Despite the current proliferation of intimate
violence studies, domestic violence advocates and policymakers in
public health and criminal justice are often confused about the
efficacy of practical interventions. Under what circumstances is a
woman at risk if she terminates an abusive relationship? In what
situations does arrest increase or decrease the risk of death? How do
stalking and other forms of harassment interact with events and
changing circumstances, such as gun ownership or pregnancy, in
affecting the risk of a lethal outcome for a victim or offender? The
Chicago Women's Health Risk Study was conceived to respond to the
increasing need for information that would be vital to medical, public
health, and criminal justice agencies, and to domestic violence
advocates and educators in building public health and public safety
strategies. To help this array of practitioners identify and
effectively intervene in potentially life-threatening intimate
violence situations, the goal of the CWHRS was to develop a reliable
and validated profile of risk factors directly related to lethal
outcomes in intimate violence, for use in agencies and organizations
working to help women in abusive relationships. Data were collected to
draw comparisons between abused women in situations resulting in fatal
outcomes and those without fatal outcomes, as well as baseline
comparisons between abused women and non-abused women, while taking
into account the interaction of numerous events, circumstances, and
interventions occurring over the course of a year or two.
The CWHRS used a quasi-experimental design to
gather survey data on 705 women at the point of service for any kind
of treatment (related to abuse or not) sought at one of four medical
sites serving populations in areas with high rates of intimate partner
homicide (Chicago Women's Health Center, Cook County Hospital, Erie
Family Health Center, and Roseland Public Health Center). Over 2,600
women patients were randomly screened for abuse using a screening
questionnaire based on the Intimate Violence Screening Tool. The
screener determined if a woman had a history of physical abuse by an
intimate partner in the past year and whether she wanted to
participate in the study. Although the format of the screening
differed slightly for each site, it always included three questions:
(1) "Has your intimate partner ever hit, slapped, kicked, or otherwise
physically hurt or threatened you?" (2) "Has your intimate partner
ever forced you to engage in sexual activities that made you
uncomfortable?" and (3) "Are you afraid of your intimate partner?" A
woman was categorized as "screened AW (abused woman)" if she answered
positively to any of the three questions and met all of the following
criteria: the abuse took place within the past year, the abuser was an
intimate partner (regardless of gender), and the woman answering was
at least 18 years old. Women answering "no" to all questions who were
at least 18 years old and who had had an intimate relationship in the
past year were categorized as "screened NAW (non-abused woman)." From
the women screened, 705 valid initial interviews were conducted, in
which supportive and understanding interviewers asked more sensitive
questions about the violence. After the interview, 497 women were
categorized as AW (they had experienced violence or a violent threat
from an intimate partner in the past year), and 208 women were
categorized as NAW. Data were collected through face-to-face
interviews conducted during 1997-1998 with the women. To ensure
anonymity and increase safety and security during the interview
process, each initial interview took place in a private and secure
room, behind a closed door, at the hospital, clinic, or health
center. The interviews, lasting an average of 45 minutes, were
conducted in both Spanish and English and touched upon details of each
abusive incident (for the abused women), harassment and stalking,
social support network, help-seeking and interventions, household
composition, mental and physical health, pregnancy, and firearm
availability. The interviewer and each abused woman worked together to
complete a 12-month retrospective "calendar history," highlighting key
events and each violent incident that had happened in the woman's life
during the past year. The data instrument was a set of blank calendars
printed with one or two months per page. First, the woman was asked to
place holidays and important life events on the calendar (e.g., having
a baby, changing jobs, or moving). The woman and the interviewer also
placed on the calendar each violent incident at the specific day or
closest approximate day of occurrence. (If a woman told an interviewer
about a violent incident that was not committed by an intimate
partner, the incident was recorded in the event calendar dataset.) If
a woman had been abused by more than one intimate partner in the
previous year, interviewers recorded each separate incident on the
calendar. Each woman received ten dollars at the initial interview as
a thank-you for her participation in the study. Data from the initial
interview sections comprise Parts 1-8. To collect prospective data on
each of the 497 women who interviewed as AW, each woman was asked, at
the end of the initial interview, for her consent to be reinterviewed
in a series of follow-up interviews. If a woman consented, she was
asked to provide one or more contacts where she might be
reached. Researchers conducted at least one follow-up interview with
approximately 323 of the 497 abused women during 1998 and 1999, via
personal and telephone interviews. Each respondent was given 20
dollars for each follow-up interview as a thank-you for her continued
participation. The follow-up interview schedule was similar to the
initial interview. In addition to the questionnaire, women who had
experienced any violence at the hand of an intimate partner since the
initial interview completed a follow-up calendar history, covering the
period since the initial interview. At the second follow-up interview,
women who had experienced any violence completed a follow-up calendar
history for the period since the first follow-up interview. In
homicides in which a woman was the homicide offender, attempts were
made to contact and interview her. Every intimate homicide that
occurred in Chicago in 1995 or 1996 involving a woman over the age of
17 was included, regardless of whether further investigation
determined the homicide to be justified. This "lethal" sample, all
such homicides that took place in 1995 or 1996, was developed from two
sources, HOMICIDES IN CHICAGO, 1965-1995 (ICPSR 6399) and the Cook
County Medical Examiner's Office. The lethal sample contained 87
homicides, including 57 women victim cases, 28 women offender cases,
and two same-sex cases. Additional data for the lethal sample were
collected using official records from the Chicago Police Department
Murder Analysis Reports, Cook County Criminal Court records, and
newspaper articles from the Chicago Tribune archive, and the Chicago
Sun-Times and Chicago Defender microfiche archives (Parts
13-15). Researchers wanted to locate individuals who could serve as
knowledgeable and credible proxy respondents about the relationship
between the victim and the offender in the year prior to the
homicide. They began with the names and addresses of relatives who had
identified the body provided from the Medical Examiner files, and the
names of witnesses provided by the Cook County Criminal Court records,
and used three primary strategies to develop the proxy contacts.
