969642ANN The Annals of the American AcademyThe Scope, Nature, and Causes of Child Abuse and Neglect

research-article2020

Child maltreatment is a complex problem affecting mil-

lions of children in the United States every year. This
article examines existing knowledge on the scope,
nature, and causes of child abuse and neglect. First, we
review the discordant definitions and conceptualiza-
tions of child maltreatment and consider the implica-
tions of broad and narrow definitions for the size and
scope of the child welfare system and for child safety.
Second, we provide an assessment of the quality and
comprehensiveness of existing data for understanding
The Scope, the incidence rates and trends in child abuse and
neglect. Third, we review theory and evidence on the
Nature, and causes of child maltreatment, with particular attention
to whether and how social policy can reduce its preva-

Causes of Child lence. Last, we provide recommendations for improv-

ing the use of data and scientific evidence in child
welfare policy and systems.
Abuse and Keywords: child maltreatment; child welfare system;
Neglect risk factors; measurement; data

O ver the past several decades, states have

developed and expanded their child
­welfare systems with growing federal oversight
By to effectively prevent and respond to child mal-
Sarah A. Font treatment. Over this same period, research and
and data to understand child maltreatment has pro-
Kathryn Maguire-Jack liferated. Yet core questions remain ­unanswered:
How many children experience maltreatment
today? How much have rates changed over time
and why? How effective are existing systems
in identifying children in need of protection?

Sarah A. Font is an assistant professor of sociology at

the Pennsylvania State University. Her research focuses
on the policies and practices of the child welfare system
and the experiences and outcomes of system-involved
children.
Kathryn Maguire-Jack is an associate professor of social
work at the University of Michigan. She studies child
maltreatment prevention with a focus on understanding
contextual risk and protective factors. She has expertise
in program evaluation and public policy.
Correspondence: Saf252@psu.edu

DOI: 10.1177/0002716220969642

26 ANNALS, AAPSS, 692, November 2020

The Scope, Nature, and Causes of Child Abuse and Neglect 27

What policies are most effective for reducing child maltreatment? These ques-
tions evoke important social values about the scope and size of government; roles
of federal, state, and local governments; children’s rights; and parental autonomy.
In this article, we review these debates and, where possible, we draw on available
research and data to make recommendations about how to move forward. We
begin with a review of how child maltreatment is defined—in federal policy, in
states’ civil and criminal statutes, and in research—and explore the implications of
broadening or narrowing definitions. Beyond definitions, we then describe the
difficulties and limitations associated with estimating rates of child maltreatment.
To this end, we draw on numerous population-level or nationally representative
survey datasets to compare estimates of the incidence and prevalence of child
maltreatment. These comparisons highlight the extent to which official statistics
on victimization (based on confirmed reports to child protective services [CPS])
underestimate exposure to abuse and neglect among U.S. children. Underestimates
of child maltreatment rates may lead to an underinvestment in resolving this prob-
lem. We then turn to a discussion of the causes of child maltreatment, with special
emphasis on factors that may be malleable through social policy changes. Last, we
provide recommendations about how investments in data and research can be
leveraged to inform policy reforms, improve the child welfare system, and better
protect children from abuse and neglect.

Scope and Nature of Child Maltreatment

This section reviews definitions of child maltreatment, describes approaches to
measuring maltreatment, and reviews existing estimates of the incidence and
prevalence of child maltreatment.

What is child maltreatment?

Child maltreatment has a range of definitions, with variability across and within
countries; between civil and criminal statutes; and across legal, lay, and academic
perspectives. Further, variability in measurement reflects differences in defini-
tions, standards of evidence, and sources of information. In the United States, the
Child Abuse Prevention and Treatment Act (CAPTA), originally passed in 1974
(P.L. 93-247), provides the federal definition of child maltreatment: “Any recent
act or failure to act on the part of a parent or caretaker, which results in death,
serious physical or emotional harm, sexual abuse, or exploitation, or an act or fail-
ure to act which presents an imminent risk of serious harm.” This definition
encapsulates a fairly broad range of actions and inactions that can be defined as
child maltreatment but narrows the focus to perpetrators in caregiving roles.
Typically, the child welfare system focuses on maltreatment perpetrated by indi-
viduals who are responsible for the child, consistent with the primary mandate of
child safety. The child welfare system receives and investigates allegations of child
maltreatment through CPS and provides in-home and foster care services to
28 THE ANNALS OF THE AMERICAN ACADEMY

families at risk. When a child is abused by a person outside the family, the child’s
parents (or legal custodians) are expected to take protective action (e.g., by elimi-
nating contact with the abuser), and law enforcement and the criminal justice
system can act to protect society at large through criminal prosecution. This leaves
no clear need for child welfare system intervention; in contrast, abuse or neglect
by a parent may require child welfare system intervention, alone or in combina-
tion with law enforcement.
States provide more specific—and sometimes more expansive—definitions in
their civil statutes (which guide child welfare system and family court actions)
and criminal statutes (which guide decisions to prosecute forms of child maltreat-
ment as a criminal offense). Consistent with a focus on child safety rather than
parental culpability, statutory definitions of child maltreatment tend to empha-
size harm or threat of harm to children that results from specific actions or inac-
tions, with comparatively little emphasis on perpetrator intent. In research,
definitions tend to be less restrictive with regard to perpetrators and may con-
sider a range of exposures that pose a threat to children’s safety or welfare, irre-
spective of whether they meet legal definitions. For example, survey-based
measures of neglect tend to include so-called involuntary neglect, or situations in
which a child experiences material deprivation but it is unknown whether the
deprivation is solely due to poverty. The widely used Parent-Child Conflict Tactic
Scales—a caregiver self-report survey instrument about the frequency of various
parenting acts or omissions over the past 12 months in the domains of psychologi-
cal aggression, physical assault, nonviolent discipline, and neglect (Straus et al.
1998)—include children lacking necessary medical care in the neglect subscale,
which may occur for reasons of negligence, poverty, or (less commonly) malice.
Under many states’ statutes, neglect that occurs solely due to poverty is not
defined as child neglect (Rebbe 2018), regardless of harm incurred to the child.
Nevertheless, definitions remain vague and subject to differential interpretation.
In addition, low-income families compose a large majority of CPS reports and
families receiving child welfare system interventions (Dolan et al. 2011), and
child welfare agencies face persistent criticism about whether they accurately
distinguish parental acts of neglect from poverty-driven material hardships
(Eamon and Kopels 2004; Milner and Kelly 2020). Yet differentiating neglect
from poverty is a rather difficult and subjective judgment. Consider a single par-
ent who leaves her toddler home alone because she lacked childcare, and while
alone, the child falls down the stairs. Injuries sustained in the fall were uninten-
tional but were a foreseeable risk given the child’s developmental stage. Whether
the incident was mostly about poverty or mostly about parental negligence is not
clear-cut. One approach is to consider whether a reasonable person might have
made the same choice given the parent’s constraints. For example, a parent who
left her child home alone to avoid being fired from her job (which would risk
homelessness and other forms of serious deprivation) arguably behaved more
reasonably than a parent who left to attend a social event. However, to make such
a determination about whether neglect was “involuntary,” the circumstances
must be investigated; as such, it is, to some extent, inevitable that impoverished
families will be overrepresented in neglect reports to CPS. In addition, whereas
The Scope, Nature, and Causes of Child Abuse and Neglect 29

the distinction between poverty and neglect can and should drive decisions about
how to intervene—in-home services versus foster care, court-mandated services
versus voluntary community supports—it is not clear such distinctions should
determine whether to intervene. It would seem reckless for the child welfare
system to ignore serious risk or harm to a child simply because the parent was not
perceived to be “at fault,” particularly in a society in which the child welfare sys-
tem may be the best or only means of accessing services or resources.

