Responding to the scarcity of published research
about advocacy services for battered women, this study evaluated
advocacy services offered to battered women in Detroit, Michigan, and
examined other aspects of coordinated community responses to domestic
violence by focusing on women named as victims in police reports.
Advocacy was defined as those services provided to support victims
during the legal process or to enhance their safety. The study sought
to answer the following questions: (1) Is advocacy at the precinct
and/or prosecutor's level associated with a higher rate of completed
prosecution of batterers? (2) Is advocacy at the precinct and/or
prosecutor's level associated with a higher rate of guilty findings
against batterers (or pleas of guilty)? (3) Does advocacy at the
precinct and/or prosecutor's level affect victims' reports of rates of
subsequent violence as well as rates of police reports and arrests?
(4) Do victims who received advocacy assess their situation as more or
less safe than victims who did not receive it? (5) How do victims view
advocacy and the criminal justice process related to the abuse?
For the Preliminary Complaint Reports Data (Part
1), a random sample of preliminary complaint reports (PCRs) was
gathered. A research assistant went weekly to the five precincts and
collected copies of all the PCRs that officers set aside as "domestic
violence incidents." She then sorted them and removed cases that did
not meet the criteria of having an adult female victim meeting
Michigan's definition of domestic violence. According to Michigan law,
an incident is domestic violence when the perpetrator and victim are
married, formerly married, live-together partners, formerly lived
together, or have a child in common. Researchers had a quota of cases
for each precinct, so if the research assistant had too many from a
single precinct, she made a selection by counting them and selecting a
case when her count matched the last two digits of a table of random
numbers. This quota enabled the researchers to be certain to gather
reports from all of the precincts throughout the data collection
period rather than gathering all the reports researchers needed from
one precinct in a short period of time. If the number of reports did
not exceed the quota, researchers used all the domestic violence
reports from the precinct for the week. A total of 1,057 cases were
selected. The researchers continued to collect PCRs from the five
precincts through March 1999 to look for additional reports pertaining
to the 1,057 victims in the sample. The subsequent PCRs were coded
according to whether the perpetrator was the same or different from
the initial incident perpetrator. For Victim Advocacy Contact Data
(Part 2), researchers obtained data from advocates' files about the
services they provided to the 1,057 victims. During the planning
stages of the proposal, the advocacy programs agreed to provide access
to advocates' records about the numbers of face-to-face and telephone
contacts they had with each victim as well as the types of services
provided. However, as the data gathering began, researchers learned
that some advocates did not keep methodical records on all of the
services they delivered to each victim. Others were apparently
unwilling to allow the researchers to have access to these data.
Therefore, researchers devised a "contact form" for advocates to
fill out after each in-person or phone contact with a victim. After
the intake of PCRs was completed, researchers compiled a list of the
1,057 names with the birth date or age of each victim. They then asked
the police department advocates and legal advocates to check the name
of each victim they served against this list and fill out a contact
form for new contacts with women on the list. While advocates reported
that they were following this procedure, it was impossible to know how
consistently they were checking the list. For Case Disposition Data
(Part 3), researchers created a prosecution outcome form and produced
a copy for each case when the prosecutor's advocate's logbook
indicated a warrant request. Researchers then asked the prosecutor's
advocates to fill out a form for cases that were assigned to them on
their intake days. Researchers collected a few completed forms, but
the prosecutor's advocates did not fill out the forms for most of the
cases that were completed. Therefore, researchers conducted a computer
search for the outcomes of the cases. Researchers looked up each
perpetrator from the list of 1,057 incidents, and determined whether
there was a warrant for the focal incident, whether it turned into a
prosecution, and the outcome. Although the computer search did not
yield a report that was as detailed as the disposition forms that
researchers gave the advocates, it did provide the outcome of the
case, and it gave some information about why cases were dismissed.