First, researchers contacted potential proxies by phone using numbers
listed in the case files, current telephone books, the HAYNES
CRISS-CROSS DIRECTORY at the Harold Washington Public Library, and
Internet directories. Next, they mailed letters to family members
introducing the study and inviting participation as potential
proxies. Third, they made field visits to the homicide sites,
particularly to neighbors of the couple who might provide potential
proxy contact information. Researchers were able to locate and
interview at least one proxy respondent for all but 11 of the 87
homicide cases. It took as many as three proxy interviews (conducted
face-to-face and via telephone) to compile the information necessary
to complete a case. Researchers were able to obtain permission from
the Illinois Department of Corrections to interview women homicide
offenders still in prison when data collection began. Each of the
proxy respondents completed, as far as possible, the same
questionnaire used for the 705 women sampled in the clinic/hospital
settings.
Convenience and random sampling.
Parts 1-7: Non-abused and abused women living in Chicago,
IL. Parts 8-12: Abused women living in Chicago, IL. Parts 13-15:
Intimate partner homicide cases in Chicago, IL, in 1995-1996.
Parts 1-8, 11, and 12: Individuals. Parts 9 and 10:
Live events and physical abuse incidents. Parts 13 and 14: Intimate
partner homicide incidents. (Part 14 includes only those incidents in
which there was at least one proxy or offender interview.) Part 15:
the individual woman victim or woman offender (woman victim in the two
same-sex cases) in each homicide.
personal interviews, telephone interviews, and official
records
survey data, aggregate data, and administrative records
data
Parts 1-12 comprise data from the clinic/hospital
sample and contain a unique identifier variable that allows users to
link each respondent across files. Part 1 includes demographic
variables describing each woman, such as age, race and ethnicity,
level of education, employment status, screening status (AW or NAW),
birthplace, and marital status. It also includes information
(relationship type and length) on the woman's intimate partner, or if
there was more than one, the partner the woman felt "closest to."
Variables in Part 2 describe the woman's living arrangements at the
initial interview and the previous year, including whether she was
homeless or living in a treatment center or shelter, the number of
people living in the household, the number of the woman's children
living in and outside the household, whether the children were the
woman's birth, adopted, or foster children, the relationship, age, and
gender of up to ten children (the woman's and others) living in and
outside the household, and the age, gender, and relationship of up to
ten adults living in the household. Any changes to the household
(e.g., someone going to jail or a baby being born) and the dates of
each change are included as well. Variables in Part 3 deal with the
respondent's physical and mental health, including general well-being,
type and duration of any physical or emotional limiting condition,
amount of bodily pain experienced, pregnancy and pregnancy outcome in
the past year, alcohol or drug use and treatment, a four-item Medical
Outcomes Study (MOS) scale of depression (Hays et al., 1995, Stewart
et al., 1988), a scale of the strength of the woman's social support
network (acceptance and support, knowledge of and access to resources,
and tangible help in emergencies), and measures of other resources,
such as personal and household incomes. Variables in Part 4 provide
information on the number and type of guns found in the woman's home,
Scales of Power and Control (Johnson, 1996) or Stalking and Harassment
(Sheridan, 1992) by any intimate partner in the past year, specific
types of physical violence or violent threat she might have
experienced by any intimate partner in the past year (using the CTS as
modified by Johnson, 1996), and the PTSD Symptom Scale (Foa, et al.,
1993) measuring Post-Traumatic Stress Disorder related to these
incidents. Part 5 contains variables from a lengthy section of the
interview that determined which intimate partner had been responsible
for the incidents just mentioned (this may not be the partner
described in Part 1). If more than one partner was responsible for any
of the incidents, the woman was asked to choose one partner to talk
about, identified as "Name." Variables in Part 5 record the woman's
partner situation (e.g., whether the partner mentioned in Part 1 was
the only abuser, how many partners and how many abusers were
mentioned), and basic information (type and length of relationship)
about each partner mentioned in addition to the "closest" partner
discussed in Part 1. Finally, Part 5 includes a series of variables
about Name (the chosen abusive partner, or the "closest" partner for
NAW women), including Name's age, gender, and race or ethnicity, the
length of the relationship between the woman and Name, the age
difference between them, Name's employment status and education,
living arrangements with Name, whether the woman had left or ended the
relationship, tried to leave, or asked Name to leave or end the
relationship in the past year, and if so her reasons, whether she or
Name had returned to the relationship after it ended, and if so her
reasons, her children with Name and where they were living in the past
year, and whether she had children who were stepchildren to Name.