Measurement of child maltreatment

There are numerous approaches to measuring maltreatment. The most promi-
nent strategies and data sources are described in Table 1. Each data source or
approach has significant limitations. Both the National Child Abuse and Neglect
Data System (NCANDS; a federal database comprising extracts of state child
welfare records) and the National Surveys of Child and Adolescent Well-Being
(NSCAW; a federally sponsored longitudinal survey of CPS investigations) only
contain information about families investigated by CPS, thus providing no infor-
mation about the nature or extent of child maltreatment that does not reach the
attention of CPS. Although NSCAW includes a variety of measures of child mal-
treatment, including parent-reported and child-reported, each measure is lim-
ited to a different subsample, thus inhibiting comparisons of measures. For
example, the parent-reported measures are only asked for the primary parent and
only if they retain custody of their children at the time of the interview, whereas
the child-reported items that directly align with the parent-reported items are
only asked of children ages 11 and older.
Outside of data collected on children and families already involved in the child
welfare system (e.g., NCANDS, NSCAW), there is little prospective data collec-
tion on child maltreatment. Beginning in the 1980s, there have been four
National Incidence Studies (NIS), which collected information about suspected
child maltreatment from professionals who have regular contact with children.
The value of these studies is that they captured child maltreatment that may not
have been reported to CPS or that was reported but screened out (not investi-
gated) by CPS. In addition, by relying on informants other than parents, these
studies are, arguably, less biased by the limitations and biases of self-reported
maltreatment. However, the last NIS was in 2005–2006, and we do not know
when funding will be appropriated to conduct another round.
Many large-scale longitudinal surveys, such as the Fragile Families and Child
Wellbeing Study (FFCWS; a longitudinal study that tracks a nationally repre-
sentative birth cohort of children born in large U.S. cities between 1998 and
2000), do not collect explicit information on child maltreatment. Although there
are a variety of reasons that surveys may include or exclude particular measures,
of particular note, affirmative disclosures of child maltreatment victimization or
perpetration introduce legal and ethical dilemmas about mandatory reporting to
CPS and informed consent (Putnam, Liss, and Landsverk 2014). If disclosures
would invoke mandatory reporting responsibilities, respondents may be unlikely
to respond truthfully, which may limit the quality and utility of any data collected.
30
Table 1
Commonly Tracked Measures (and Proxies) of Child Maltreatment

Measure Type Found Ina Examples of Measures Utility as a Measure of Maltreatment

Child welfare NCANDS; Investigation, alternative response, Benefits: National coverage, annual collection, consistency of the
involvement NSCAW and substantiation of maltreat- variables included each year
ment allegations, by type
Concerns: Detection bias (Brown et al. 1998), ongoing debate
regarding meaningfulness of substantiation (Font, Maguire-Jack,
and Dillard 2020), state and temporal variation in statutory defi-
nitions and policies (Child Welfare Information Gateway 2016),
variation in completeness and quality of state reports
Professional proxy- NIS Professional descriptions of scenar- Benefits: Addresses some problems related to detection bias;
reports ios coded by type and severity of accounts for degrees of severity
maltreatment (Sedlak et al. 2010)
Concerns: Small samples; concerns about methodology (proxy-
reported estimation)
Parent-reported NSCAW, NSCEV Parent-Child Conflict Tactic Scales Benefits: Addresses some problems related to detection bias;
maltreatment (Straus et al. 1996); Juvenile includes more nuanced questions (e.g., frequency, severity)
Victimization Questionnaire
(Finkelhor et al. 2005)
Concerns: Response bias due to social desirability and mandatory
reporting requirements; low agreement between parent and child
reports (Font and Cage 2018); inadequate measures of neglect
Child-reported NSCAW Child version of Conflict Tactic Benefits: May reduce detection and social desirability biases;
maltreatment Scales; Exposure to Violence allows for cross-comparisons of multiple sources of information
Questionnaire (Fox and Leavitt (e.g., parent and child reports)
1995) Concerns: Ethical issues around reporting/risk to children who dis-
close; not available for young children; typically excludes neglect

(continued)
 Table 1 (Continued)

Measure Type Found Ina Examples of Measures Utility as a Measure of Maltreatment

Substandard par- Numerous popula- Commonly: physical punishment, Benefits: Parents may be more truthful about substandard parent-
enting tion-based severe material hardships, yelling/ ing than maltreatment; responses do not trigger mandatory
household sur- criticizing child, substance use reporting duties; substandard parenting and maltreatment are
veys (e.g., and domestic violence in the similarly associated with child wellbeing; often combine parent
FFCWS) household reports and interviewer observations; relative measures of how
far below “average” parenting quality a family is in a particular
domain
Concerns: Not exact measures of maltreatment (as legally
defined); may still have underreporting due to social desirability
bias; limited ability to track population-level change over time
Adult retrospec- BRFSS; Add- Commonly: Adverse Childhood Benefits: By asking adults, there is no concern about mandatory
tive reports of Health Experiences (ACEs) question- reporting, which may allow for more truthful responses
maltreatment naire (Felitti et al. 1998), which
includes some maltreatment
items
Concerns: Recall bias is likely significant, resulting in high rates of
false negatives (Hardt and Rutter 2004); inadequate measures of
neglect

NOTE: NCANDS: National Child Abuse and Neglect Data System (https://www.ndacan.acf.hhs.gov/datasets/datasets-list-ncands-child-file.cfm).
NSCAW: National Survey of Child and Adolescent Wellbeing (https://www.ndacan.acf.hhs.gov/datasets/datasets-list-nscaw.cfm). NIS: National
Incidence Studies (https://www.ndacan.acf.hhs.gov/datasets/datasets-list-nis.cfm). NSCEV: National Survey of Children’s Exposure to Violence
(https://www.icpsr.umich.edu/icpsrweb/NACJD/studies/36523). FFCWS: Fragile Families and Child Wellbeing Study (https://fragilefamilies.prince-
ton.edu/documentation). BRFSS: Behavioral Risk Factor Surveillance System (https://www.cdc.gov/brfss/data_documentation/index.htm). Add-
Health: National Longitudinal Study of Adolescent to Adult Health (https://www.cpc.unc.edu/projects/addhealth/documentation).
a. Not an exhaustive list.

31
32 THE ANNALS OF THE AMERICAN ACADEMY

Thus, some surveys may include measures of “substandard parenting” or

“behaviorally-approximated” maltreatment measures—parenting behaviors or
­
environments that post risks to child safety and well-being but are not considered
to trigger mandatory reporting responsibilities. Such measures may include chil-
dren’s exposure to high-frequency or harsh physical discipline, domestic violence
or parental substance use, or deprivation of basic needs (Font and Berger 2015;
Schneider, Waldfogel, and Brooks-Gunn 2016; Berger et al. 2017).
Child victims may also underreport experiences of maltreatment for fear of
the consequences to their parents or the unknown environment that may await
them in foster care, or due to poor recall of early experiences or difficulty con-
ceptualizing neglect. Moreover, interviews with young children require signifi-
cant training to ensure quality, which may be cost-prohibitive in much survey
research. Child-reported maltreatment is thus relatively uncommon in surveys,
and when it is used, measures vary in depth and scope (Amaya-Jackson et al.
2000). Retrospective studies of child maltreatment are far more common, in part
because they are used to examine long-term outcomes associated with early
childhood experiences. However, such measures are prone to underreporting
due to recall bias (Hardt and Rutter 2004), which may pose particular concerns
for identifying less-severe maltreatment, maltreatment in early childhood (versus
middle childhood or adolescence), and child neglect (versus physical, emotional,
or sexual abuse).
More generally, existing data sources include inadequate and inconsistent
measures of neglect. Child neglect can include a large number of domains, but
commonly used scales tend to focus disproportionately on physical neglect
(unmet basic needs). Such measures tend to conflate poverty-driven material
deprivation with negligent parenting by sidestepping complex issues of parental
capacity, intent, and culpability, and by excluding more difficult-to-measure
experiences such as inappropriate supervision or failure to protect children from
harm. In contrast, CPS records are, on the whole, likely to substantially under-
capture emotional maltreatment and, by design, do not capture most physical or
sexual abuse committed by noncaregivers. Moreover, low rates of substantiation
for all forms of maltreatment are suggestive of substantial undercounting, and it
is not possible to differentiate between genuinely false CPS reports and reports
where CPS deemed the evidence insufficient or CPS did not consider the inci-
dent severe enough to warrant substantiation. Thus, researchers must use cau-
tion when using CPS records—and CPS victimization rates based on substantiated
reports—to make claims about the incidence or prevalence of child maltreat-
ment. In addition, substantial variability in the rate at which reports are screened-
in for investigation or substantiated make cross-state or cross-year comparisons
highly suspect. Notwithstanding, CPS investigation records are appropriate for
studies of the causes or consequences of child maltreatment: CPS investigations
are associated with short- and long-term adverse outcomes for children even
when no intervention occurs, indicating that the adverse outcomes cannot be
attributed solely to system intervention, but rather to the alleged maltreatment
and related family circumstances (Hussey, Marshall, Knight, et al. 2005; Font and
Maguire-Jack 2020). Moreover, children known to the child welfare system are
The Scope, Nature, and Causes of Child Abuse and Neglect 33

Table 2
Child Maltreatment Incidence Estimates (Annual Rate per 1,000 Children)

National Child Abuse

National Incidence Study and Neglect Data National Survey
IV, 2005–2006 System, 2016 of Children’s
Exposure to
Harm Endangerment Investigated Confirmed Violence,
Standard Standard Children Victims 2013–2014

Any maltreatment 17.06 39.46 46.74 9.10 152

Neglect 10.48 30.58 29.78 7.01 51
­(composite)
Physical neglect 4.01 16.19 —
Medical neglect 1.29 0.20
Emotional neglect 2.63 15.94 —
Educational 4.9 4.9 —
neglect
Physical abuse 4.39 6.47 11.83 1.66 50
Sexual abuse 1.84 2.45 3.68 0.78 1
Psychological 2.02 4.11 3.62 0.52 93
­maltreatment /
emotional abuse

NOTE: NIS-IV estimates drawn from NIS-IV final report Appendices B and C (Sedlak et al.
2010); NCANDS estimates generated from the 2016 child file version 3; NSCEV estimates
from Finkelhor et al. (2015).

those positioned to receive intervention, and thus studies of their circumstances

and outcomes are necessary to inform policies and practices. Nevertheless,
researchers should be diligent about explaining the limitations of such measures,
including likely contamination of the comparison group (i.e., undetected mal-
treatment among those without CPS investigations). Ideally, however, questions
of critical significance for social policy would be addressed with multiple meas-
ures of child maltreatment.

