Child Neglect Article Review
Mi-Youn Yang Louisiana State University
Kathryn Maguire-Jack Ohio State University
Individual and Cumulative Risks for Child Abuse
and Neglect
Objective: The present study investigated whether risk factors vary by subtypes of child maltreatment investigations, with particular attention to the role of cumulative risks affecting child maltreatment. Background: Building and expanding on prior work finding that the accumulation of risk fac- tors puts children at risk for maltreatment, this work examines child physical abuse and neglect investigations separately. Method: A sample of 1,181 low-income fam- ilies was randomly selected from one state’s public assistance caseload. Multinomial logistic regression was used to model risk factors asso- ciated with 3 subtypes of investigated child mal- treatment reports. Results: Risk factors for each type of child mal- treatment were different. As the number of risk factors families experienced increased, the like- lihood of child maltreatment increased across all subtypes of maltreatment. Specifically, families with 5 or more risk factors were at greater risk of maltreatment than families with 2 or fewer risk factors. Conclusion: The threshold effect of cumulative risks demonstrates that families may be able to tolerate a moderate number of risk factors; however, beyond a certain number of risks,
School of Social Work, 209 Huey P. Long Field House, Louisiana State University, Baton Rouge, LA 70803-2804 (myyang7@gmail.com).
Key Words: child abuse, child neglect, cumulative risk.
families may not be able to cope adequately with the stress, and the likelihood of maltreating their children dramatically increases. Implications: To alleviate child maltreatment, prevention programs need to address diverse risks simultaneously, rather than focus on a par- ticular risk factor.
In 2014, an estimated 6.6 million children in the United States were reported to Child Protective Services (CPS), 3.2 million children received an investigation or assessment following the report of child maltreatment, and about 702,000 children were found to be victims of child maltreatment (U.S. Department of Health and Human Services, 2016). During the same year, CPS agencies provided 1.3 million children with services after a report of child maltreat- ment, including both in-home and foster care services (U.S. Department of Health and Human Services, 2016). Furthermore, 2.9 million chil- dren received services, such as home-visiting programs or childcare services, that aimed to prevent child maltreatment (U.S. Depart- ment of Health and Human Services, 2016). To provide these prevention and postresponse services, states spent $29 billion in state fiscal year 2014 (Rosinsky & Connelly, 2016). Child maltreatment has both short- and long-term consequences for a child’s development, aca- demic achievement, delinquency, substance abuse, and mental and physical health and has been associated with chronic victimization, perpetration of violence, and unemployment
Family Relations 67 (April 2018): 287–301 287 DOI:10.1111/fare.12310
288 Family Relations
(Brown, Cohen, Johnson, & Smailes, 1999; Corso, Edwards, Fang, & Mercy, 2008; Dube et al., 2003; Renner & Slack, 2006; Whitfield, Anda, Dube, & Felitti, 2003). To protect vulner- able children from maltreatment and to prevent further detrimental effects of child maltreatment on society, efficient and effective identification of children and families at risk is critical.
Previous studies have often relied on indi- vidual risk models that explore whether a specific risk factor is associated with child maltreatment (for a review, see Stith et al., 2009). However, risk factors tend to cluster together. Prior research on adverse childhood experiences showed that having one risk factor greatly increased the likelihood that another risk factor would be experienced (Felitti et al., 1998). Recognizing the co-occurrence of risk factors and the complex interaction between risk factors, cumulative risk models have sug- gested that the individual risk factors are less important than the total number of risk fac- tors families face (Appleyard, Egeland, van Dulmen, & Sroufe, 2005; Gutman, Sameroff, & Eccles, 2002; Nair, Schuler, Black, Ket- tinger, & Harrington, 2003; Popp, Spinrad, & Smith, 2008; Sameroff, Seifer, Barocas, Zax, & Greenspan,1987). A few studies have recently begun to focus on the predictive relationship between cumulative risks and child maltreat- ment. Empirically investigating two theoretical models of the developmental–ecological and cumulative risk models, Begle, Dumas, and Hanson (2010) found that a cumulative risk model predicts child abuse potential better than a specific risk model. However, the authors used parents’ self-report measures of child abuse potential rather than abusive behaviors from parental or administrative CPS reports. The authors additionally used cross-sectional data that measured risk factors and child abuse potential simultaneously, making it difficult to unravel the temporal association between presumed risk factors and child maltreatment. Similarly, MacKenzie, Kotch, and Lee (2011) found that the capacity of a cumulative risk model to predict future CPS involvement was more powerful than an individual risk model. MacKenzie et al. (2011) used a prospective lon- gitudinal design to examine predictors of CPS reports. However, MacKenzie et al. (2011) did not differentiate between types of maltreatment and instead relied on any CPS report. As a result, it was not possible for the authors to ascertain
whether the effects of tested cumulative risks varied by type of child maltreatment.