Variables in Part 6 cover the woman's help-seeking and interventions
in the past year. Women interviewed as NAW (who answered "no" to each
of the violence or violence threat questions) were not asked these
questions. Part 6 has detailed sets of questions about informal
help-seeking (talking to someone about the violence), and about three
kinds of formal help-seeking (contacting an agency or counselor,
seeking medical help, and contacting the police). A summary variable
counts the types of help mentioned. Information on each of these four
types of help-seeking includes whether the woman had tried to seek
that type of help, her reasons for doing so, what type of help she
received, and whether it was helpful. For this public-use version of
the data the narrative responses have been deleted, but the variables
with coded responses still remain. Part 6 also includes general
questions on interventions, such as jail time, probation, orders of
protection, and types of counseling or treatment involving Name, the
seriousness of incidents with Name (whether the woman felt her life
was in danger, whether Name tried to force her to do something
illegal), and coded summaries of the woman's responses to open-ended
questions about what she needed but did not get from formal types of
help, as well as her advice to other women in a similar situation.
Variables in Part 7 include questions comprising the Campbell Danger
Assessment (Campbell, 1993), such as whether Name used drugs or
alcohol, tried to control the woman's daily activities, beat her while
she was pregnant, or had ever choked or tried to strangle her. In
addition, Part 7 includes questions about Name's violence outside the
home, whether Name had ever been arrested, and if so, the type of
charge(s), and whether the woman had had any other intimate partner
who was violent toward her. Part 8 is provided to make analysis easier
for the user. It pulls together all of the information about Name (the
chosen abusing intimate partner, or for women who interviewed as NAW,
the closest intimate partner). In addition, Part 8 includes a number
of summary variables created for analysis. Variables in Part 8 include
the type and length of relationship with Name, Name's age, gender, and
racial/ethnic group, age disparity between the woman and Name, Name's
education and occupation, living arrangements with Name, whether the
woman left, or tried to leave, or ended the relationship with Name in
the past year, and if so her reasons, whether she returned to the
relationship after it ended, and if so her reasons, children with Name
and where they were living, and her children who were Name's
stepchildren. Part 9 includes information about the events women
mentioned in the 12-month retrospective "calendar history," or in
their response to interview narrative questions. Part 9 is an
event-level file, with a unique identifier linking each woman to her
set of events. Part 9 variables include a coded description of the
event, the date of the event, and for events in which the date spanned
a period of time, the beginning and ending dates. Violent incidents
not committed by an intimate partner are included as events in Part
9. Part 10 contains information about each violent incident or threat
of violence that the woman mentioned in the 12-month retrospective
calendar history. Part 10 is an incident-level file, with a unique
identifier linking each woman to her set of incidents. Variables in
Part 10 include two aggregate summaries of the number and maximum
severity of the incidents over the past year (see Part 11), as well as
information on each incident. Variables include the severity code,
date, whether a weapon was involved, whether there were other people
present, and if so, who, if anyone else was injured, whether the woman
suffered a miscarriage as a result of the incident, whether the woman
was restrained or tied down in the incident, whether Name tried to
choke or strangle her, whether the woman or Name were using drugs or
alcohol at the time of the incident, and the number of days between
the incident and the interview. Part 11 consists of aggregate
person-level variables created from the incidents in Part
10. Variables for Part 11 record the total number of incidents during
the year, the number of days before the initial interview that the
most recent incident occurred, the maximum severity of any incident in
the past year, the total number of incidents coded under each category
of severity, the number of incidents involving a weapon, a gun, or a
knife, the number of incidents resulting in miscarriage, the number of
incidents in which the woman was restrained or tied down, the number
of incidents in which other people, including children, were present,
the number of incidents in which the woman had used alcohol or drugs,
the number of incidents in which Name had used alcohol or drugs, and
the number of incidents in which both had used alcohol or drugs. Part
12 is a person-level file containing variables summarizing violent
incidents at the hands of any intimate partner (Name or any other
intimate partner) the woman experienced on follow-up. Variables in
Part 12 include the dates of each follow-up interview, the number of
days from the initial interview to each follow-up and the total number