Estimating the incidence and prevalence of child maltreatment

The type of measure used has significant implications for estimating the inci-
dence and prevalence of child maltreatment. Table 2 shows different estimates
of child maltreatment incidence (annual rate per 1,000 children) from the NIS-4,
NCANDS, and the National Survey of Children’s Exposure to Violence (NSCEV).
The estimates for NCANDS are based on the authors’ analysis of the NCANDS
child files, whereas the other estimates are pulled from published research or
reports (Finkelhor et al. 2015; Sedlak et al. 2010). It is evident from Table 2 that
CPS substantiations are likely to grossly understate all forms of child maltreat-
ment, but especially physical abuse. The NIS-4 includes two measures: a harm
34 THE ANNALS OF THE AMERICAN ACADEMY

Table 3
Lifetime Prevalence Estimates by Source

Data Source NCANDS BRFSS NSCEV

Measure Substantiated Investigated as Adult retrospec- Caregiver report

CPS victims CPS victims tive self-report for children
by age 18 by age 18 ages 14–17
Data type 2014 population data 2011–2014; 2013–2014
23-state nationally
­sample ­representative
sample
(Source) (Kim et al. 2017) (Centers for (Finkelhor et al.
Disease 2015)
Control and
Prevention
2019)
Any maltreatmenta 11.8% 37.4% — 38.1%
Neglect 8.0% 25.2% — 18.4%
 Physical abuse 2.0% 11.5% 17.9% 18.1%
Sexual abuse 0.9% 4.1% — 0.2%
­(narrow perpetrator
definition)
 Psychological 0.6% 3.5% 34.4% 23.9%
­maltreatment/
emotional abuse
Witness to family — — 17.5% 19.5%
­violence
Sexual abuse (broad — — 11.6% 3.2%
perpetrator
­definition)

a. The types of maltreatment included in the composite “any maltreatment” measure vary by
source.

standard and an endangerment standard. The harm standard is more stringent

and requires evidence of specific harms to the child victim, whereas the endan-
germent standard includes any maltreatment that poses a significant risk to
children.
Table 3 depicts estimates of lifetime prevalence (from birth to age 18) of child
maltreatment for children in the United States, drawing from national population
or nationally representative survey datasets. We show the overall estimate as well
as subtypes due to lack of consistency across data source in types of maltreatment
included. Unsurprisingly, estimated prevalence based on CPS-substantiated mal-
treatment is far lower than estimates based on CPS investigations or caregiver-
reported maltreatment. For example, 2 percent of U.S. children have a
CPS-substantiated allegation of physical abuse by age 18, whereas 18 percent of
The Scope, Nature, and Causes of Child Abuse and Neglect 35

children are estimated to have experienced physical abuse based on retrospective

reports of adults (Behavioral Risk Factor Surveillance System; BRFSS) or car-
egiver report (NSCEV).
Both CPS-substantiated and CPS-investigated maltreatment rates suggest
emotional maltreatment is relatively rare, whereas emotional maltreatment is the
most common form of maltreatment identified in the BRFSS and NSCEV. The
differential findings related to emotional maltreatment may be driven by the fact
that emotional maltreatment is difficult to demonstrate and “prove.” However,
forms of maltreatment most likely to have physical indicators, like physical abuse,
also have very low rates of substantiation when investigated by CPS. Rather, low
rates of CPS investigation for emotional maltreatment may indicate underreport-
ing, which may reflect inadequate understanding of the signs or symptoms of
emotional abuse among mandatory reporters, or reporters’ ambivalence about
the threshold at which harsh or withdrawn parenting becomes emotional mal-
treatment. Additionally, maltreatment subtypes commonly co-occur, but a single
form of maltreatment is sufficient to justify intervention. Thus, CPS investigators
may choose to pursue or focus on the allegations of abuse or neglect that are easi-
est to document. Scholars have long noted the difficulty in identifying and
responding to emotional maltreatment, given the child welfare system’s emphasis
on physical safety (English et al. 2015). Moreover, in surveys, measures of emo-
tional abuse tend to focus on the child’s experiences of verbal abuse, whereas
legal definitions (commonly, “mental injury”) require evidence that the child suf-
fered serious and observable injury to their psychological capacity or emotional
stability (Child Welfare Information Gateway 2016) as a result of the experiences.
(Of course, since researchers are often interested in the effects of maltreatment,
defining maltreatment based on its effects would be tautological.)
Notably, CPS-substantiated sexual abuse is more prevalent than caregiver-
reported sexual abuse by a parent or caregiver. This may reflect underreporting
by caregivers, in addition to differences in what each measure captures.
Depending on the state, CPS measures of sexual abuse may include a broader
definition of perpetrator (e.g., not limited to parents or caregivers; may include
figures such as babysitters, temporary household members, or siblings).
Differences may also reflect some states’ practice of designating “failure to pro-
tect from sexual abuse” in the same category as perpetration of sexual abuse.

Prior research on concordance across measures of child maltreatment

Prior research has also sought to compare estimates of child maltreatment
using multiple sources or approaches to measurement. Overall, such research
suggests low agreement between child and parent reports of children’s experi-
ences of abuse, neglect, and other adversities (Chan 2015; Font and Cage 2018;
Schneider et al. 2014) and between child self-reports and CPS or court records
(Pinto and Maia 2013; Swahn et al. 2006). Given lack of consensus about the
definition and operationalization of neglect, it is perhaps unsurprising that agree-
ment across reporting sources appears lowest for neglect (McGee et al. 1995).
36 THE ANNALS OF THE AMERICAN ACADEMY

The implications of low agreement across measures for scientific understand-

ing of the antecedents and outcomes of child maltreatment are not fully appar-
ent. Some research studies indicate that the correlates of victimization (both risk
factors and outcomes) are similar irrespective of the source of information (Font
and Berger 2015; Font and Cage 2018; Schneider et al. 2014). However, other
research suggests that children’s reports of their prior victimizations are more
predictive of social-emotional functioning reports from social workers or infor-
mation from child welfare case records, especially for child neglect and emo-
tional maltreatment (McGee et al. 1995).
The discordance across measures of maltreatment raises critical questions
about the extent to which child welfare agencies are providing services to the
optimal number of families and whether families receiving services are those
most in need. A large number of children are reported to CPS each year, and
hundreds of thousands of children enter foster care (U.S. Department of Health
and Human Services 2020), but are these numbers disproportionate to the size
of the underlying problem of child maltreatment? It seems unavoidable that
some number of nonmaltreated children will be reported to CPS if mandatory
reporters are acting appropriately—suspicious events, such as unexplained inju-
ries or a young child exhibiting inappropriate sexual behavior, should be reported
but may ultimately have nonmaltreatment explanations. Some may point to the
large proportion of “unsubstantiated” CPS investigations as an indication that too
many investigations are occurring. However, there is little reason to believe that
most unsubstantiated investigations are false allegations. Moreover, research has
documented that there is little consistency across caseworkers or agencies in
decision-making (Doyle 2007; English et al. 2002; Font, Maguire-Jack, and
Dillard 2020), and there is substantial evidence that children with an unsubstan-
tiated maltreatment report face heightened long-term risk of a host of negative
outcomes (Font and Maguire-Jack 2020; Hussey, Marshall, English, et al. 2005),
including future maltreatment (Kohl, Jonson-Reid, and Drake 2009) and death
(Putnam-Hornstein 2011). In addition, annual and cumulative incidence rates of
investigated maltreatment are not substantially greater than rates of child mal-
treatment reported in surveys (Tables 2–3). Another data point to consider is the
approximately 2.7 million U.S. children informally raised by relatives (Annie E.
Casey Foundation 2012) in what is referred to as the “hidden foster care system”
(Gupta-Kagan 2020). This suggests that a rather large number of children
require temporary or permanent substitute care, far exceeding the number of
children placed in formal foster care. Collectively, these data points suggest that
the number of children involved with the child welfare system may not be exces-
sive relative to the number of children exposed to substantial risk in their familial
environments.
However, these data also imply that the child welfare system may incorrectly
identify which children are in need of intervention beyond the CPS investigation
(i.e., in-home services or foster care). Certainly, not all children who experience
maltreatment are in need of intervention, and part of the investigative process is
to determine what, if anything, is needed to reduce future risk of harm. When a
nonparent perpetrates maltreatment, and children have a parent or legal
The Scope, Nature, and Causes of Child Abuse and Neglect 37

caregiver willing and able to protect them from future harm, intervention may
unnecessarily usurp or undermine protective action taken by parents or caregiv-
ers. Similarly, in cases involving a single incident of nonsevere maltreatment,
caregivers may be able to successfully address the precipitating factors (mental
health, substance use, parenting skills) through voluntary community services.
However, it would be a misstep for child welfare agencies to merely delay action
until serious harm occurs—it is critical to identify (1) when a low-risk event is a
signal of a deteriorating family environment or incipient crisis and (2) whether
parents in need of support are able and willing to follow up on voluntary com-
munity services without oversight or court mandate. Unfortunately, there is rela-
tively limited evidence to bring to bear on these issues, and subjective assessment
of issues like parental cooperation raise concerns about implicit bias. Predictive
risk modeling (discussed by Drake et al., this volume) may provide a path forward
for more effective targeting of limited resources.

