This research project was designed as a
replication and extension of earlier research on how childhood
victimization relates to delinquency, adult criminality, and violent
criminal behavior (CHILD ABUSE, NEGLECT, AND VIOLENT CRIMINAL BEHAVIOR
IN A MIDWEST METROPOLITAN AREA OF THE UNITED STATES, 1967-1988 (ICPSR
9480)). This study had four major goals: (1) to document the
prevalence of delinquency, adult criminality, and violence in a new
cohort of abused and/or neglected children and matched controls,
representing a different geographic area (Northwest), time period
(1980s), and ethnic composition (to include Native American youth) as
compared to earlier studies, (2) to examine the extent to which there
are gender and ethnic differences in the relationship between
childhood victimization and crime and violent offending, (3) to
determine the extent to which different types of maltreatment
(physical and sexual abuse and neglect) are associated with increased
risk of subsequent delinquent, adult, and violent criminal behavior,
and (4) to examine the extent to which placement experiences mediate
delinquent and criminal consequences.
This study consisted of a sample of substantiated
cases of abused and neglected children who were made dependents of the
Superior Court of a large urban county in the Northwest between 1980
and 1984, and a matched control group of children. Control match
criteria were collected from Department of Health birth records
data. Dependency records were obtained from the county court
house. Type of abuse/neglect precipitating the dependency petition was
coded using a modified version of the Maltreatment Classification
Coding Scheme (MCS). The MCS provides a mechanism for classifying
maltreatment by subtypes with up to six levels of severity for each
maltreatment allegation type. Data on juvenile arrests from juvenile
court records, including both number and types, were collected for
each abused and/or neglected youth and each matched control
subject. Adult criminal arrests, excluding routine traffic offenses,
for all abused and neglected subjects and matched controls were
collected from local, county, state, and federal law enforcement
sources. Each jurisdiction has its own coding system for crimes, so a
coding scheme was developed for this study to allow arrest comparisons
across jurisdictions. Duplicate charges were removed from the dataset
through individual examination of each case and comparison to charges
from each data source. It is possible that some crimes were not
attributed to an individual due to matching issues, but this was
minimized by a thorough examination of each name, other identifying
information, and charge.
All dependency petitions filed between 1980 and 1984 on
children aged from birth to 11 years were included in the initial
sample. A total of 2,262 dependency petitions were identified as
meeting the study criteria. After initial examination, 187 cases were
excluded because the case was transferred out of the area, the child
died, or the dependency record was not available. Of the remaining
2,075 dependent children, 1,198 were excluded from the study because
they either were not born in the state, were not made a dependent in
the county of interest, or their dependency record was not
located. This left a final group of 877 abused and neglected
children. To locate a matched control group, a file of all youth born
within the jurisdiction of interest in the same time period as the
abuse and neglect group was obtained from the Department of Health
(DOH). DOH birth records were searched until each abuse and/or neglect
case in the study had a corresponding matched control based on age
(within three months), gender, ethnicity, and approximate
socioeconomic status. Approximate socioeconomic status matching for
the controls was based on birth in the same or similar census tract as
the abuse and neglect group. Census tracts were clustered based on
median family income, percent on welfare, percent persons below
poverty level, percent persons over 25 with a high school diploma, and
percent single female head of household. Once a "matched pool" was
identified, child abuse and neglect registry records were reviewed in
order to exclude any of the matched controls found to have a child
abuse and neglect report history from the sample pool. These
procedures produced a one-to-one matched control group of 877 cases.
All children aged from birth to 11 years during the period
1980-1984 in a large urban county in the Northwest.
Individuals.
Data were obtained from county court house dependency
records for January 1, 1980, through December 31, 1984, from
Department of Health birth records data for January 1, 1969, through
December 31, 1984, and from the arrest records of local, county,
state, and federal law enforcement offices for January 1, 1980,
through December 31, 1997. Socioeconomic data were obtained from the
United States Census Bureau.
administrative records data and census/enumeration data
Demographic variables for all subjects include
gender, ethnicity, and age on March 1, 1999. Criminal record variables
include age at first offense, number of juvenile and adult offenses,
number of charges at different jurisdictional levels, and number and
type of offenses. A variable for violent crimes was created and
provides counts of arrests for attempted murder, rape and sexual
assault, kidnapping, homicide, assault, child abuse, burglary and
trespassing, theft and robbery, and vehicular homicide. Dependency
record variables, which only apply to subjects abused/neglected as
children, include age at dependency, reason for dependency, type of
maltreatment, type of order at dependency, reason for resolution of
dependency, final disposition, length of dependency, caregiver's
relationship to child, caregiver problems at dependency, identified
child problems at dependency, whether a child was placed as a result
of the dependency, type of placement, length of placement, and whether
the child remained in placement at the resolution of dependency.
Variables based on MCS are included for several types of maltreatment,
and their values reflect the severity of the maltreatment. Types of
maltreatment included are physical abuse, which is divided into eight
subtypes according to the location of the abuse on the child's body,
emotional abuse, which is divided into 26 subtypes, sexual abuse,
neglect, which is divided into eight subtypes, including failure to
provide food, clothing, shelter, medical treatment, or hygiene, lack
of supervision, a poor supervision environment, educational
maltreatment, use of an unsuitable substitute caregiver, involving
child in illegal activities, and whether the child used
drugs/alcohol. Census variables included are census tract in which
child was born, population of the census tract, percent of the
population under 14 years, percent of households with a single female
head, percent of single female heads of household with children,
percent of families below the poverty line, percent of African
Americans below the poverty line, percent of Native Americans below
the poverty line, percent of Asians below the poverty line, mean and
median family income, percent of families receiving federal
assistance, percent of people over 25 with a high school diploma, and
socioeconomic status of the census tract. A number of derived
variables are also included.
Not applicable.
The study used the Maltreatment Classification Coding
Scheme (MCS) developed by Barnett, Manly, and Ciccheti.