrests in the Spotlight category (2.18 versus 2.45), with no differences in the median number of re-arrests in either category. Although we found that a longer mean time to first re-arrests for second-year successful program defendants in the eligible category was statistically significant in comparison with their pre-program counterparts, this did not appear to have a meaningful impact on lowering the frequency of re-arrests among recidivist defendants in the second year’s eligible category.

FREQUENCY OF RE-ARRESTS FOR FIRST-YEAR DCP-SUCCESSFUL DEFENDANTS AND MATCHED COMPARISON GROUP DEFENDANTS Defendants in First-Year Defendants in Pre-Program Successful Program Cases Matched Cases PROGRAM- Mean Median Mean Median ELIGIBILITY N of Re- N of Re- Number Number Number Number STATUS arrested arrested of Re- of Re- of Re- of Re- Defendants Defendants arrests arrests arrests arrests Eligible 139 1.65 1.0 141 1.78 1.0 Spotlight 92 1.89 1.0 98 2.24 2.0 FREQUENCY OF RE-ARRESTS FOR SECOND-YEAR DCP-SUCCESSFUL DEFENDANTS AND MATCHED COMPARISON GROUP DEFENDANTS Defendants in Second-Year Defendants in Pre-Program Successful Program Cases Matched Comparison Cases PROGRAM- Mean Median Mean Median ELIGIBILITY N of Re- N of Re- Number Number Number Number STATUS arrested arrested of Re- of Re- of Re- of Re- Defendants Defendants arrests arrests arrests arrests Eligible 114 1.69 1.0 138 1.73 1.0 Spotlight 72 2.18 2.0 86 2.45 2.0

This section of the report has attempted to examine whether successful program completion had an affect on recidivism within five months of the 91

program compliance date. Recidivism was measured in terms of the prevalence, time to, and frequency of new prosecuted arrests, using matched comparison groups of defendants and cases from the pre-program calendar year 2004 period. For successful program defendants in the Spotlight category cases we found no statistically significant differences with defendants in the matched comparison group cases in any of our measures of recidivism. There also appeared to have been no differences in patterns of recidivism among successful program clients in the eligible category in the first program year. The analysis does indicate that the lower re-arrest rate and longer mean time to first re-arrest for successful program clients in the eligible category was statistically significant in comparison with defendants in the matched comparison group for this category. However, no statistical difference was found in the mean number of re-arrests for defendants in the eligible category in either program year when comparing successful program clients with those in each year’s matched comparison group. In addition, because we only examined new prosecuted re-arrests within a five month interval, there is no way of anticipating whether there is any longer term impact on recidivism among second program year successful program clients in the eligible category. In addition, nothing in these data can help inform what aspects of the program may be affecting the lower rate and average time to first new prosecuted arrests among these defendants. 92

SUMMARY AND CONCLUSIONS The Day Custody Program (DCP) is a project of the Center for Alternative Sentences and Employment Services (CASES) offered in conjunction with the New York City Department of Correction (DOC). It is available as an alternative sentence for recidivist misdemeanants in cases arraigned on misdemeanor charges at the Downtown Manhattan Criminal Court location who otherwise would be expected to receive a jail sentence for an arraignment conviction. For defendants in program-sentenced cases, DCP offers the court a sanction combining accountability, needs assessment and early intervention programs, and discharge planning services, over the course of three days in a DOC facility adjacent to the courthouse. However, unlike traditional jail sentences, defendants are not held in a correctional facility for the entirety of the sentence and, instead, are responsible for returning to the program location each day. At sentencing an intermittent jail sentence is imposed with the understanding that failure to satisfactorily complete the program’s requirements will result in the imposition of a jail sentence, usually of ten days’ duration. Defendants have ten weekdays in which to fulfill the three program days’ activities, and a court compliance date is scheduled shortly thereafter. Depending on the day of the week and the time of day at which the sentence is imposed some defendants begin their sentence immediately upon imposition; in other instances defendants must report to the program location on a subsequent day to begin their program sentence. DCP is a selective program with a series of eligibility criteria for the cases and defendants it will accept. Among these criteria are that defendants must already have at least three prior misdemeanor convictions and be in custody awaiting Criminal Court arraignment for a misdemeanor crime in a non-Spotlight targeted case. Under the conditions of funding CASES’ is supposed to exclude defendants in Spotlight-targeted cases, a separate City initiative designed to increase the use and length of jail sentences imposed citywide on recidivist defendants actively cycling through the Criminal Courts who are arrested and arraigned on misdemeanor charges. However, DCP may accept defendants in