Are rates of child maltreatment increasing or decreasing?

Rates of child sexual abuse and child physical abuse appear to have declined
substantially throughout the 1990s and early 2000s according to numerous data
sources (victimization surveys, crime reports, CPS substantiations) and consistent
with declines in violent crime broadly (Finkelhor and Jones 2006). However, in
recent years, CPS-substantiated rates of sexual abuse have increased once again,
and rates of physical abuse have stagnated (Finkelhor, Saito, and Jones 2020).
Factors that may have contributed to declines in physical and sexual abuse spe-
cifically include increased public awareness and prevention efforts focused on
physical abuse (e.g., shaken baby prevention campaigns) and sexual abuse (e.g.,
proliferation of “safe touch” training), increased treatment and medication
options for child behavior problems, and more aggressive prosecution and sen-
tencing of violent crimes (Jones, Finkelhor, and Halter 2006).
It is more difficult to ascertain whether neglect rates have changed during this
period: CPS data suggest little net decline, with enormous variability across states
(Finkelhor, Saito, and Jones 2020), and there are few other data sources to negate
or confirm. Moreover, because neglect comprises a broad range of parental acts
and omissions, it is possible that some forms of neglect have increased and others
have declined. More detailed categorization of neglect in national data systems is
essential for testing the impacts of various social, demographic, and governmen-
tal changes on rates of child neglect.
Moreover, critical questions about how and for whom maltreatment rates have
changed remain unanswered. For example, has the cumulative incidence (preva-
lence) of perpetrators and victims declined? A decline in annual incidence of
maltreatment coupled with no change in prevalence of victims or perpetrators
may suggest effective interventions (success in reducing revictimization and reof-
fending) rather than effective prevention. Available data suggest slight declines
in prevalence of substantiated maltreatment between 2004 and 2009, and little
change between 2009 and 2016 (Wildeman et al. 2014; Yi, Edwards, and
Wildeman 2020). Again, however, declines in the prevalence of substantiated
38 THE ANNALS OF THE AMERICAN ACADEMY

maltreatment may reflect, in full or in part, changes to child welfare system policy
and practice.

Risk Factors for Child Maltreatment

This section reviews theories surrounding the causes of child maltreatment, risk
factors for maltreatment, and limitations of the current work.

Theories related to child maltreatment etiology

Several theories from a range of scientific disciplines help to explain differ-
ences in children’s risk of experiencing maltreatment. These theories focus on
child and parent characteristics, relationships within the family, and the environ-
ment surrounding the parent; some theories additionally emphasize the interplay
among multiple factors. Although early frameworks for understanding child
maltreatment focused largely on the pathologies of perpetrators, modern theo-
ries tend to emphasize a broader range of individual, family, environmental, and
societal factors (Garbarino 1977; Gelles 1973).
A large body of research has focused on the intergenerational transmission of
child maltreatment, or the processes through which individuals’ experiences of
maltreatment victimization increase the risk that their children experience mal-
treatment victimization (Schelbe and Geiger 2017). Attachment theory (Bowlby
1969) suggests that insecure attachment—or a phenomenon in which the parent-
child relationship is contaminated with fear and distrust—explains intergenera-
tional transmission of maltreatment. Parents who have not experienced secure
attachments with their own caregivers have difficulty forming them with their
own children (Morton and Browne 1998). Other research refers to social learning
theory (Bandura 1978) to suggest that intergenerational transmission occurs
because children learn how to be parents from their own parents (Muller,
Hunter, and Stollak 1995). Both theories, as applied to maltreatment, suggest
that the propensity to perpetrate child maltreatment is rooted in parents’ own
childhood relationships and experiences.
Looking beyond explanations related to prior parenting experiences, the family
stress model of economic hardship (Conger and Elder 1994) was proposed to
understand the pathway through which economic hardships, such as debt burden,
income loss, or economic insecurity, negatively affect child and adolescent devel-
opment. The model proposes that economic hardship may not only deprive chil-
dren of critical material needs, but also adversely impact family dynamics. The
model posits that difficulties in or inability to meet family economic needs results
in economic pressure, which in turn produces psychological distress, relationship
conflict, and changes in parental affect and behavior (Conger and Elder 1994). In
these family environments, children may be subjected to harsh and inconsistent
discipline practices, or parents may be withdrawn (Conger et al. 1994). In extreme
cases, harsh or withdrawn parenting can escalate to child abuse or neglect.
The Scope, Nature, and Causes of Child Abuse and Neglect 39

Social disorganization theory was originally put forth to explain geographic

variation in crime and delinquency (Shaw and McKay 1942), but it has since been
widely applied to geographic variation in child maltreatment rates. The theory
postulates that there is something unique about the communities in which indi-
viduals live that can increase the rates of crime and delinquency, and that by and
large these differential rates are not driven by differences across individual peo-
ple. Shaw and McKay (1942) referred to this phenomenon as social disorganiza-
tion and proposed three community-level factors that they posited led to
increased crime and delinquency: concentrated disadvantage, ethnic heterogene-
ity, and residential mobility. This theory has been adapted by child maltreatment
researchers to understand geographic variation in maltreatment. These research-
ers have suggested that parents face multiple stressors, are unable to access
resources, and do not have the necessary social norms to prevent maltreatment
because of the neighborhood in which they live (Coulton et al. 2007; Coulton,
Korbin, and Su 1999).
Theorists have also considered how child maltreatment is defined and the
extent to which that reflects the norms of those in position to set and enforce
parenting standards (Hutchinson 1990). Social deviance/labeling theory posits
that there is no objective behavior that can be called “child abuse”; it is only
through perceiving deviance from a socially accepted norm that it is labeled as
such (Gelles 1975). Specifically, Gelles (1975) refers to child abuse as a social
construction by which “a) a definition of abuse is constructed, b) certain judges
or ‘gatekeepers’ are selected for applying the definition, and c) the definition is
applied by designating the labels ‘abuse’ and ‘abuser’ to particular individuals and
families” (p. 365). Thus, some argue that, although violence occurs in a range of
family environments, it is predominantly labeled as “child abuse” when it occurs
within lower-income or otherwise socially marginalized families (Newberger,
Newberger, and Hampton 1983).
Last, ecological systems theory (Bronfenbrenner 1976) combines many of
these theories, asserting that to truly understand individuals, one must consider
all of the factors that occur at multiple levels of the social ecology, conceptualized
as systems. Ecological systems theory is depicted as a series of concentric circles,
with the child at the center of these circles surrounded by the microsystem (the
child’s most consistent and frequent connections: parents, siblings, peers, and
school), mesosystem (relationships between the child’s microsystems), exosystem
(institutions and public policies), and macrosystem (social norms and cultural
values). This theory posits that there is no single cause of maltreatment, but there
are a range of independent and interactive factors that when combined may lead
to abuse or neglect.
Although not originally created to explain child maltreatment, ecological sys-
tems theory is prolific in child maltreatment research and theory. An early adop-
tion of this framework was Belsky’s efforts to integrate child, family, and
environmental influences into a holistic model of parenting and child maltreat-
ment (Belsky 1980, 1984). Belsky theorized that a parent’s own history of mal-
treatment as a child increases the likelihood that they will engage in abusive or
neglectful parenting themselves. Within the microsystem, Belsky suggested that
40 THE ANNALS OF THE AMERICAN ACADEMY

the family and child him/herself contributes to maltreatment—that children

influence their parents’ behavior, while simultaneously being influenced by it. At
the exosystem level, Belsky pointed to places of employment and the neighbor-
hood. Parents who are unemployed or have low job satisfaction may have a
greater propensity to maltreat. Further, being socially isolated within one’s neigh-
borhood can both contribute to a greater level of stress/depression and increase
the opportunity to maltreat because of a lack of oversight from others. Finally, at
the macrosystem level, Belsky suggested that society’s attitudes toward violence,
corporal punishment, and the rights of children all contribute to maltreatment.
As the body of theoretical work proliferated, so too did empirical studies
examining the extent to which data supported the theories. Applying attachment
and social learning theories, studies have focused on factors about individual
parents’ own experiences of maltreatment in childhood (Muller, Hunter, and
Stollak 1995). Applying the family stress model of economic hardship, research-
ers have focused on economic shocks in families and their relation to harsh par-
enting (Conger et al. 2002; Parke et al. 2004). Social disorganization theory has
been applied by child maltreatment researchers to hone in on the impact of
neighborhood characteristics on parenting (Coulton et al. 2007). Still other stud-
ies have relied on the more holistic ecological systems theory to focus on the
array of characteristics at different levels of the social ecology that might contrib-
ute to maladaptive parenting behaviors (Mulder et al. 2018).
Taken together, these theories highlight that child maltreatment is a complex
phenomenon with multiple conditions that must coalesce for abuse and neglect
to occur. The theories range from unidimensional, focusing on specific factors of
individual relationships (e.g., attachment theory) to the multidimensional, focus-
ing on a broad array of factors (e.g., ecological systems theory). Theories that fall
at different points on this continuum have complementary strengths and weak-
nesses, with those that are more narrowly focused being easier to examine and
test, but possibly missing many other causes; and those that are broader being
much more difficult to test, but more holistic in their consideration of the causes.

