Predicting the Future and the Codification of Poverty
Assessing the risk of future harm to a child "is central to practice
in the field of child welfare" (McDonald & Marks, 1991). Risk assessment
is "the driving force for any child protection service delivery"
(Cuyahoga County Department of Children and Family Services). Risk assessment
consists of "the systematic collection of information to determine
the degree to which a child is likely to be abused or neglected at some
future point in time" (Doueck, et al., 1993; English & Pecora,
1994). In practice, "risk refers to the prediction of future events" (Pennsylvania Department of Public
To these ends, by the early 1990s at least 42 states had adopted one or
more of a wide variety of available risk assessment models (Berkowitz, 1991).
Complicating matters is that one state may use more than one system, or
modify an existing system. Some states use the same model for different
purposes. Of the 42 states using some form of risk assessment, 15 use a
matrix approach, 13 use a scale for entering ordinal values, and 11 use
different approaches in different areas of the state (Lyons, Doueck, &
Wodarski, 1996). Many agencies use the same instrument to assist caseworkers
in making several different types of decisions, including whether to divert
a case following substantiation, whether to remove a child during the investigation
stage, whether to place a child in foster care after adjudication, and whether
to reunify a child with her family (Wald & Woolverton, 1990). Yet, as
McDonald and Marks (1991) point out, less than half of the variables measured
by these instruments have been empirically tested, much less validated,
by even the weakest of research designs. Moreover, there also exists a clear
lack of consensus in the field as to what one should consider when assessing
Like others among the many "solutions" devised to improve the
poor decision making which plagues the child welfare system, the risk assessment
device only appears to have made things far worse for families and children.
"Many agencies have acted prematurely, implementing risk assessment
instruments that have not been adequately designed or researched,"
note Wald and Woolverton (1990). "It is not possible to make highly
accurate predictions of risk with existing instruments." Nevertheless,
they have gained an almost uncritical acceptance in the field of child welfare,
as Wald and Woolverton explain:
Unfortunately, some child protective services (CPS) agencies appear to be
using risk-assessment instruments in an unjustifiable manner, given the limited
knowledge base regarding the validity of these instruments. Moreover, we
are concerned that many agencies are adopting risk-assessment instruments
in lieu of addressing fundamental problems in existing child protection
systems, such as the excessive number of inexperienced or incompetent workers
and the lack of adequate resources. In fact, use of inadequately designed
or researched risk-assessment instruments may result in poorer decisions,
because workers will rely on mechanical rules and procedures instead of
trying to develop greater clinical experience.
While an in-depth review of risk assessment instruments is beyond the scope
of this work, a brief review of some of the risk factors typically assessed
by these instruments is in order given their role in the foster care decision-making
process. Among the specific factors which indicate a "high contribution
to risk" in the Cuyahoga County, Ohio, Risk Assessment, Safety and
Case Planning Manual (1997) are "diseases affecting motor coordination
(cerebral palsy, muscular dystrophy)," "confinement to wheelchair,"
"juvenile diabetes," and "severe to profound mental retardation."
Behavioral indicators in the high risk category include "autistic child"
and "child who lacks any emotional attachments." As for the characteristics
of the parent, a lack of regular social activities (e.g., church or other
organization) is considered a high risk factor, as is the refusal to avail
oneself "of community resources or appropriate resources available."
But if there is one thing many of these assessment tools have in common,
it is that they almost invariably define poverty conditions as indicating
neglect on the part of parents. Among the countless factors typically included
are "dirty or unkempt home," "children's clothing torn or
dirty," "lack of pride in neighborhood," "poor and unsafe
living conditions," "family can only afford inadequate housing,"
"leaky faucets," "no heat in home in winter," and "exposed
wiring." Indicators listed as among physical hazards in the high risk
category in the Cuyahoga County matrix include "uncovered holes in
exterior doors, walls, windows," "steep stairway," "steps
shaky," "boards missing from steps," and "exposed nails
The results are not surprising. Dana Mack (1997) notes that a 1986 federal
study evaluating child welfare caseworkers found that up to two-thirds of
substantiated cases of child maltreatment involved no actual wrongdoing
on the part of parents. Many removals of children into foster care are "capricious
actions of 'preventive intervention'-undertaken on a caseworker's presumption
that though a child's home situation poses no immediate dangers or deprivations,
it might sometime in the future" (p. 59). In examining studies conducted
by the American Humane Association during the mid-1980s, Mack found that
half of the families child welfare agencies compelled to undergo therapeutic
services for child maltreatment had never mistreated their children at all.
These therapeutic services are foisted on innocent families on a massive
scale. Besharov (1985) notes that even after the extensive screening of
reports that takes place, as of the mid-1980s roughly 400,000 families across
the country were being "supervised" by child protective agencies,
compelled to accept such "treatment services" under threat of
Confronted with criticism over the death of Joey Wallace, the Illinois Department of Children
and Family Services enlisted the aid of outside advisors to
draft the Child Endangerment Risk Assessment Protocol, a checklist which
reportedly guides caseworkers through home visits. The protocol asks caseworkers
to evaluate the child's surroundings, from the behavior of the parent or
guardian to the possibility of drug or alcohol abuse by the caretaker. Workers
are instructed that they can't leave unless they are convinced-through their
observations and checklist-that the child will be safe. "We tell workers
if any of those factors are there, you don't leave without the kid,"
explains Ed Cotton, deputy director of the Division of Child Protection
(Associated Press, 1997d).
Reduced to its essence, the child welfare system operates on the promise
of an eventuality being realized such that unsophisticated caseworkers,
once provided with the proper combination of tools and training, will develop
the ability to predict the course of future events, thus "preventing"
child maltreatment. No similar proposals have been advanced on behalf of
law enforcement officers such that they would be authorized to intervene
in order to prevent a crime before it occurs. No civilized society would
tolerate the arrest and detention of its citizens on the basis of a checklist
coupled with an "intuition" indicating that a crime is likely
to be committed at some point in the future. Yet society tolerates an entirely
similar situation in child protection today.
Over the course of the last decades, some few millions of children have
been irreparably harmed by arbitrary and capricious removals into state
care based on the theory of future-based preventive intervention. To the
extent that efforts at developing more accurate models of risk assessment
should continue, they should aim toward identifying and eliminating ecological
risks in the child's environment through the provision of hard services
subject to the uncoerced acceptance of such services by prospective clients.
If the field is to fulfill the promise of effectively addressing the problem
of child maltreatment, it must begin by divesting itself of empirically-unvalidated
models and methodologies, as they demonstrably do more harm than good.