Spotlight-targeted cases if solicited to do so by other court personnel. Further, DCP and other court personnel may be unaware of Spotlight targeting if the court papers are not appropriately marked. In addition, defendants must be able to immediately begin the program sentence if convicted at arraignment, so that defendants with outstanding bench warrants, or who are on parole, cannot receive a program sentence unless these impediments can be resolved before the conclusion of the arraignment proceedings. Court papers are reviewed by CASES’ staff prior to Criminal Court arraignment to determine program eligibility. Defendants found to be “paper eligible” are then interviewed by CASES’ staff upon the consent of the defense attorney and after consultation with the prosecuting attorney to determine if there will be a plea offer of jail time greater than time served. In some instances of cases of otherwise ineligible defendants an interview may be conducted at the request of other court personnel. Once interviewed prior to arraignment by CASES staff, defendants may be rejected by the program for a variety of reasons including issues beyond the program’s capacity to address, or by the refusal of the judge or even by the defendant him/herself to consent to the Day Custody sentence. As described in this report, in the program’s second full year of operation, October 1, 2006 through September 30, 2007, CASES’ court staff screened the court papers in a total of 3,751 cases of which 496 cases, or approximately 13%, received a program sentence. CJA identified the NYSID numbers (New York State fingerprint identification) of 3,142 individuals among these 3,751 cases. In the second program year there were 485 defendants in the 496 DCP-sentenced cases, with 11 defendants receiving the program sentence in two separate cases. CASES’ selection criteria were met in 61.9% of the program-sentenced cases in the “Eligible” category. Over a third of the second-year DCP-sentenced cases were Spotlight-targeted cases, but they represented only a very small percentage of all Spotlight-targeted cases arraigned in the downtown Manhattan 94

Criminal Court in the same time period. In only a handful of the cases were there defendants with insufficient numbers of prior misdemeanor convictions. Defendants in the DCP-sentenced cases were mostly male, with an average age of close to forty, and approximately two-thirds were non-Hispanic Blacks with about another quarter Hispanic. An overwhelming majority of defendants in program-sentenced cases had far more than the minimum three prior misdemeanor convictions, with an average of 14 prior convictions for misdemeanor crimes. Over two-thirds (67.7%) of the defendants in the DCP- sentenced cases also had previously been convicted of at least one felony crime. Overwhelmingly CJA rated defendants in the DCP-sentenced cases as poor risks for recognizance release had the cases been continued at arraignment and, among defendants in these interviewed cases, only one-quarter reported being engaged full-time in work or other activities at the time of the arrest. Many of the characteristics of defendants in second-year DCP-sentenced cases were the same or very similar with those in the first program year. However, there were some differences. For example, in responses to CJA interview items a greater percentage of defendants in second-year in comparison with first-year cases informed CJA that they lived with others, but a smaller percentage expected someone to attend their Criminal Court arraignment appearance. Some of the differences found in defendant characteristics between the two program years may have been associated with some small shifts in case composition. However, the extent to which these differences reflect changes in the characteristics of misdemeanor cases and defendants being prosecuted in the downtown Manhattan Criminal Court over time could not be determined from the program data, and pursing this issue was beyond the scope of this research. However, changes over time in defendant characteristics among DCP-sentenced cases could affect the successful completion rates and needs of the client population. Well over a third of the DCP-sentenced cases had defendants arrested for drug crimes and another third for property crimes—mostly petit larceny, with most of the remaining cases involving fraud crimes, or public disorder offenses 95

