No Slide Title

Download Report

Transcript No Slide Title

1
Research Synthesis on Child Welfare
Disproportionality and Disparities:
Child Welfare Entries
John Fluke, Ph.D.
Child Protection Research Center
Child Welfare
American Humane Association
Race & Child Welfare: Disproportionality, Disparity, Discrimination:
Re-Assessing the Facts, Re-Thinking the Policy Options -Working Conference
Harvard Law School , Cambridge, MA
January 28, 2011
Acknowledgements
•
Supported by:
– Center for the Study of Social Policy
– Annie E. Casey Foundation
•
Co-authors:
– Brenda Jones Harden, Ph.D.
• Department of Human Development, University of Maryland
– Molly Jenkins, M.S.W.
– Ashleigh Ruehrdanz
• Child Protection Research Center, Child Welfare, American
Humane Association
•
This presentation represents the presenter’s perspective on the topic, and
not necessarily the views of the co-authors.
3
Overview
• The Review and Definitions
• Explanatory Framework
• Incidence
• Enumerating Entries
• Explaining Entries
• What Research is Needed?
Review Scope
• Roughly based on research since 2006
• Intention was to be objective
• Peer review and fugitive literature
• Review contains over 200 references
• Over 400 documents were considered for
inclusion
• Definitions
– Disproportionality
– Disparity
– Discrimination
Disproportionality and Disparity (Myers, 2010)
• Disproportionality is the ratio of the percent of persons of a certain
race or ethnicity in a target population (e.g., children who are
substantiated for maltreatment) to the percentage of persons of
the same group in a reference (or base) population.
• Disparity is the comparison of the ratio of one race or ethnic group
in an event to the representation of another race or ethnic group
who experienced the same event. A disparity exists when the ratios
being compared are not equal.
• Discrimination is the unequal treatment of identically situated
groups. Such groups can be considered identical with respect to the
most important features related to the situation being analyzed;
however, being able to define circumstances as perfectly identical
is probably not possible. Nevertheless, the identification of
disparities may be the first step in identifying possible
discrimination.
Explanatory Framework
RACIAL BIAS/DISCRIMINATION
•Reporter Bias
•Caseworker Bias
•Cultural Sensitivity and
Competence
•Institutional Racism
SYSTEM PROCESSES AND
RESOURCES
•Lack of Agency Resources
•Caseworker Factors
•Lack of Culturally Appropriate
Services
•Caseworker/family race
matching
•Multiplicative processes
FAMILY RISKS
GEOGRAPHICAL CONTEXT
•Poverty
•Substance Abuse
•Single Parenthood
•Mental Health Challenges
•Intimate Partner Violence
•Parent Incarceration
•Child Age and Functioning
•Neighborhood Segregation
•Neighborhood Economic and
Social Resources
•Concentration of poverty
•Jurisdiction-specific child welfare
policy and practice
Racial
Disproportionality
and Disparity in
Child Welfare
8
Incidence/Prevalence Studies
• Dietz (2002)
– Household survey /Conflict Tactics Scale
– Differences (bi-variate and multi-variate) in harsh
disciplinary practices between African American and
White caregivers
• Finkelhor, et al. (2005)
– Household survey/Juvenile Victims Questionnaire
(multiple forms of victimization)
– No bi-variate differences between White, African
American, and Hispanics
– Very preliminary findings at this stage
9
Incidence/Prevalence Studies
• Westat, NIS
– Applies to sentinels (professionals in contact with
children) and Child Protective Services
– NIS 3 and Prior NIS indicated no difference in race
ethnicity
– NIS 4 finds differences across overall maltreatment
and several specific categories comparing Whites
and African Americans (no differences between
Whites and Hispanics)
– Explanations
• Income gaps
• Sample precision
• Multivariate SES (> differences in not low SES), Family
Structure
10
Where does this leave us?
•
Existence or Absence of Disparities in Incidence and Prevalence Remain
Unclear
– Methods are a problem
– Some populations are not addressed
– Tendency of research is that some disparities do exist for African
Americans
• Causes are subtle and explained to some extent, but
research is very weak
•
Relationship to Child Welfare
– If disparities in incidence exist for African Americans, then we might
expect more African American children to be in the system regardless
of what the system is doing
– Underlines the importance of comparing differences using decision
point based denominators
11
Incidence?
12
Enumerating Disparities and Disproportionality in
Decision Points
• Population Based Denominator Ratios
– Based on data from one child welfare
decision (e.g., new placements/population)
– Easiest to obtain
• Decision Based Denominator Ratios
– Based on data from at least two child welfare
decisions (e.g., new placements/opened
cases)
• Relationship
Population Based Denominatork = e( ∑ ln(Decision Based Denominatori))
Comparison of Populationa and Decision Based Diparity Ratios - Colorado 2009 African American Children with Respect to White Children
2.69
2.34
2.34
2.42
2.16
1.35
1.03
0.89
Population Denominator
Decision Denominator
Explaining Removals at the Worker Level
(DME/GADM) (Baumann, et al. 2010)
Disparity Index
-
Percent on
Caseload
-
Community
Resources
(worker)
Skills (AA only)
+
AA
Only
-
Workload
(worker)
Removals
+
Family Poverty
+
Family Risk
Level
Explaining Removals
 The relationship between the case factors risk, race, and poverty
may be difficult for caseworkers to understand because they are
intertwined.
 The perception of lower interpersonal skills, an individual factor,
is related to greater disparities in the removal of African
American children.
 Having fewer African American or Hispanic families on one’s
caseload, an organizational factor, is associated with greater
disparities in the removal of African American or Hispanic
children.
 Removals themselves are increased when the caseworker
believes the services in the areas in which they work to be
inadequate (organizational). Removals are decreased when the
worker believes their caseloads are too high (organizational).
16
Summary: Many aspects of the issue need to be
better understood from research
• Incidence/prevalence
• Absence of key groups (e.g., American Indians, Hispanics)
• Issues with race coding and enumeration particularly in secondary
analyses
• For entries Issues of Use of Population or Decision Based
Denominators
• Most of what we know is emergent, not planned research.
• Few multivariate and multi-level analysis of explanatory factors
• No experiments, except vignettes
• Formulation of hypotheses:
– how and when disparities are inequities (either under or over
served)
– how to effectively reduce inequities
– Impact of resource gaps and the capacities of communities