Therefore, researchers were able to categorize cases according to
whether they were dismissed for lack of evidence or lack of a
complaining witness. The Initial Victim Interview (Part 4) and
Follow-Up Victim Interview (Part 5) were conducted from April 1998 to
July 1999. Before the initial interview, a passive consent (refusal)
form was used in precincts by domestic violence counselors and by the
prosecutor's advocates. Advocates were asked to describe the study
when they met with the victims. They explained the survey and had
women sign the refusal form if they were unwilling to be contacted for
the survey. Two questionnaires, one for the initial telephone
interview and the other for the follow-up, were developed. At the
beginning of each interview, interviewers read the consent form to the
respondents and informed them of potential risks and benefits. Victims
were contacted again three months after the first interview to make
sure they were still at the same telephone number and to ask them to
notify the interviewers if their number changed. For the six-month
interviews, all of the women who were reached at three months were
called, except the few who had refused further contact. Researchers
also attempted to contact women who were initially interviewed but
could not be reached with the three-month call. During the same period
that researchers were completing the second interviews, they also
interviewed 23 women (Victim Comparison Group Interview, Part 6) from
the list of 1,057 whom they had been unable to reach during the first
interviews. They compared these 23 women to the 63 who had second
interviews to determine if there were any differences in use of
services, or views toward or participation in prosecution.
Random sampling.
Women named as domestic violence victims in police
incident reports in Detroit, Michigan.
individual
administrative records, and telephone interviews
survey data, and administrative records data
Variables in Part 1 focus on whether alcohol and
abuse were involved, previous incidents, the suspect's psychological
aggressions and physical assaults, if a weapon was used, if the victim
was hurt, if property was damaged, if the victim sought medical
attention, and the severity of physical abuse or injury. Variables in
Part 2 provide information on the role of the advocate, methods of
contact, types of referrals made, and services provided. Variables in
Part 3 include the type of charge, outcome of resolved case, why the
case was dismissed, if applicable, and if the suspect was sentenced to
probation, costs, confinement, no contact with the victim, a batterer
program, or community service. The initial, follow-up, and comparison
group interviews (Parts 4-6) all collected similar information.
Variables about the incident include how well the respondent
remembered the incident, if police arrived promptly, if the respondent
was advised to file charges, if police told the respondent that a
counselor was available, and if the respondent's partner had been in
jail since the incident. Variables concerning advocacy include whether
the victim contacted advocates, and if advocates provided legal help
and referrals. Legal system variables include whether the respondent
felt pressured by anyone to drop charges or pursue charges, if the
respondent received help for preliminary examination or trial, and if
contact with the legal system helped the respondent. Variables about
services include whether the respondent received assistance in
temporary shelter, food/money resources, child care, employment,
education, a lawyer for divorce/custody, support or self-help group,
or a substance abuse treatment program. Variables concerning what
happened in the previous six months cover the number of times the
respondent had called police because of danger, left home because of a
violent incident, partner had been arrested because of violence, and
partner physically abused respondent. Variables about events that
occurred while the respondent and abuser were separated include how
often the partner harassed the respondent on the phone, wrote
threatening letters, violated legal restrictions, refused to leave the
respondent's home, failed to pay child support, and threatened to take
the children. Demographic variables include respondent's race or
ethnic background, education, marital status, number of children
respondent had, number of children who lived with respondent, and
employment status and income at the time of the interviews.
Initial interview: 242 initial interviews were
completed from the list of 1,057 selected from the PCRs, which was a
response rate of 22.8 percent. One hundred and ninety of the phone
numbers were disconnected, and 182 were numbers for homes with no
resident females or where the victim listed on the PCR denied that
the incident had happened. Eighty-one women refused to be
interviewed. Three-month follow-up: Out of 242 women who were
interviewed initially, phone interviewers were able to reach 153
women for the three-month follow-up. They were able to make an
appointment for the six-month interview with 126 of them. During the
three-month calls, nine women refused to be contacted for the
six-month interview. At 18 of the numbers, interviewers were told
that the respondent was not there or it was the wrong locale. Of the
remainder of the telephone numbers, 62 were not in service. Six-month
interview: 63 interviews were completed. There were 67 telephone
numbers that were not in service. Twenty-four women refused the
second interview, 34 were the wrong locale or no respondent, and
three claimed to be ineligible or that there was no incident that
occurred on the PCR date.
Several Likert-type scales were used.