such as criminal trespass. However, at Criminal Court arraignment the property crime category had the largest percentage of cases, over six percentage points greater than the percentage of cases in the drug category. This change in charge composition is important because almost all convictions in DCP- sentenced cases were to the arraignment charge, and conviction for petit larceny tends to have a greater likelihood of jail time and comparatively more of it than almost all other misdemeanor charges. In addition, 92.3% of all DCP-sentenced cases were arraigned (and convicted) for crimes of A-misdemeanor severity. Our research found that almost 80% percent of defendants in the 496 second-year cases sentenced to DCP completed the program requirements. A successful completion rate of 84.4% was found among defendants in cases that fit the program’s specific eligibility requirements, with a lower successful completion rate for defendants in Spotlight-targeted cases sentenced to the program. Among the 102 cases in which defendants failed to satisfactorily complete the program sentence, 85.3% had been returned to court for re-sentencing as of May 30, 2008, the study cut-off date for tracking the cases. In almost all of the 87 re-sentenced cases this research found that defendants returned to court and re-sentenced after program failure were held accountable by appropriate sentences to jail time. Most of the defendants in the re-sentenced cases were returned to court as a result of new arrests; in only seven cases were defendants returned and re-sentenced in the DCP cases without a new prosecuted arrest. Among the 87 cases re-sentenced subsequent to being returned to court after a new arrest, 41 were re-sentenced in the DCP cases separately from the outcome and sanction on the new arrest, and the remaining 46 had the re- sentence in the DCP case imposed at the same time the new prosecuted arrest also was being disposed. As a result, it is not possible to precisely account for the penalty assessed for the program failure as in many instances the fact of a new arrest, or the actual imposition of the sentence in the re-arrest case, invariably influenced the DCP cases’ re-sentence. In addition, in a few instances the alternative jail time set with the DCP sentence was for more than ten days so 96

that not all re-sentences for program failure necessarily would have been for exactly ten days. What we did find, however, was that in all but three of the re- sentenced cases the defendant received a jail sentence. In addition, over half of the jail sentences imposed were for ten days, and another 22.6% received a sentence greater than ten days, with the largest number of these being sentences of 20- or 30-days. In addition to the DCP-sentenced cases, there were 3,255 cases involving 2,750 defendants screened but not program sentenced during the program’s second year. CASES’ court staff was the source of the rejection for 62.8% of the cases, with defense attorneys and judges being the next largest sources for rejections. Among those rejected by CASES’ staff, over three-fifths were excluded because of Spotlight-program targeting (as were over a third of the rejected cases overall). Among the remaining cases rejected by DCP staff, homelessness/insufficient community ties and an active hold were the next two most frequent reasons given for rejection. The characteristics of defendants in rejected cases closely resembled those in program-sentenced cases: overwhelmingly male, non-White, and 35-years-of- age or older. Most had far more than the requisite three prior misdemeanor convictions, and only about 30% did not also have prior cases with conviction to a crime of felony severity. Further, like their program-sentenced counterparts, defendants in the rejected cases overwhelmingly were rated by CJA as being of high risk of failing to appear if released on recognizance were the case to be continued at the Criminal Court arraignment. These characteristics were fairly consistent regardless of the source or reason for rejection, although percentages among rejection categories varied. However, there were some differences in the top arrest charges and their severities between the sentenced and rejected cases. Drug crime had the largest percentage of arrest charges for both groups’ cases, but there was a higher percentage found at arrest among the rejected cases. Similarly, while the property crime category had the second largest percentage of top arrest charges for both sentenced and rejected cases, the percentage was smaller among the 97 rejected cases. In addition, a smaller percentage of rejected cases had top arrest charges of felony severity, although to some extent this reflects the large proportion of Spotlight cases among the rejects all of which can only have misdemeanor-severity top arrest charges. Somewhat over two-thirds of all rejected cases were disposed at Criminal Court arraignment, almost all with a conviction, and 84.2% had some form of a jail sentence imposed for the arraignment conviction. However, almost two-thirds of these cases received a jail sentence of less than ten days (including time- served sentences). Examining the characteristics of court charges and severities of rejected cases with arraignment convictions revealed some differences in comparison to the program-sentenced cases. The largest percentages of all rejected cases convicted and sentenced at arraignment were prosecuted for drug charges with the second largest percentage found in the property crime category. Among DCP-sentenced cases this order was reversed, with a larger percentage of arraignment charges falling into the property crime category than in the drug category. In addition, a somewhat greater percentage of rejected cases convicted at arraignment were prosecuted for crimes of B-misdemeanor severity, but this charge severity category had a relatively small percentage of cases for both the rejected and sentenced groups. These overall differences in charge and severity composition between rejected cases convicted at arraignment and program-sentenced cases strongly suggest that sentenced cases of program-eligible defendants would more likely have had jail sentences imposed, and of longer duration, than the rejected cases with arraignment convictions (excluding Spotlight-targeted cases). This study did find some differences among the rejected cases depending on the source or reason for program rejection. For example, Spotlight program cases and those rejected by judges received comparatively longer jail sentences at arraignment in comparison to cases rejected from other sources or for other reasons. Conversely, cases rejected by prosecuting and defense attorneys, or 98