Prior reviews and meta-analyses on risk factors for child maltreatment

Four components are required to establish a causal effect of a presumed risk
factor on an outcome: (1) there must be a logical relationship, (2) there must be
an empirical association, (3) the temporal ordering must be accurate, and (4) the
relationship must not be spurious or due to an omitted variable (National
Research Council 2014). Most prior work has identified correlates of child mal-
treatment, as opposed to causal factors that meet all of the aforementioned
requirements for causality. There has been a large focus on specific characteris-
tics of the child, parents, and family that might contribute to abuse and neglect.
Parent characteristics consistently associated with child maltreatment include
parental experiences of childhood maltreatment (Assink et al. 2019; Mulder et al.
2018; Stith et al. 2009; van IJzendoorn et al. 2020); parental experiences of inti-
mate partner violence (Assink et al. 2019; Korbin and Krugman 2014; van
IJzendoorn et al. 2020); and parental behavioral health characteristics such as
The Scope, Nature, and Causes of Child Abuse and Neglect 41

anger, depression, alcohol and drug abuse, and psychopathology (Mulder et al.
2018; Stith et al. 2009; van IJzendoorn et al. 2020). In addition, numerous studies
indicate that parental and family socioeconomic status is associated with child
abuse and neglect (Erickson, Labella, and Egeland 2018; Kolko and Berkout
2018; Korbin and Krugman 2014; National Research Council 2014; Stith et al.
2009) as well as family structure, with single-parenthood being associated with a
greater risk for abuse and neglect (Mulder et al. 2018; Stith et al. 2009).
Research to identify the causal antecedents of child maltreatment has largely
focused on factors that can be addressed through public policy. In particular,
reforms and expansions of public benefit programs or economic policies provide
bountiful opportunity for natural experiments. This body of research generally
suggests that increased income (through the Earned Income Tax Credit,
increased minimum wage, child care subsidy, welfare, and child support policies)
reduces CPS reports, child neglect, and foster care caseloads (Berger et al. 2017;
Biehl and Hill 2018; Cancian, Yang, and Slack 2013; Maguire-Jack et al. 2019;
Raissian and Bullinger 2017; Yang et al. 2019), whereas reduced access to
employment and public benefits may increase rates of CPS reports (Fein and
Lee 2003; Paxson and Waldfogel 1999; Raissian 2015). Estimated effect sizes vary
across studies and are difficult to compare due to differences in samples, meth-
ods, and measures. Broadly, however, studies of income changes suggest that
relatively modest increases in income may reduce an individual’s risk of a CPS
report by about 10 percent (Berger et al. 2017; Cancian, Yang, and Slack 2013);
state-level studies of policies that increase family income similarly suggest reduc-
tions in rates of CPS reports or foster care entries by 7 to 10 percent (Biehl and
Hill 2018; Raissian and Bullinger 2017).
There is a related small but growing body of research that suggests that expan-
sive social policies—particularly, those that increase access to income, child care,
parental leave, and health care—may also have unintentional benefits vis-à-vis
child maltreatment prevention, such as decreased child welfare reports,
decreased self-reported maltreatment, and decreased foster care caseloads
(Campbell 2019). This research is not conclusive, however, and the issues plagu-
ing this body of research are complex and difficult to address. For example,
policy impact studies often leverage differences in policies across states or
changes in policy over time. As with other studies of the causes of maltreatment,
many policy impact studies use CPS records to measure child maltreatment, and
both temporal and geographic variability in child welfare systems practices and
policies may confound policy variation. Thus, estimating the causal effect of a
particular policy is challenging. Also, many studies rely on state-level data, mak-
ing it difficult to identify the mechanisms through which policy changes impact
rates of CPS-reported maltreatment. Because substantiation status is not a reli-
able means of distinguishing true versus false reports of child maltreatment,
many studies focus on changes in the rate of CPS investigations (also referred to
as “screened in reports”). Although there are plausible mechanisms through
which such policies or programs could reduce child maltreatment (e.g., reducing
parental stress, ensuring children’s basic needs are met), it is also possible for
CPS reports to decline with no true change in child maltreatment, given that
42 THE ANNALS OF THE AMERICAN ACADEMY

some subset of children reported to CPS have not experienced maltreatment.

Expanded economic benefits, for instance, may reduce low-risk CPS reports and
those that are unlikely to be substantiated, in particular, by reducing suspi-
cion of neglect among families for whom poverty is the sole or primary risk
factor. Thus, claims about reducing the occurrence of child maltreatment
on the basis of changes in CPS reports should be made and interpreted
­cautiously. Notwithstanding, if such policies or programs reduce the rate at
which impoverished families are unnecessarily reported to CPS, that frees up
resources to aid families where child maltreatment is occurring, which may
­ultimately improve the quality and functioning of the child welfare system.
Beyond income or antipoverty programs, few policy levers have been widely
studied. There has been some effort by researchers to understand whether policy
changes can reduce substance use, with results indicating that reductions in the
supply of or access to illicit drugs and prescription drugs with high potential for
dependence may reduce foster care caseloads (Cunningham and Finlay 2013;
Gihleb, Giuntella, and Zhang 2019; Markowitz et al. 2011; Quast, Storch, and
Yampolskaya 2018). Overall, the capacity of public policy to prevent and treat
substance use and addiction is not altogether clear. However, substance use
drives a substantial proportion of foster care entries and child maltreatment cases
(Child Welfare Information Gateway 2014; U.S. General Accounting Office
1998), and its prevention and treatment should be a component of any compre-
hensive child maltreatment prevention strategy.
Another potential policy lever that has received little attention is expansion of
health insurance coverage and affordable health care. Child welfare agencies
have limited resources to spend on services for families, and there is little evi-
dence that the services provided are effective at preventing and ameliorating the
effects of maltreatment (Jonson-Reid et al. 2017). Yet many of the evidence-
based services for reducing child maltreatment or associated risk factors are
covered by health insurance, including various mental health and substance use
services and home visiting programs. Child welfare agencies have limited
resources and contract with a limited number of service providers, which may
constrain the quality of services as well as the number of families to whom such
services are offered (Child Welfare Information Gateway 2014). Insured parents
are better positioned to seek treatment from providers outside of the narrow
network of child welfare system contractors, which may allow parents to access
effective treatments prior to the onset of child maltreatment and also increase
the quality of services available to families in the child welfare system. However,
approximately 11 million parents are uninsured (Karpman et al. 2016), and many
more may be underinsured or face high out-of-pocket costs. Expansion of health
care coverage and reduction of out-of-pocket health care costs may increase
uptake and quality of services for at-risk parents.

