Research Brief No. 2 NEW YORK CITY CRIMINAL JUSTICE AGENCY, Inc.
April 2003
A series of reports summarizing current research from CJA Executive Director, Jerome E. McElroy Director, Research Dept. Richard R. Peterson, Ph.D. Research Brief Editor, Mary T. Phillips, Ph.D. Graphics & Production, Raymond Caligiure Administrative Associate,
Bernice Linen-Reed CJA is a not-for-profit corporation that provides a variety of criminal justice services under a contract with the City of New York. CJA staff interview defendants arrested in New York City, make recommendations for pretrial release, and notify released defendants of upcoming court dates. Within the Agency, the Research Department conducts studies covering a broad array of criminal justice policy concerns. The Research Brief series summarizes the results of some of these studies.
New York City Criminal Justice Agency 52 Duane Street New York, NY 10007 PHONE: 646 213-2500 FAX: 646 213-2650 WEB: www.nycja.org 2003 NYC Criminal Justice Agency
The Impact of Felony ATI Programs On Recidivism By Jukka Savolainen, Ph.D.
Are ATI participants released into the community less likely to be re-arrested than other felony offenders? New York City funds a variety of Alternative-to-Incarceration (ATI) programs for defendants charged with misdemeanor or felony offenses. Focusing on adult offenders participating in felony-level ATI programs, the purpose of this study is to evaluate the effectiveness of six of these programs from the point of view of recidivism. Specifically, using re-arrest as the measure of recidivism, we ask: how do felony defendants who have completed an ATI program compare with similar defendants who did not partici-
pate in such programs? What difference, if any, do these programs make in terms of reducing criminal behavior? The six ATI programs discussed in this report are described briefly in the sidebar on the following page. Two of them, Women’s Day Treatment (WDT) and DAMAS, serve female offenders; WDT, El Rio, and FlameTree are drug treatment programs; Court Employment Project (CEP) and Freedom provide educational and vocational services (CEP is for 16to-19-year-old offenders).
This Research Brief is adapted from: Criminal Recidivism Among Felony-Level ATI Program Participants in New York City (August 2002) by CJA Senior Research Analyst Jukka Savolainen, Ph.D. An Executive Summary is also available. For more detailed information, see the full report on the CJA web site: www.nycja.org/research/research.htm Systems programming: Wayne Nehwadowich, Aïda Tejaratchi.
R Reesseeaarrcchh B Brriieeff ATI Programs Included in the Research Court Employment Project (CEP) of the Center for Alternative Sentencing and Employment Services (CASES) offers educational and vocational services to young (16-to-19-year-old) felony offenders. Freedom Program of the Fortune Society offers educational and vocational services to adult (19 and older) felony offenders. El Rio Day-Treatment Program of the Osborne Association offers substance abuse treatment to adult felony offenders in Manhattan and the Bronx. FlameTree Program of the Fortune Society offers substance abuse treatment to adult felony offenders in Manhattan, Brooklyn, and Queens. DAMAS Program of the Fortune Society offers educational and vocational services to female felony offenders. Women’s Day Treatment (WDT) Program of the Project Return Foundation offered outpatient substance abuse treatment to female felony offenders at the time of this research. The program was discontinued early in 2002 after the City significantly reduced its funding. ____________
These programs are funded by the Mayor’s Office of the Criminal Justice Coordinator (OCJC). An additional OCJC program, the Youth Advocacy Project (YAP) of the Center for Community Alternatives, is a felony ATI program for juvenile offenders. Because the database used in our research for tracking recidivism is not able to measure juveniles with sufficient accuracy, we are not discussing our findings for YAP in this report. ____________
ATI programs operating in New York City with funding from sources other than the OCJC were not included in this research.
2
Post-Program Focus Before we can measure any effects on recidivism a program may have, we must first let it run its full course. As the programs are designed to last about 6 months to a year, it would be unreasonable to expect any significant effects as soon as participants begin to receive their treatments and/or services. Thus, any act of recidivism that takes place while a client is still in the program is not counted in this study. Our sole focus is on what happens once the participant has completed the program. Because of this exclusive focus on postprogram recidivism, this study examined only those ATI participants who were released into the community upon program completion. A significant portion of felony program participants are sentenced to either prison or jail as a result of inprogram re-arrest or some other violation of a program requirement. As these incarcerated clients are not at risk of post-program recidivism, they cannot be included in our analysis. At the same time, a large number of ATI clients remained in the community despite a lessthan-perfect record of participation. In our sample, 37% of the cases were terminated from a program due to poor attendance or some other violation (such as in-program re-arrest). Indeed, we compare results separately for successful and unsuccessful ATI clients towards the end of this report.