defendants themselves, had the lowest percentages of jail-time sentences imposed for an arraignment conviction. In an initial examination of re-arrest patterns among DCP-sentenced cases, CJA’s research found that successful program clients had lower rates and longer times to the first re-arrest, within five months of the compliance date of the DCP- sentenced cases, than unsuccessful program clients. This examination of re- arrests also found that if defendants were re-arrested, there was virtually no difference in the frequency with which re-arrests occurred between successful and unsuccessful program completers within five months of the compliance date. However, defendant and case characteristics between the successful and unsuccessful program completion groups were too dissimilar to form any basis for drawing conclusions about the program’s impact on recidivism among defendants in the successful program completion group. In order to begin to more closely examine recidivism among DCP- sentenced cases and defendants, comparison groups were created from pre- program cases in calendar year 2004 with similar case and defendant characteristics to those in program-sentenced cases. Among the selection criteria in the creation of comparison groups were that the cases had recidivist defendants held for arraigned in the downtown Manhattan Criminal Court, prosecuted for a misdemeanor charge, or a charge outside of the Penal and Vehicle and Traffic Laws of unknown severity (almost all of which are of misdemeanor-equivalent severity), with an arraignment conviction accompanied by a jail sentence of between 3 and 365 days. Pre-program cases were divided into groups that mirrored the program-eligibility categories of DCP-sentenced cases: eligible, insufficient prior misdemeanor convictions, and Spotlight-targeted cases. The first stage of the recidivism analysis was to compare the prevalence of new prosecuted arrests, and the time to and frequency of new arrests among cases of re-arrested defendants, between pre-program comparison groups and first- and second-year program sentenced cases. For each pre-program case a projected jail release date was computed and the CJA data was searched for 99

new prosecuted re-arrests within five months of jail release. For program- sentenced cases a five-month interval from the program’s compliance data was used. Among the pre-program cases the research found an overall re-arrest rate within five months of jail release of 63.0 percent, with a lower re-arrest rate for cases of defendants in the ‘eligible’ category (55.9%) and higher re-arrest rate for defendants in the Spotlight cases’ category (73.0%). There were virtually no differences from these expected rates among first-year program-sentenced cases. There were, however, somewhat lower re-arrest rates overall and in the eligible and Spotlight categories among second-year program-sentenced cases. Because of the small numbers of cases in the insufficient priors category no meaningful comparisons could be made, and they had little effect on overall re- arrest rates. These patterns were unchanged when the data regarding the prevalence of new prosecuted arrests was analyzed using only new prosecuted arrests in the same five-month interval for the first case for each defendant, as opposed to using all cases which included multiple arrests for the same defendants. For re-arrested cases and defendants we examined the mean (mathematical average) and median (midpoint) times to first arrests, within program eligibility categories for all cases and then only for the first case for each defendant The findings in regard to average times to first re-arrests for cases in the pre-program and first-program year periods were almost identical regardless of whether the computation was based on all cases or only on time to a new prosecuted arrest after the first (or only) case. There were, however, somewhat longer average times for re-arrested second-program year cases and defendants. On average, the time to first arrests for defendants in the pre-program eligible category of cases was approximately 50 days after jail release, in comparison with an average of about 55 days for first-year DCP-sentenced cases and defendants, and approximately 60 days for second-year program- sentenced cases and defendants. Among Spotlight category cases the average 100

time to first re-arrests in both the pre-program period and first year was approximately 43 days, and approximately 50 days in the second program year. For re-arrested cases and defendants we also examined the mean and median frequencies of new arrests, within the five-month period studied. These were examined for all cases within program eligibility categories and then only for the first case for each defendant. The mean number of new arrests within the five-month period studied, overall and for cases and defendants in the eligible category was less than two, with a median of one re-arrest. In the Spotlight category the average was approximately two with a median of two. Overall, and for the eligible and Spotlight categories, the largest average number of re-arrests was in the pre-program period, but the differences with program-sentenced counterparts were only fractionally different, and the medians within categories were the same. These data appear to indicate that there were few differences in patterns of recidivism for defendants in the first-year’s program-sentenced cases in comparison with those in the pre-program period, and some greater variation for defendants in second-year program sentenced cases. However, this analysis did not distinguish between cases of successful and unsuccessful program clients. To further refine our analysis, and to focus on whether successful program completion affected the different facets of recidivism studied, matched comparison groups were created. For the first case for each successful program client in each of the two program years a composite score was created based on defendant, criminal history, and prosecuted charge characteristics. The same factors were used to create a composite score among pre-program period cases. These scores were then matched, for successful defendants in the eligible and Spotlight categories in each program year, with the scores of pre-program defendants’ cases within the equivalent program eligible categories. The matched comparison groups were used to examine difference within categories between the pre-program period and each year’s groups of program successful clients. Tests of statistical significance were used, as appropriate, to determine if differences found were likely to have occurred as a result of a program affect or 101