Quality of data on risk factors for child maltreatment

The most common approaches to investigating risk factors for child maltreatment
are to examine characteristics associated with child welfare system involvement, or
The Scope, Nature, and Causes of Child Abuse and Neglect 43

to rely on parent self-report of maltreatment behavior or substandard parenting

(i.e., concerning parent practices or environments that may not reach the legal
threshold for reporting to CPS).
The only annual national data available on CPS cases, NCANDS, contain several
variables related to risk factors for maltreatment, but there are significant limitations
to the data. First, the dataset includes only those cases of maltreatment that are
reported to CPS. As such, it is not possible to understand individual-level causal
processes leading to maltreatment, because there is no counterfactual. Second, the
quality of data is poor. Many states report no data on risk factors, whereas others
report information that is highly suspect. Data on parental drug use illustrates these
concerns (Seay 2015). In NCANDS in 2017, several states reported parental drug
use as a risk factor in less than 5 percent of substantiated cases, despite strong indi-
cations that substance use is a major factor in child maltreatment and child welfare
system involvement in particular (Berger et al. 2010; Brook and McDonald 2009;
Child Welfare Information Gateway 2014; Murphy et al. 1991; Ross 1997). Although
data quality problems are well known, states have little incentive to improve the
quality of their NCANDS submissions. NCANDS submission is voluntary, and the
Children’s Bureau1 indicates that the purpose of the data is “to examine trends in
child abuse and neglect across the country”—a purpose that is of little direct value
or consequence to the day-to-day operations of state agencies.
In addition, survey data may not provide reliable information on critical risk
factors. Returning to the example of parental substance use, the accuracy of
­self-reported substance use is low in populations already involved with public
systems (Garg et al. 2016; Peters, Kremling, and Hunt 2015; Rendon et al. 2017).
In a review of different estimates of parental substance use among child welfare
­system–involved parents, parent self-report produced lower rates than case-
worker report or case record reviews (Seay 2015).
In addition to underreporting, reliance on parent or caregiver reports, as is
common in surveys, may produce findings that are driven or inflated by common
method variance. Common method variance occurs when measures are pro-
duced using the same reporter: in survey data, the parent may be the source of
information for both the risk factors and the maltreatment measures. It is diffi-
cult to ascertain how much of a problem this is; and for some constructs, alterna-
tive sources of measurement may be unavailable or unreliable. Recent
meta-analyses of risk factors for neglect (Mulder et al. 2018) and sexual abuse
(Assink et al. 2019) did not find that the source of information on neglect (child
welfare records or self-report) explained heterogeneity in effects of various risk
factors. Other meta-analyses of risk factors for child maltreatment have not
addressed this question (Stith et al. 2009).

Conclusion
Child maltreatment is a complex problem with far-reaching consequences for
victims. Decades of research from around the world has provided significant
44 THE ANNALS OF THE AMERICAN ACADEMY

information about the ways in which maltreatment affects victims and the cir-
cumstances that increase the likelihood for maltreatment to occur. Despite this
wealth of knowledge, significant issues with defining and measuring child abuse
and neglect remain, which critically limit the knowledge to be gained. Official
child abuse and neglect definitions vary by state, and these definitions are not
necessarily consistent with those used in research, which also commonly vary
from one study to the next.
We offer several recommendations for improving science, policy, and practice
related to child maltreatment. First, we propose standardizing definitions of mal-
treatment within child welfare systems. Current variation in such definitions across
states limits the utility of cross-state comparative studies and undermines national
estimates of prevalence, incidence, and trends in child maltreatment. Further,
although child abuse is commonly understood to have subtypes (e.g., sexual, physi-
cal, psychological), child neglect is not typically separated in this same way, despite
the diverse array of parental acts and omissions that are captured by the overarch-
ing term “neglect” (e.g., supervisory neglect, emotional neglect, physical needs
neglect). Though researchers have long urged the capture of this information
(Slack et al. 2003), there has been no improvement to measures in NCANDS or
other administrative datasets. Given that neglect is the most common form of mal-
treatment within the U.S. child welfare system, more nuanced information is
needed on the causes and consequences of its various subtypes.
Second, we propose improving the measurement of child maltreatment, given
the set of standardized definitions. All states are required to document child
abuse and neglect through their systems, but submitting data to NCANDS is a
voluntary process, and limited information is available. Notably, however, com-
pulsory federal reporting (such as the Adoption and Foster Care Reporting and
Analysis System) is also plagued with problems, which may in part reflect the
dysfunctional and antiquated data systems used in many states (Font 2020).
Current plans for the Comprehensive Child Welfare Information Systems
(CCWIS) and corresponding data quality protocols (Federal Register 2016)
should emphasize the quality and timeliness of NCANDS submissions and
require complete and valid data on all NCANDS elements.
Third, social welfare programs within the United States gather a wealth of
information about families. Commonly, such information is unable to be easily
linked across systems, and concerns about privacy of individuals has hindered
progress in making such linkages. If linked, such data would facilitate evaluation
of state and federal policy or programmatic changes (e.g., expansions or retrac-
tions of public benefits). Advances in technology facilitate matching between
systems and careful planning and security protocols can address concerns about
privacy (Jonson-Reid and Drake 2008). Despite their limitations, administrative
or systems data are an invaluable tool in understanding the experiences and out-
comes of maltreated children who child welfare systems identify.
However, administrative data are insufficient to inform policy and practice,
given that not all cases of child maltreatment reach the attention of the child wel-
fare system. Thus, our final recommendation is that surveys include multiple-
source measures of maltreatment to bound estimates of the prevalence, antecedents,
The Scope, Nature, and Causes of Child Abuse and Neglect 45

and effects of child maltreatment. Further, there is a need for studies comparing
maltreatment models across datasets that use the same measures, to ascertain the
reliability of estimates across different sample parameters and time frames.
Moreover, existing survey measures could be greatly improved. Slack and col-
leagues (2003) critique overreliance on studies that predict child welfare involve-
ment or reinvolvement to understand the causes and correlates of child
maltreatment, given limitations and potential biases of system involvement as a
proxy for child maltreatment. To augment and validate current data focused on
child welfare involvement, prospective survey designs, including birth cohorts,
provide a viable option to track experiences and outcomes throughout childhood.
To understand the causes of maltreatment, a universal sample of children, rather
than focusing on children deemed to be “high risk” from prior research, is needed.

Note
1. See https://www.acf.hhs.gov/cb/research-data-technology/reporting-systems/ncands.

References
Amaya-Jackson, Lisa, Rebecca R. S. Socolar, Wanda Hunter, Desmond K. Runyan, and Rom Colindres.
2000. Directly questioning children and adolescents about maltreatment: A review of survey measures
used. Journal of Interpersonal Violence 15 (7): 725–59.
Annie E. Casey Foundation. 2012. Stepping up for kids: What government and communities should do to
support kinship families. Baltimore, MD: Annie E. Casey Foundation. Available from https://www.aecf
.org/resources/stepping-up-for-kids/.
Assink, Mark, Claudia E. van der Put, Mandy W. C. M. Meeuwsen, Nynke M. de Jong, Frans J. Oort,
Geert Jan J. M. Stams, and Machteld Hoeve. 2019. Risk factors for child sexual abuse victimization: A
meta-analytic review. Psychological Bulletin 145 (5): 459–89.
Bandura, Albert. 1978. Social learning theory of aggression. Journal of Communication 28 (3): 12–29.
Belsky, Jay. 1980. Child maltreatment: An ecological integration. American Psychologist 35 (4): 320–35.
Belsky, Jay. 1984. The determinants of parenting: A process model. Child Development 55 (1): 83–96.
Berger, Lawrence M., Sarah A. Font, Kristen S. Slack, and Jane Waldfogel. 2017. Income and child mal-
treatment in unmarried families: Evidence from the Earned Income Tax Credit. Review of Economics
of the Household 15 (4): 1345–72.
Berger, Lawrence M., Kristen S. Slack, Jane Waldfogel, and Sarah K. Bruch. 2010. Caseworker-perceived
caregiver substance abuse and child protective services outcomes. Child Maltreatment 15 (3): 199–210.
Biehl, Amelia M., and Brian Hill. 2018. Foster care and the earned income tax credit. Review of Economics
of the Household 16 (3): 661–80.
Bowlby, J. 1969. Attachment and loss. Vol. I: Attachment. London: Hogarth Press.
Bronfenbrenner, U. 1976. The ecology of human development: History and perspectives. Psychologia
Wychowawcza 19 (5): 537–49.
Brook, Jody, and Tom McDonald. 2009. The impact of parental substance abuse on the stability of family
reunifications from foster care. Children and Youth Services Review 31 (2): 193–98.
Brown, Jocelyn, Patricia Cohen, Jeffrey G Johnson, and Suzanne Salzinger. 1998. A longitudinal analysis
of risk factors for child maltreatment: Findings of a 17-year prospective study of officially recorded and
self-reported child abuse and neglect. Child Abuse & Neglect 22 (11): 1065–78.
Campbell, Kristine. 2019. Prevention of child maltreatment as an unexpected benefit of social policies.
JAMA Network Open 2 (6): e195521.
46 THE ANNALS OF THE AMERICAN ACADEMY