Matched comparisons In order to provide a meaningful context for our research findings, the recidivism rates of the ATI participants are analyzed in comparison with a similar group of non-ATI felony offenders in three sentencing categories: 1. Probation: felons serving a straight probationary sentence; 2. Prison: felons released from a New York State correctional facility; 3. Jail: felons discharged from a New York City jail. April 2003
R Reesseeaarrcchh B Brriieeff Each offender in each comparison group was released or — in the case of probationers — sentenced into the community concurrently with the ATI participants: between July 1, 1998, and March 31, 2000. As a key aspect of the research design, the comparison groups were assembled by finding an individually selected match for each ATI participant from each of the three sentencing categories. For example, a 27-year-old female ATI participant with one prior misdemeanor conviction was matched with a female felony convict from the jail, prison, and probation populations sharing these (and several other) characteristics. The variables used in the process of finding suitable matches are described in the accompanying box. To the extent that a suitable match was not available, the relevant ATI participant was not included in the analyses involving that particular comparison group. Fortunately, we were able to find matches in each comparison group for nearly all ATI cases. In fact, 80% of the ATI cases had a suitable match in every comparison group, and only 9 ATI cases (less than 1%) had no suitable match in any of the three sentencing categories. The proportion of unmatched cases is the lowest in comparisons involving prisoners (3%), followed by probationers (6%), and those released from jail (18%). Since every ATI client did not have a match in each comparison group, the exact size and composition of the ATI sample varies somewhat across comparisons.
Measures of recidivism To collect information about recidivism, each individual included in the study was tracked for one year. Our measures rely on official data for re-arrests and convictions in New York City. In this research, both re-arrest and conviction data were analyzed and reported. However, as the analysis of conviction data did not add to the results obtained with re-arrest data, the only indicators of recidivism discussed here are the ones based on re-arrests. We analyzed these data from the perspective of prevalence, incidence, and timing: Research Brief #2
How matching was done 1. We considered the following nine characteristics in the process of finding suitable comparisons for the ATI participants: A. Demographic profile · Age at the time of instant arrest · Sex · Borough of prosecution (instant arrest) B. Instant arrest · Detention status leaving criminal court arraignment · Top indictment charge type (property, drug, violent, weapon, etc.) · Top indictment charge severity (A-, B-, C-, D-, or E-felony) C. Criminal Record · Prior adult arrests · Prior misdemeanor convictions · Prior felony convictions 2. A simultaneous application of these nine characteristics generates a detailed, and in some instances an entirely unique, matching profile for each ATI participant. The comparison groups were selected by choosing one individual sharing this same profile from each sentencing category (probation, prison, jail).
Prevalence refers to the percentage of those who recidivated at least once during the tracking period; incidence indicates the average number of times this event occurred per person for the group; and timing measures how long, on average, it took until an offender recidivated. The analyses of timing were limited to the data on the offender’s first re-arrest. All three measures were used in both aggregate and program-specific comparisons. In this brief summary we report one result from each of these three measurements representing the core pattern of findings. 3
R Reesseeaarrcchh B Brriieeff Analysis The main research question of this study is: are there any differences in the rates of recidivism between the ATI participants and matched cases in the three comparison groups? Because the comparison groups were selected to resemble their ATI counterparts as closely as possible, any differences in re-arrest rates are likely to reflect true differences between these groups rather than some unconsidered characteristic of the defendants. However, as the ninevariable matching profile does not guarantee perfect similarity between the ATI sample and its comparison groups, we verified our findings in multivariate analyses that took into account six additional variables (area of residence, employment/schooling status, race/ethnicity, previous Youthful Offender status, representation by private/public defense attorney, and the timing of the re-arrest.) Since the results from these analyses were the same as the results from our simpler comparisons of the matched ATI, jail, prison and probation groups, we will discuss only the latter results in this Research Brief.