of follow-up days, whether there had been any violent incidents since
the last interview, the total number of violent incidents during the
follow-up period, the maximum severity of the incidents at each
follow-up and overall, and the status of the follow-up (including
whether the woman consented or had died). Parts 13-15 contain data
from official records sources and information supplied by proxies for
victims of intimate partner homicides in 1995 and 1996 in
Chicago. Part 13 contains information about the homicide incidents
from the "lethal sample," along with outcomes of the court cases (if
any) from the Administrative Office of the Illinois Courts. Variables
for Part 13 include the number of victims killed in the incident, the
month and year of the incident, the gender, race, and age of both the
victim and offender, who initiated the violence, the severity of any
other violence immediately preceding the death, whether leaving the
relationship triggered the final incident, whether either partner was
invading the other's home at the time of the incident, whether
jealousy or infidelity was an issue in the final incident, whether
there was drug or alcohol use noted by witnesses, the predominant
motive of the homicide, location of the homicide, relationship of
victim to offender, type of weapon used, whether the offender
committed suicide after the homicide, whether any criminal charges
were filed, and the type of disposition and length of sentence for
that charge. Parts 14 and 15 contain data collected using the proxy
interview questionnaire (or the interview of the woman offender, if
applicable). The questionnaire used for Part 14 was identical to the
one used in the clinic sample, except for some extra questions about
the homicide incident. The data include only those 76 cases for which
at least one interview was conducted. Most variables in Part 14
pertain to the victim or the offender, regardless of gender (unless
otherwise labeled). For ease of analysis, Part 15 includes the same 76
cases as Part 14, but the variables are organized from the woman's
point of view, regardless of whether she was the victim or offender in
the homicide (for the same-sex cases, Part 15 is from the woman
victim's point of view). Parts 14 and 15 can be linked by ID
number. However, Part 14 includes five sets of variables that were
asked only from the woman's perspective in the original questionnaire:
household composition, Post-Traumatic Stress Disorder (PTSD), social
support network, personal income (as opposed to household income), and
help-seeking and intervention. To avoid redundancy, these variables
appear only in Part 14. Other variables in Part 14 cover information
about the person(s) interviewed, the victim's and offender's age, sex,
race/ethnicity, birthplace, employment status at time of death, and
level of education, a scale of the victim's and offender's severity of
physical abuse in the year prior to the death, the length of the
relationship between victim and offender, the number of children
belonging to each partner, whether either partner tried to leave
and/or asked the other to stay away, the reasons why each partner
tried to leave, the longest amount of time each partner stayed away,
whether either or both partners returned to the relationship before
the death, any known physical or emotional problems sustained by the
victim or offender, including the MOS scale of depression, drug and
alcohol use of the victim and offender, number and type of guns in the
household of the victim and offender, Scales of Power and Control
(Johnson, 1996) or Stalking and Harassment (Sheridan, 1992) by either
intimate partner in the year prior to the death, a modified version of
the Conflict Tactics Scale (CTS) (Johnson, 1996) measuring the type of
physical violence experienced by either intimate partner, and the
Campbell Danger Assessment for the victim and offender. In addition,
Part 14 contains a number of summary variables about the fatal
incident, most of which are also in Part 13. These include questions
related to the circumstances of the incident, time, place, witnesses,
who had initiated the violence, outcome for the offender (e.g.,
suicide or other death, arrest, sentence, etc.), and outcome for
children and others who witnessed the violence or found the body. Part
15 contains the same data as Part 14, except that each variable is
presented from the woman's point of view, regardless of whether she
was the victim or offender in the homicide. Additional summary
variables were added regarding the overall nature of any prior
physical abuse in the relationship, as well as the overall pattern of
leaving and returning to the relationship in the year prior to the
death.
The response rate for Part 12 (Summary of Abuse
on Follow-up Data) was 65 percent. Response rates for Parts 1-11 and
Parts 13-15 are not applicable.
Several Likert-type scales were used as well as the
Medical Outcomes Study (MOS) scale of depression (Hays et al., 1995,
Stewart et al., 1988), Scales of Power and Control (Johnson, 1996) and
Stalking and Harassment (Sheridan, 1992), a modified version of the
Conflict Tactics Scale (CTS) (Johnson, 1996), the Post-Traumatic
Stress Disorder (PTSD) Symptom Scale (Foa et al., 1993), and the
Campbell Danger Assessment (Campbell, 1993).