Cancian, Maria, Mi-Youn Yang, and Kristen S. Slack. 2013. The effect of additional child support income
on the risk of child maltreatment. Social Service Review 87 (3): 417–37.
Centers for Disease Control and Prevention. 2019. Behavioral risk factor surveillance system ACE data.
Available from https://www.cdc.gov/violenceprevention/childabuseandneglect/acestudy/ace-brfss.html.
Chan, Ko Ling. 2015. Are parents reliable in reporting child victimization? Comparison of parental and
adolescent reports in a matched Chinese household sample. Child Abuse & Neglect 44 (June): 170–83.
Child Welfare Information Gateway. 2014. Parental substance use and the child welfare system.
Washington, DC: U.S. Department of Health and Human Services, Children’s Bureau. Available from
https://www.childwelfare.gov/pubpdfs/parentalsubabuse.pdf.
Child Welfare Information Gateway. 2016. Definitions of child abuse and neglect. State statutes.
Washington DC: U.S. Department of Health and Human Services. Available from https://www
.childwelfare.gov/pubpdfs/define.pdf.
Conger, Rand, and G. Elder. 1994. Families in troubled times: Adapting to change in rural America. New
York, NY: de Gruyter.
Conger, Rand, X. Ge, G. Elder Jr., F. Lorenz, and R. Simons. 1994. Economic stress, coercive family
process, and developmental problems of adolescents. Child Development 65 (2): 541–61.
Conger, Rand, L. Ebert Wallace, Y. Sun, R. Simons, V. McLoyd, and G. Brody. 2002. Economic pressure
in African American families. Developmental Psychology 38 (2): 179–93.
Coulton, Claudia J., David S. Crampton, Molly Irwin, James C. Spilsbury, and Jill E. Korbin. 2007. How
neighborhoods influence child maltreatment: A review of the literature and alternative pathways. Child
Abuse & Neglect 31 (11–12): 1117–42.
Coulton, Claudia J., Jill E. Korbin, and Marilyn Su. 1999. Neighborhoods and child maltreatment: A multi-
level study. Child Abuse & Neglect 23 (11): 1019–40.
Cunningham, Scott, and Keith Finlay. 2013. Parental substance use and foster care: Evidence from two
methamphetamine supply shocks. Economic Inquiry 51 (1): 764–82.
Dolan, Melissa, Keith Smith, Cecilia Casanueva, and Heather Ringeisen. 2011. NSCAW II baseline report:
Introduction to NSCAW II final report. Washington, DC: Office of Planning, Research and Evaluation,
Administration for Children and Families, U.S. Department of Health and Human Services.
Doyle, Joseph J., Jr. 2007. Child protection and child outcomes: Measuring the effects of foster care.
American Economic Review 97 (5): 1583–1610.
Eamon, Mary Keegan, and Sandra Kopels. 2004. “For reasons of poverty”: Court challenges to child wel-
fare practices and mandated programs. Children and Youth Services Review 26 (9): 821–36.
English, Diana J., Sherry C. Brummel, J. Christopher Graham, and Laura K. Coghlan. 2002. Factors that
influence the decision not to substantiate a CPS referral. Phase II: Mail and telephone surveys of child
protective services social workers. Olympia, VA: Department of Social and Health Services.
English, Diana J., Richard Thompson, Catherine Roller White, and Dee Wilson. 2015. Why should child
welfare pay more attention to emotional maltreatment? Children and Youth Services Review 50
(March): 53–63.
Erickson, M. F., M. H. Labella, and B. Egeland. 2018. Child neglect. In The APSAC handbook of child
maltreatment, eds. J. B. Klika and J. R. Conte, 127–44. Thousand Oaks, CA: Sage Publications.
Federal Register. 2016. Comprehensive child welfare information system. 81 (106). Washington, DC: U.S.
Department of Health and Human Services, Administration for Children and Families, Administration
on Children, Youth and Families. Available from https://www.govinfo.gov/content/pkg/FR-2016-06-02/
pdf/2016-12509.pdf.
Fein, David J., and Wang S. Lee. 2003. The impacts of welfare reform on child maltreatment in Delaware.
Children and Youth Services Review 25 (1–2): 83–111.
Felitti, M. D., J. Vincent, M. D. Anda, F. Robert, M. D. Nordenberg, M. S. Williamson, F. David, et al.
1998. Relationship of childhood abuse and household dysfunction to many of the leading causes of
death in adults: The adverse childhood experiences (ACE) study. American Journal of Preventive
Medicine 14 (4): 245–58.
Finkelhor, David, and Lisa Jones. 2006. Why have child maltreatment and child victimization declined?
Journal of Social Issues 62 (4): 685–716.
Finkelhor, David, Kei Saito, and Lisa Jones. 2020. Updated trends in child maltreatment 2018. Durham,
NH: Crimes Against Children Research Center.
The Scope, Nature, and Causes of Child Abuse and Neglect 47

Finkelhor, David, Heather A. Turner, Anne Shattuck, and Sherry L. Hamby. 2015. Prevalence of childhood
exposure to violence, crime, and abuse: Results from the national survey of children’s exposure to
violence. JAMA Pediatrics 169 (8): 746–54.
Font, Sarah A. 2020. Data challenges and opportunities in child welfare. Washington, DC: American
Enterprise Institute. Available from https://www.aei.org/research-products/report/data-challenges-
and-opportunities-in-child-welfare/.
Font, Sarah A., and Lawrence M. Berger. 2015. Child maltreatment and children’s developmental trajec-
tories in early to middle childhood. Child Development 86 (2): 536–56.
Font, Sarah A., and Jamie Cage. 2018. Dimensions of physical punishment and their associations with
children’s cognitive performance and school adjustment. Child Abuse & Neglect 75:29–40.
Font, Sarah A., and Kathryn Maguire-Jack. 2020. It’s not “just poverty”: Educational, social, and economic
functioning among young adults exposed to childhood neglect, abuse, and poverty. Child Abuse &
Neglect 101 (March). Available from https://doi.org/10.1016/j.chiabu.2020.104356.
Font, Sarah A., Kathryn Maguire-Jack, and Rebecca Dillard. 2020. The decision to substantiate allegations
of child maltreatment. In Decision making and judgement in child welfare and protection: Theory,
research, and practice, eds. John D. Fluke, Monica Lopez, Rami Benbenishty, Erik J. Knorth, and
Donald J. Baumann. New York, NY: Oxford University Press.
Fox, N. A., and L. A. Leavitt. 1995. The violence exposure scale for children-revised (VEX-R). College
Park, MD: University of Maryland.
Garbarino, James. 1977. The human ecology of child maltreatment: A conceptual model for research.
Journal of Marriage and the Family 39 (4): 721–35.
Garg, Mahek, Laura Garrison, Lawrence Leeman, Ajna Hamidovic, Matthew Borrego, William F.
Rayburn, and Ludmila Bakhireva. 2016. Validity of self-reported drug use information among pregnant
women. Maternal and Child Health Journal 20 (1): 41–47.
Gelles, Richard J. 1973. Child abuse as psychopathology: A sociological critique and reformulation.
American Journal of Orthopsychiatry 43 (4): 611–21.
Gelles, Richard J. 1975. The social construction of child abuse. American Journal of Orthopsychiatry 45
(3): 363–71.
Gihleb, Rania, Osea Giuntella, and Ning Zhang. 2019. The effect of mandatory access prescription drug
monitoring programs on foster care admissions. Journal of Human Resources. Available from https://
doi.org/10.3368/jhr.57.1.0918-9729R2.
Gupta-Kagan, Josh. 2020. America’s hidden foster care system. Stanford Law Review 72. Available from
https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3437849.
Hardt, Jochen, and Michael Rutter. 2004. Validity of adult retrospective reports of adverse childhood
experiences: Review of the evidence. Journal of Child Psychology and Psychiatry 45 (2): 260–73.
Hussey, Jon M., Jane Marie Marshall, Diana J. English, Elizabeth Dawes Knight, Anna S. Lau, Howard
Dubowitz, and Jonathan B. Kotch. 2005. Defining maltreatment according to substantiation:
Distinction without a difference? Child Abuse & Neglect 29 (5): 479–92.
Hussey, Jon M., Jane Marie Marshall, Elizabeth Dawes Knight, Anna S. Lau, Howard Dubowitz, and
Jonathan B. Kotch. 2005. Defining maltreatment according to substantiation: Distinction without a
difference? Child Abuse & Neglect 29 (5): 479–92.
Hutchinson, Elizabeth D. 1990. Child maltreatment: Can it be defined? Social Service Review 64 (1):
60–78.
Jones, Lisa M., David Finkelhor, and Stephanie Halter. 2006. Child maltreatment trends in the 1990s:
Why does neglect differ from sexual and physical abuse? Child Maltreatment 11 (2): 107–20.
Jonson-Reid, Melissa, and Brett Drake. 2008. Multisector longitudinal administrative databases: An indis-
pensable tool for evidence-based policy for maltreated children and their families. Child Maltreatment
13 (4): 392–99.
Jonson-Reid, Melissa, Brett Drake, Patricia Kohl, Shenyang Guo, Derek Brown, Timothy McBride,
Hyunil Kim, and Ericka Lewis. 2017. What do we really know about usual care child protective ser-
vices? Children and Youth Services Review 82:222–29.
Karpman, Michael, Jason Gates, Stacey McMorrow, and Genevieve M. Kenney. 2016. Uninsurance among
parents, 1997–2014: Long-term trends and recent patterns. Washington, DC: Urban Institute.
Kohl, Patricia L., Melissa Jonson-Reid, and Brett Drake. 2009. Time to leave substantiation behind:
Findings from a national probability study. Child Maltreatment 14 (1): 17–26.
48 THE ANNALS OF THE AMERICAN ACADEMY

Kolko, D. J., and O. V. Berkout. 2018. Physical abuse of children. In The APSAC handbook on child mal-
treatment, eds. J. B. Klika and J. R. Conte, 111–26. Thousand Oaks, CA: Sage Publications.
Korbin, Jill E., and Richard D. Krugman. 2014. Handbook of child maltreatment. New York, NY: Springer.
Maguire-Jack, Kathryn, Kelly M. Purtell, Kathryn Showalter, Sheila Barnhart, and Mi-Youn Yang. 2019.
Preventive benefits of U.S. childcare subsidies in supervisory child neglect. Children & Society 33 (2):
185–94.
Markowitz, Sara, Alison Evans Cuellar, Ryan M. Conrad, and Michael Grossman. 2011. The effects of
alcohol policies in reducing entry rates and time spent in foster care. Cambridge, MA: National Bureau
of Economic Research.
McGee, Robin A., David A. Wolfe, Sandra A. Yuen, Susan K. Wilson, and Jean Carnochan. 1995. The
measurement of maltreatment: A comparison of approaches. Child Abuse & Neglect 19 (2): 233–49.
Milner, Jerry, and David Kelly. 17 January 2020. It’s time to stop confusing poverty with neglect. Chronicle
of Social Change. Available from https://chronicleofsocialchange.org.
Morton, Nicola, and Kevin D. Browne. 1998. Theory and observation of attachment and its relation to
child maltreatment: A review. Child Abuse & Neglect 22 (11): 1093–1104.
Mulder, Tim M., Kimberly C. Kuiper, Claudia E. van der Put, Geert-Jan J. M. Stams, and Mark Assink.
2018. Risk factors for child neglect: A meta-analytic review. Child Abuse & Neglect 77 (March): 198–
210.
Muller, Robert T., John E. Hunter, and Gary Stollak. 1995. The intergenerational transmission of corporal
punishment: A comparison of social learning and temperament models. Child Abuse & Neglect 19 (11):
1323–35.
Murphy, J. Michael, Michael Jellinek, Dorothy Quinn, Gene Smith, Francis G. Poitrast, and Marilyn
Goshko. 1991. Substance abuse and serious child mistreatment: Prevalence, risk, and outcome in a
court sample. Child Abuse & Neglect 15 (3): 197–211.
National Research Council. 2014. New directions in child abuse and neglect research. Washington, DC:
The National Academies Press. Available from https://doi.org/10.17226/18331.
Newberger, E. H., C. M. Newberger, and R. L. Hampton. 1983. Child abuse: The current theory base and
future research needs. Journal of the American Academy of Child Psychiatry 22 (3): 262–68.
Parke, R., S. Coltrane, S. Duffy, R. Buriel, J. Dennis, Justina Powers, Sabine French, and Keith F.
Widaman. 2004. Economic stress, parenting, and child adjustment in Mexican American and European
American families. Child Development 75 (6): 1632–56.
Paxson, Christina, and Jane Waldfogel. 1999. Parental resources and child abuse and neglect. American
Economic Review 89:239–244.
Peters, Roger H., Janine Kremling, and Elizabeth Hunt. 2015. Accuracy of self-reported drug use among
offenders: Findings from the Arrestee Drug Abuse Monitoring–II program. Criminal Justice and
Behavior 42 (6): 623–43.
Pinto, Ricardo J., and Ângela C. Maia. 2013. A comparison study between official records and self-reports
of childhood adversity. Child Abuse Review 22 (5): 354–66.
Putnam, Frank W., Marsha B. Liss, and John Landsverk. 2014. Ethical issues in maltreatment research
with children and adolescents. In Ethical issues in mental health research with children and adoles-
cents, eds. Kimberly Hoagwood, Peter S. Jensen, and Celia B. Fisher, 113–29. New York, NY:
Routledge.
Putnam-Hornstein, Emily. 2011. Report of maltreatment as a risk factor for injury death a prospective
birth cohort study. Child Maltreatment 16 (3): 163–74.
Quast, Troy, Eric A. Storch, and Svetlana Yampolskaya. 2018. Opioid prescription rates and child removals:
Evidence from Florida. Health Affairs 37 (1): 134–39.
Raissian, Kerri M. 2015. Does unemployment affect child abuse rates? Evidence from New York State.
Child Abuse & Neglect 48 (October): 1–12.
Raissian, Kerri M., and Lindsey Rose Bullinger. 2017. Money matters: Does the minimum wage affect
child maltreatment rates? Economic Causes and Consequences of Child Maltreatment 72 (January):
60–70.
Rebbe, Rebecca. 2018. What is neglect? State legal definitions in the United States. Child Maltreatment
23 (3): 303–15.
The Scope, Nature, and Causes of Child Abuse and Neglect 49

Rendon, Alexis, Melvin Livingston, Sumihiro Suzuki, Whitney Hill, and Scott Walters. 2017. What’s the
agreement between self-reported and biochemical verification of drug use? A look at permanent sup-
portive housing residents. Addictive Behaviors 70 (July): 90–96.
Ross, Jane L. 1997. Parental substance abuse: Implications for children, the child welfare system, and
foster care outcomes. Washington, DC: U.S. General Accounting Office.
Schelbe, L., and J. M. Geiger. 2017. Intergenerational transmission of child maltreatment. SpringerBriefs
in Social Work. New York, NY: Springer.
Schneider, William, Michael MacKenzie, Jane Waldfogel, and Jeanne Brooks-Gunn. 2014. Parent and
child reporting of corporal punishment: New evidence from the Fragile Families And Child Wellbeing
study. Child Indicators Research 8 (2): 347–58.
Schneider, William, Jane Waldfogel, and Jeanne Brooks-Gunn. 2016. The Great Recession and risk for
child abuse and neglect. Children and Youth Services Review 53:471–505.
Seay, Kristen. 2015. How many families in child welfare services are affected by parental substance use
disorders? A common question that remains unanswered. Child Welfare 94 (4): 19–51.
Sedlak, Andrea J., Jane Mettenburg, Monica Basena, Ian Peta, Karla McPherson, Angela Greene, and
Spencer Li. 2010. Fourth National Incidence Study of Child Abuse and Neglect (NIS-4). Washington,
DC: U.S. Department of Health and Human Services.
Shaw, C., and H. McKay. 1942. Juvenile delinquency in urban areas. Chicago, IL: University of Chicago
Press.
Slack, Kristen Shook, Jane Holl, Lisa Altenbernd, Marla McDaniel, and Amy Bush Stevens. 2003.
Improving the measurement of child neglect for survey research: Issues and recommendations. Child
Maltreatment 8 (2): 98–111.
Stith, Sandra M., Ting Liu, L. Christopher Davies, Esther L. Boykin, Meagan C. Alder, Jennifer M. Harris,
Anurag Som, Mary McPherson, and J. E. M. E. G. Dees. 2009. Risk factors in child maltreatment: A
meta-analytic review of the literature. Aggression and Violent Behavior 14 (1): 13–29.
Straus, Murray A., Sherry L. Hamby, David Finkelhor, David W. Moore, and Desmond Runyan. 1998.
Identification of child maltreatment with the Parent-Child Conflict Tactics Scales: Development and
psychometric data for a national sample of American parents. Child Abuse & Neglect 22 (4): 249–70.
Swahn, Monica H., Daniel J. Whitaker, Courtney B. Pippen, Rebecca T. Leeb, Linda A. Teplin, Karen M.
Abram, and Gary M. McClelland. 2006. Concordance between self-reported maltreatment and court
records of abuse or neglect among high-risk youths. American Journal of Public Health 96 (10):
1849–53.
U.S. Department of Health and Human Services. 2020. Child maltreatment 2018. Washington, DC.
Available from https://www.acf.hhs.gov/cb/resource/child-maltreatment-2018.
U.S. General Accounting Office. 1998. Foster care: Agencies face challenges securing stable homes for
children of substance abusers. GAO/HEHS-98-182. Washington, DC: U.S. General Accounting Office.
Available from https://www.gao.gov/assets/230/226295.pdf.
van IJzendoorn, Marinus H., Marian J. Bakermans-Kranenburg, Barry Coughlan, and Sophie Reijman.
2020. Annual research review: Umbrella synthesis of meta-analyses on child maltreatment antecedents
and interventions: Differential susceptibility perspective on risk and resilience. Journal of Child
Psychology and Psychiatry 61 (3): 272–90.
Wildeman, Christopher, Natalia Emanuel, John M. Leventhal, Emily Putnam-Hornstein, Jane Waldfogel,
and Hedwig Lee. 2014. The prevalence of confirmed maltreatment among US children, 2004 to 2011.
JAMA Pediatrics 168 (8): 706–13.
Yang, Mi-Youn, Kathryn Maguire-Jack, Kathryn Showalter, Youn Kyoung Kim, and Kristen Shook Slack.
2019. Child care subsidy and child maltreatment. Child & Family Social Work 24 (4): 547–54.
Yi, Youngmin, Frank R. Edwards, and Christopher Wildeman. 2020. Cumulative prevalence of confirmed
maltreatment and foster care placement for US children by race/ethnicity, 2011–2016. American
Journal of Public Health 110 (5): 704–9.