Figure 1 The Prevalence of Recidivism Among ATI Participants and Their Matched Comparison Groups (% Ever Re-arrested in the One-Year Follow-Up Period) A) ATI vs. Probation (N = 943, each group) Percent
0
10
20
4
40
50
Re-arrested First re-arrest is felony
ATI
Probation
B) ATI vs. Prison (N = 970, each group) Percent
0
10
20
30
ATI
Prison
40
50
Re-arrested First re-arrest is felony
Findings As shown in the first panel (A) of Figure 1, the percentage of offenders who were re-arrested at least once during the tracking period is about the same for probationers (42%) and ATI participants (41%). The next comparison in panel A concerns only those first re-arrests that were associated with a felony-level charge. The difference between probationers and ATI participants is equally trivial in this regard. As indicated by the next panel (B), this same basic pattern of findings prevails in comparisons between the ATI participants and the prison group: the prisoners are no different from ATI participants in terms of the prevalence of re-arrests. Felons released from jail constitute the only comparison group with a re-arrest rate that is different from the ATI participants (Panel C of Figure 1). Close to 53% of cases in the jail group were rearrested within one year of their discharge. This rate is about 9 percentage points higher than in the corresponding ATI sample. As there is no mean-
30
C) ATI vs. Jail (N = 827, each group) Percent
0
10
20
30
40
50
60
Re-arrested First re-arrest is felony
ATI
Jail
ingful difference between these two groups in felony re-arrests, we may conclude that the initial difference stems from a higher rate of misdemeanorlevel re-arrests among the jail group. The comparison with the jail group regarding the prevalence of re-arrest is the only difference presented in Figure 1 that is statistically significant, which means that it is highly unlikely to have occurred by chance alone. The small differences in April 2003
R Reesseeaarrcchh B Brriieeff re-arrest rates found in Panels A and B were not statistically significant, so they are more likely to have occurred by chance alone. A large difference is more likely to be statistically significant, as are differences found in large samples. Program-specific findings In light of the above comparisons, ATI participants are no different from probationers or prisoners but they are somewhat less likely to be rearrested than felons released from a New York City jail. These findings were based on aggregate comparisons pooling six different ATI programs into a single group. In the next series of charts, displayed in Figure 2, we present findings from program-specific comparisons of the frequency of re-arrests. Whereas prevalence was the measure presented for the aggregate comparisons, incidence is the measure presented for the program-specific comparisons. As it happens, the program-specific analyses tend to reinforce the ATI-wide results described earlier. None of the individual ATI programs differs from a matched group of probationers by a significant margin (Panel A). In comparisons with prisoners (Panel B), CEP is associated with a slightly higher rate of rearrests than its comparison group. This is the only prison comparison that meets our standard of statistical significance. Thus, any differences in the comparisons involving the five other programs are too minor to be viewed as genuine. Finally, as indicated in Panel C, each ATI program is associated with a lower average number of re-arrests per participant than its matched jail comparison group. Most (4) of these differences are statistically significant. (Despite the fact that the incidence of re-arrests in WDT is little more than half that of the matched jail sample, the difference is not statistically significant because of the small number of cases.) These program-specific comparisons of incidence sustain the basic conclusion that the felony ATI participants do not recidivate any more or any less than felons sentenced to probation or felons released from prison. The program-specific analyses also confirm that felony convicts discharged from jail are more likely to be re-arrested than a comparable group of post-program ATI participants. Research Brief #2
Figure 2 Program-Specific Differences Between ATI Participants and Their Matched Comparison Groups in the Incidence of Re-arrests A) Comparisons with Probation Average number of re-arrests 0
0.2
0.4
0.6
0.8
1
1.2
CEP DAMAS El Rio FlameTree Freedom WDT ATI
Probation
B) Comparisons with Prison Average number of re-arrests 0
0.2
0.4
0.6
0.8
1
1.2
CEP* DAMAS El Rio FlameTree Freedom WDT
ATI
Prison
C) Comparisons with Jail Average number of re-arrests 0
0.2
0.4
0.6
0.8
1
1.2
1.4
CEP* DAMAS El Rio* FlameTree* Freedom* WDT^ ATI
Jail
*Statistically significant at .10 level. ^ N <50
5
R Reesseeaarrcchh B Brriieeff Does post-release supervision matter?