by chance. Because there were only a few first year program cases of defendants with insufficient priors, and even fewer among the second program year’s cases, no attempt was made to measure differences in this eligibility category. We found little if any differences in re-arrest rates between first program year successful defendants and their pre-program counterparts in the eligible and Spotlight categories, and when there were differences they were not statistically significant. We did find somewhat greater differences among second-year successful program defendants in these categories in comparison with their matched comparison groups from calendar year 2004. However, only for the nine percentage point lower rate of re-arrests within five months of the compliance date for defendants in the eligible category (44.9%), in comparison with the rate of re-arrests within five months of jail release for pre-program defendants (54.3%), was the difference statistically significant. In examining the mean times to the first re-arrests for re-arrested defendants, we found longer times to first re-arrests for successful defendants in both the eligible and Spotlight categories in both program years, but only the longer time for second-year defendants in the eligible category was statistically significant. There were almost no differences in the average numbers of re- arrests within the five month interval for either eligible or Spotlight defendants in either program year, in comparison with their matched comparison group cases, and none of these very small differences was statistically significant. The goal of the matched comparison group analysis was to examine if successful program completion had a measurable impact on various aspects of recidivism. Among first-year successful program clients this analysis found little if any differences on the measures of recidivism between successful defendants in the eligible or Spotlight categories in comparison with their pre-program matched comparison group counterparts and, when differences were found, none were statistically significant. In the second program year greater differences were found, with lower re- arrest rates and longer average times to first re-arrest for defendants, in both the 102 eligible and Spotlight categories in comparison with their matched comparison groups. However, only the differences for second-year successful defendants in the eligible category were found to be statistically significant. In addition, there were little differences in the mean and median numbers of re-arrests in the five month interval for any of the matched comparison group categories in either program year, and none of the differences were found to be statistically significant. These data suggest that there may have been some measurable positive effect on recidivism only for successful program defendants in the eligible category in the second program year. However, without knowing more about defendant needs, and services provided, this research cannot determine what aspects of the Day Custody Program may have had an affect on recidivism for this group of defendants. As a result, more information would be needed about “what works” before the Day Custody Program experience could provide guidance to other program providers and government agencies as they seek to promote policies and programs to better serve the discharge planning and community-service needs of offenders during re-entry and re-integration upon release from jail.

103

APPENDIX

CJA’S CRIME-TYPE CATEGORIES. The categories and principal charges found in CJA’s typology of offenses are as follows: • The harm-to-persons category consists of some of the most serious felony crimes such as murder, manslaughter and rape, as well as various types of assault including those of misdemeanor severity. • The harm-to-persons-and-property category consists almost exclusively of violent felony offenses such as robbery. • The weapon category includes all charges relating to use, possession and sale of weapons found in Article 265 of the New York State Penal Law. • The property crime category includes petit larceny at the misdemeanor level, and grand larceny at the felony level, and also includes related offenses such as possession of stolen property or burglar’s tools. This category also includes criminal mischief charges that can be of either felony or misdemeanor severity. • The drug category includes all non-marijuana charges found in Article 220, and all marijuana charges found in Article 221, of the New York State Penal Law. • The sex-crime category contains primarily prostitution-related offenses such as promoting prostitution, soliciting for the purposes of prostitution, prostitution, or patronizing a prostitute. • The fraud category principally contains the theft-of-services misdemeanor charge, the vast majority of which involve fare beating, and other types of theft- by-deception activities such as forgery, credit card and welfare fraud, or trademark infringement. • The misconduct category contains a variety of public-order offenses such as criminal trespass, harassment, disorderly conduct, and loitering, and other charges such as illegal gambling or obstructing governmental administration. • The obstruction-of-justice category includes charges such as resisting arrest and criminal contempt charges including violating protection orders. • VTL refers to offenses contained in the Vehicle and Traffic Laws, the vast majority of which are considered to be of unclassified misdemeanor severity. • The other/unknown crime category contains mostly charges from sources outside the Penal Law or VTL, such as the City’s Administrative Code or the State Tax Code (e.g. untaxed cigarettes), and small numbers of charges not recognized in the CJA database.