Previous research has shown that distinct risk factors are associated with the various subtypes of child maltreatment. For example, poverty is strongly associated with neglect but not with sexual abuse, and intimate partner violence is more strongly associated with physical abuse than with other types of child maltreatment (Brown, Cohen, Johnson, & Salzinger,1998; Finkelhor & Jones, 2006). In addition, despite the accumulated evidence that indicates chil- dren who experience multiple types of child maltreatment (e.g., both physical abuse and neglect) have greater adverse effects than those who only experience a single child maltreat- ment type (e.g., either physical abuse only or neglect only), little attention has been given to identifying distinct risk factors that are associated with multiple types of child maltreat- ment (Arata, Langhinrichsen-Rohling, Bowers, & O’Farrill-Swails, 2005; Edwards, Holden, Felitti, & Anda, 2003; Hahm, Lee, Oxonoff, & Van Wert, 2010; Higgins & McCabe, 2001).
Present Study
The present study was designed to investigate whether risk factors vary by subtypes of child maltreatment investigations (i.e., physical abuse only, neglect only, and investigations of both types), with particular attention to the role of cumulative risks affecting child maltreat- ment, using a prospective longitudinal research design. We focus on a sample of families who were receiving Temporary Assistance for Needy Families (TANF) benefits. Children in low-income families are more likely to expe- rience child maltreatment (Sedlak et al., 2010) and understanding the factors that contribute to maltreatment is critical for understanding this link further (Slack, Font, Maguire-Jack, & Berger, 2017). The risk factors examined were selected from prior studies investigating samples of TANF benefits. We seek to answer three main research questions: (a) How are individual risk factors related to investigated reports of physical abuse only, investigated reports of neglect only, and investigated reports that contain allegations of both? (b) Is the linear accumulation of risks associated with the three outcomes? and (c) Are there thresholds of risk associated with the three outcomes? The study is modeled after a semi- nal study from Appleyard et al. (2005), which
Risk for Child Abuse and Neglect 289
examined the linear and threshold aspects of risk related to child behavioral outcomes. The two key prior studies in this area examined the linear cumulative risks related to maltreatment and used child abuse potential and official reports of either abuse or neglect reports. The present study expands on the prior literature in two important ways. First, we examine two compet- ing methods for examining cumulative risk: the linear count and threshold effects. Second, we separately examine the outcomes of investigated physical abuse only, investigated neglect only, and investigations of both maltreatment types to determine whether there are differential effects by subtype of maltreatment.
Method
Sample
Data for the present study were derived from the Illinois Families Study (IFS). The sample of IFS was selected from TANF recipients of Illinois in the fall of 1998. In June 1998, the TANF caseload in Illinois was 122,720 (Lewis et al., 2000). To select a representative sample of TANF recipients, IFS carefully selected nine counties (Cook, St. Clair, Peoria, Tazewell, Know, Fulton, Woodford, Marshall, and Stark) considering population size, the level of urban- ization, and characteristics of demographics. A stratified random sampling design was used to ensure a sufficient number of TANF recipients from smaller and nonurban Illinois counties, as well as the large, metropolitan Chicago area were included in the sample (Lewis et al., 2000). The participants of IFS constituted a sample (N = 1,899) of 1998 TANF recipients in nine Illinois counties that together comprised 75% of the state’s TANF population. Potential respondents were contacted by phone to set up the initial face-to-face interview. Participants who strongly preferred a phone interview were provided with that option. Interviews took approximately 70 minutes to complete and were conducted in English or Spanish. Respondents received $30 for their participation, which was mailed to their home in the form of a money order (Lewis et al., 2000). The present study uses data from the first IFS interview, con- ducted between October 1999 and September 2000. Although 1,363 participants took part in the first interview (a 72% response rate), this analysis limits the sample to the caregivers
who consented to link their survey data with administrative data (n = 1,261). There were no statistically significant differences between the participants who chose to allow or not to allow their data to be linked (Lewis et al., 2002). The probabilistic match used personally identifying information of the caregiver and child (e.g., Social Security numbers, names, birthdates) to link survey respondents (caregivers) to the administrative dataset (Lewis et al., 2002). In addition, the final sample was limited to 1,181 caregivers who had at least one child of age 15 years or younger at the time of the first inter- view to allow at least 2 years of CPS observation periods before children became legal adults.