6
Figure 3 Timing of First Re-arrest Among Felony-ATI Program Participants and a Matched Group of Felons Discharged from Jail 100%
90%
80%
Percent Not Rearrested
What accounts for the higher rate of recidivism in the jail group? As our research was not designed with this question in mind, we can only speculate about possible reasons. That said, one characteristic that seems to distinguish jail cases from the three other groups is the absence of post-release supervision. Discharged prisoners have to report to a parole officer and probationers to a probation officer; probation is also a typical sentence for an ATI participant who remains in the community. Could this factor account for the difference between jail and the other sentencing alternatives? As a way of evaluating this hypothesis, it may be informative to examine the data concerning the timing of re-arrest. If post-release supervision matters, it is reasonable to assume that this effect increases with time. Upon release into the community, each convicted felon in our study has had a recent, prolonged, and fairly intense contact with the criminal justice system. We would expect this experience to have some deterrent effect on recidivism. However, as time goes on, the deterrent effect of punishment is likely to wear out. We hypothesize that the deterrent effect of punishment deteriorates more quickly among felons who are left with no contact with the criminal justice system compared with those who remain under the supervision of a probation or parole officer. It is our assumption that regular contact with a criminal justice representative extends the deterrent effect of punishment. The following prediction can be derived from this hypothesis: In the beginning stages of the one-year tracking period, there is little or no difference between the ATI and jail groups; the higher re-arrest rate associated with the jail group begins to emerge only after the first weeks or months of the tracking period have gone by. While this hypothesis is not specific enough to enable us to estimate the precise timing for this difference to start showing, it does predict a clear pattern, which can be empirically tested. Figure 3 describes the relevant data concerning the timing of re-arrests among ATI participants and a matched sample of felons discharged from jail. The statistics described in this chart
70%
60%
50%
40%
30% 0 30 60 90 120 150 180 210 240 270 300 330 360 365
ATI Jail
Days
indicate the proportion of felons who have â&#x20AC;&#x153;survived the risk of recidivismâ&#x20AC;? at given points in time. As everybody starts with a clean record, the survival rate equals 100% for both groups on the first day. At the end of tracking, only 47% in the jail group remained without a new arrest, compared to nearly 55% in the ATI category. The most critical aspect of Figure 3 is the fact that during the first three months of tracking, there are no differences between felons discharged from jail and those released from an ATI program. While the true meaning of this finding cannot be ascertained in our research, this pattern is consistent with the hypothesis that post-release supervision helps to prolong the deterrent effects of punishment. (We conducted the same set of analyses with both probationers and prisoners and found that their curves were indistinguishable from that of the ATI group.) April 2003
R Reesseeaarrcchh B Brriieeff Does program success matter?
Conclusions
As the final step in the analyses, we separated those ATI participants who completed their program “with success” as defined by the programs from those who were unsuccessfully terminated. (Reasons for “unsuccessful” termination varied from case to case and program to program, and included nonattendance, other program violations, and re-arrest while in the program.) Assuming that the services and treatments these programs offer have any effect on recidivism, it is reasonable to expect that this effect would be stronger for participants who successfully fulfilled the program requirements. The prevalence of re-arrests was somewhat lower in the successful segment of the ATI group than in the ATI group as a whole. However, we were surprised to find this difference to be less than 2 percentage points. More importantly, there was no difference between the successful graduates and their matched comparison groups. The results from these comparisons were identical to the results concerning the full ATI group, which includes unsuccessful terminations. Thus, success in the program, as defined by the programs themselves, does not translate into reduced levels of recidivism. (These analyses were not conducted with program-specific data because the samples would have been too small for reliable statistical generalizations.)
The purpose of this study was to evaluate the effects on recidivism of six ATI programs targeting felony-level offenders in the New York City criminal justice system. The post-program recidivism of the ATI participants was compared to matched groups of convicted felons from three traditional sentencing alternatives: those serving a straight probationary sentence, those discharged from state prison, and those discharged from jail. We found no differences among felony-ATI participants and felons serving a probationary sentence or felons released from a state prison. However, the ATI group was systematically less likely to recidivate than a similar group of offenders discharged from a City jail. Specifically, felons in the jail group were more likely to have been re-arrested for a misdemeanor offense; there were no differences in felony re-arrests between these two groups. In light of these findings, these ATI programs do not appear to serve any better or any worse in reducing their participants’ recidivism than the New York State Department of Corrections (prison group) or the New York City Probation Department. However, they do seem to reduce the criminal behavior of their clients somewhat more effectively than the New York City Department of Corrections (jail group).
POLICY IMPLICATIONS ·
The felony ATI programs do not have any advantage over a prison sentence or probation; these three options appear to be of practically equal value from the perspective of offender recidivism.
·
To the extent that they are viewed as alternatives to a jail sentence, these ATI programs can be recommended as more effective in reducing recidivism.
·
The higher level of recidivism associated with felons released from jail may stem from the absence of post-release supervision. More research is needed to address this question properly.
Research Brief #2
7
Already published in this series: No. 1 (December 2002): Jail Displacement for ATI Programs by Mary T. Phillips, Ph.D.
Forthcoming: Combating Domestic Violence in New York City by Richard R. Peterson, Ph.D. This article will summarize the results from three recent studies of domestic violence by CJA’s Associate Director for Research, Richard R. Peterson, Ph.D.: ·
Comparing the Processing of Domestic Violence Cases to Non-Domestic Violence Cases in New York City Criminal Courts, CJA (December 2001).
·
Cross-Borough Differences in the Processing of Domestic Violence Cases in New York City Criminal Courts, CJA (July 2002).
·
The Impact of Case Processing on Re-arrests Among Domestic Violence Offenders in New York City, (March 2003).
The full reports (and Research Briefs published to date) are available on our web site: www.nycja.org/research/research.htm
The New York City Criminal Justice Agency 52 Duane Street New York, NY 10007
TO: