The purpose of this study was to provide a
definitive test of whether batterer programs and varying intensities
of judicial monitoring reduce reoffending among domestic violence
offenders. The study was designed to ensure that the effects of
batterer programs and judicial monitoring could be distinguished, so
that the impact of one practice would not be mistakenly attributed to
the other. The study sought to answer three primary questions:
Do batterer programs work?
Is judicial monitoring more effective if implemented in a more
flexible or "graduated" manner?
Is judicial monitoring more effective than the complete
absence thereof?
This study was designed to distinguish the effects
of batterer programs and judicial monitoring on the future violent
behavior of domestic violence offenders. In Part 1, Batterer Program
Experiment Data, convicted male domestic violence offenders were
randomly assigned to four experimental conditions. Study enrollment
took place between July 23, 2002, through February 27, 2004. The
process of random assignment began once the prosecutor and defense
attorney in court parts AP10 (Pretrial Appearances) or TAP2 (Trials)
indicated that they were prepared to accept a plea to a violation and
a sentence of a conditional discharge, order of protection, and a
batterer program with monthly monitoring (the preexisting status
quo). Provided that the judge did not choose to exclude the case from
the study, the judge then read from a standard study allocation script
that the judges in the domestic violence court parts developed
themselves. The script described in detail each of the four groups to
which the defendant could be sentenced. While the original script was
a two-page, five-minute long explanation, it was abbreviated over
time, with each judge making personal revisions. If the defendant
agreed to accept the offer at this point, understanding that there
were four possible sentencing outcomes, a bench conference was
held. During this conference, either the resource coordinator or the
project senior research associate performed the random assignment by
shaking and then removing from a cup one of four pieces of folded
paper on which were written numbers 1 through 4, representing the four
study groups. The four conditions were defined as:
Batterer program plus monthly judicial
monitoring (n = 102): Offenders in this group were assigned to attend
a batterer program and to make monthly appearances before the Judicial
Hearing Officer (JHO) in court part post-sentencing compliance
monitoring (DVC). The JHO monitored the defendant's progress in the
program, compliance with orders of protection, avoidance of
rearrests, and compliance with other conditions imposed by the
sentencing judge.
Batterer program plus graduated monitoring (n = 100):
Offenders in this group were assigned to attend a batterer program.
They were also required to appear before the JHO, but on a graduated
schedule. Those in compliance throughout the monitoring period were
required to report to DVC at four weeks post-sentence for their
initial appearance, and then at 10 weeks, 18 weeks, and 26 weeks.
Those out of compliance at any time were required to reappear in DVC
two weeks and then four weeks after the determination of noncompliance.
They were then returned to the graduated schedule.
Monthly monitoring only (n = 109): Offenders in this group
were required to appear monthly before the JHO, as in the first group,
but were not required to attend a batterer program. The JHO monitored
attendance at scheduled court appearances and compliance with all
existing orders of protection, and would respond to any serious
infraction or rearrest.
Graduated monitoring only (n = 109): As with offenders in
the second group, offenders in this group were required to appear for
graduated monitoring before the JHO, but were not required to attend a
batterer program. Also, as with the third group, compliance monitoring
focused on attendance at scheduled court appearances, compliance with
orders of protection, and avoidance of rearrests.
Before the defense attorney and defendant left the courtroom, the
resource coordinator or research associate gave each a sheet,
available in English or Spanish, summarizing the offender's
responsibilities. After receiving other paperwork from the court,
offenders assigned to the two groups required to attend a batterer
program proceeded to the defendant monitors' office. One of the
defendant monitors conducted a brief intake interview and offenders
assigned to attend a batterer program (groups 1 and 2) were mandated
to either the Domestic Violence Accountability Program (DVAP) run by
Safe Horizon or the Men's Choices Program run by the Fordham Tremont
Community Mental Health Center. The research associate also created a
research case file including the date of the first scheduled
appearance in DVC, case and offender identifiers, charge information,
disposition, sentencing date, experimental assignment, victim name,
and victim contact information. The offenders were tracked for at
least twelve months after sentencing, and for up to eighteen months
for most of the men, to determine whether they fulfilled the
conditions of their sentence, were rearrested for domestic violence,
or were reported by the victim to have engaged in new incidents of
abuse. Using each offender's New York State criminal identification
number, complete criminal record files, including prior criminal
history and recidivism, were obtained from the New York State Division
of Criminal Justice Services (DCJS). The DCJS records did not indicate
whether arrests in the criminal record involve domestic violence, so
this information was obtained separately, with the assistance of staff
from the Office of the Chief Administrative Judge of New York City.
Victims were interviewed about new domestic incidents committed
within one year of sentencing. Research staff at Safe Horizon, the
local victim services organization, were trained in a daylong
orientation sessions on how to interact with victims and protect
victim safety. Interviewers were instructed not to leave messages on
answering machines and to ensure that victims were in a safe
environment before asking them to complete the 10-minute interview.
Victims who completed telephone interviews were given 25-dollar
stipends. In cases where victims could not be reached by phone, Safe
Horizon staff attempted home visits to encourage victims to call
research staff for an interview. When that procedure yielded little in
the way of additional interviews, it was discarded in favor of sending
letters to unreachable victims, offering a 50-dollar incentive for
victims to participate in interviews.
In Part 2, a quasi-experimental study was used to compare
recidivism outcomes between the randomized offenders in Part 1 and a
control group of offenders convicted of identical offenses, but who,
as a result of the normal sentencing process, received neither a
batterer program nor any form of monitoring. Researchers identified
712 conditional discharge (CD) only cases sentenced in the court part
AP10 or TAP2 during the same time period as Part 1. This regression
yielded a propensity score for each offender. The propensity score
predicted the probability of inclusion in the randomized trial (Part
1) sample, based not on actual membership in that sample, but on the
statistical probability of membership in it, as computed from the
observed set of background characteristics. Each offender in the
randomized trial was then matched to the offender in the CD only group
with the nearest propensity score. Sometimes multiple offenders from
the the initial trial were matched to the same CD only offender. This
method lead to 387 offenders from the randomized trial being matched
to 219 CD only offenders.
For Part 1, eligible offenders were convicted of a
violation and sentenced to a conditional discharge in court parts AP10
(Pretrial Appearances) or TAP2 (Trials) of the Bronx Misdemeanor
Domestic Violence Court from July 23, 2002 through February 27, 2004,
with acceptance of a batterer program condition by both prosecution
and defense prior to randomization. The sentencing judge retained the
discretion to exclude particular offenders from randomization. One
year after the offender's sentencing, interviewers attempted to
establish phone contact with all victims except those who those who
had earlier indicated to Safe Horizon that they did not want to be
contacted. Attempts to contact each eligible victim were made at least
five times or more, with at least three attempts coming during evening
or weekend hours. For Part 2, of the 712 conditional discharge (CD)
cases, 113 were immediately discarded because they did not involve an
intimate partner relationship. Reseachers performed t test comparing
the experimental and CD only samples on all known background
characteristics to determine, what, if any, significant differences
existed. Then all characteristics on which differences existed at the
.10 level or better were entered into a logistic regression, for which
the the dependent variable was sample membership. This regression
yielded a propensity score for each offender. The propensity score
predicted the probability of inclusion in the randomized trial (Part
1) sample, based not on actual membership in that sample, but on the
statistical probability of membership in it, as computed from the
observed set of background characteristics.
All offenders convicted of a violation sentenced to
conditional discharge in court parts AP10 (Pretrial Appearances) or
TAP2 (Trials) of the Bronx Misdemeanor Domestic Violence Court from
July 23, 2002, through February 27, 2004, and their victims.
individual
Data for this collection were obtained from the New
York State Division of Criminal Justice Services (DCJS), the Office of
the Chief Administrative Judge of New York City, and telephone
interviews with victims of domestic violence.
administrative records data, and survey data
experimental data
record abstracts
telephone interview
Both Part 1 and Part 2 contain defendant
demographic and arrest variables. Demographic variables include
defendant age, gender, race, and relationship to the victim. Other
demographic variables include living situation, length of time at
current residence, education level, if employed, and length of time at
current employment. Arrest variables include date of arrest, sentence
type, injury to the victim, if the defendant was assigned to a
batterer program, and type of judicial monitoring. Variables on the
batterer program include which program the defendant was assigned to
and date of program termination. Variables on judicial monitoring
include the number of court appearances, dates of court appearances,
status at each court appearance, and length of monitoring. Also
included are variables on the number of warrants issued during the
monitoring program, the number and dates of noncompliant events, and
if the defendant was resentenced. Variables on the number and type of
prior arrests, and case processing time are included. Both Parts 1 and
2 also contain variables from the victim interviews. Demographic
variables, such as relationship to defendant, date of birth, education
level, source of income, place of birth, and race are included. Other
variables from the victim interviews include how often the victim saw
the defendant at the time of arrest, if the victim wanted the
defendant arrested, what the victim wanted instead of arrest, if the
Judge issued an order of protection, and if victim wanted the order of
protection. Other variables asked about the effect of the sentence on
the relationship between the victim and the defendant, if the victim
has had any contact with the defendant since the arrest, and if the
defendant had been violent in any way with the victim since the arrest
and sentencing.
During the sampling period for Part 1, 489
defendants were identified as eligible for the study. Judges used
their discretion to exclude 69 (14 percent) of this eligible pool,
resulting in a final sample of 420 offenders. The contact rate for
victim interviews was 25 percent (n = 106).
Several Likert-type scales were used, as well as Straus'
(1979) Conflict Tactics Scale