Table 1 depicts the sample characteristics for the full sample and for families by types of CPS involvement. The majority of the sample was non-Hispanic Black (78.8%) and not liv- ing with a partner (84.8%). One of every five caregivers was a teenager at the time of their first child’s birth. Large minorities of the care- givers had three or more children (42.4%) and did not have a high school diploma or General Educational Development (GED; 41.7%). At the time of the initial survey interview, about half of the caregivers were experiencing economic strain such as TANF receipt (55.3%), unemploy- ment (49.2%), or reported material hardship, such as rent arrears or utility shutoffs (56.1%). About one in six caregivers (16.7%) felt that their financial situation was much worse than the norm in their neighborhoods. Nearly a quarter (22.5%) had at least one child who had a chronic health condition that limited the child’s activi- ties. About one in 20 of the caregivers reported alcohol or drug use problems (4.9%), severe physical intimate partner violence (4.7%), or a learning disability (4.9%). A lack of social sup- port was reported by 40.0%, and 15.5% reported difficulty arranging emergency childcare when it was needed. Just under half (48.49 %) reported having spanked their children as a form of dis- cipline. Two years after the first interview date, 13.4% of the survey respondents had been inves- tigated by CPS: 4.3% for physical abuse only, 5.0% for neglect only, and 4.1% for investiga- tions of both types. There were statistical differ- ences in many of the risk factors by type of CPS involvement. Families who were investigated by any type of child maltreatment were experienc- ing more risks than families who were not inves- tigated by CPS.
290 Family Relations
Table 1. Weighted Sample Characteristics (N = 1,181)
Full sample Type of maltreatment (%)
Variables n % None
(n = 1,022)
Only physical (n = 51)
Only neglect (n = 59)
Both (n = 49) 𝜒2 (3) p
Overall 1,181 100.0 86.6 4.3 5.0 4.1 Demographic characteristics
Race or ethnicity Non-Hispanic Black 931 78.8 78.9 78.3 83.4 71.7 2.24 .677 Non-Hispanic White 89 7.6 6.5 10.0 10.3 24.5 22.73 <.001 Hispanic and others 161 13.6 14.6 11.7 6.3 3.8 7.67 .245
Marital status Single 1002 84.8 85.5 78.9 82.9 78.5 3.43 .535 Married 119 10.1 10.0 15.2 8.3 9.5 1.71 .785 Unmarried, living with partner 60 5.1 4.5 5.9 8.8 12.0 7.09 .178
Teen mom (first birth) 257 21.8 20.3 35.3 20.6 40.9 17.26 .013 ≥3 children 501 42.4 40.0 62.5 62.3 47.5 20.90 .006 No high school degree or GED 492 41.7 40.4 43.2 58.1 46.3 7.71 .204
Risk factors Economic characteristics
TANF 653 55.3 54.1 65.4 61.8 61.3 4.43 .461 No work 582 49.2 47.2 52.8 72.2 61.7 17.47 .012 Any material hardship 662 56.1 54.4 75.7 58.3 68.7 12.43 .048 Perceived hardship 197 16.7 15.3 26.2 32.0 16.3 14.74 .035
Child and mother’s well-being Child’s health condition that limits activities 266 22.5 21.0 23.9 31.1 42.3 14.76 .029 Depressive symptoms 278 23.5 21.5 19.0 32.6 59.3 40.10 <.001 Alcohol/drug use 58 4.9 4.6 4.8 4.6 11.5 4.85 .325 Severe intimate partner violence 56 4.7 3.4 11.9 7.9 20.7 38.62 <.001 Learning disability 58 4.9 4.2 10.3 6.6 12.0 10.09 .094 Lack of social support 472 40.0 37.2 52.3 38.9 65.8 18.11 .013 Childcare concern 183 15.5 14.3 20.1 29.4 19.4 11.21 .087
Parenting characteristics Parenting stress 217 18.4 17.3 17.0 29.3 30.0 9.96 .109 Spanking 573 48.5 46.6 59.2 67.0 55.2 12.88 .055
Cumulative levels of risks Low risks (0–2 risks) 402 34.0 35.3 25.0 11.9 8.5 29.03 <.001 Average risks (3–4 risks) 425 36.0 35.6 24.5 25.3 28.1 5.93 .297 High risks (≥5 risks) 354 30.0 29.1 50.5 62.8 63.4 58.08 <.001
Note. GED = General Educational Development; TANF = Temporary Assistance for Needy Families.
Measures
Child maltreatment investigation. Given that the present study included CPS reports in the State of Illinois, the following background infor- mation is provided for context. In the state of Illinois, there is no universal mandated reporting law for all adults. Instead, individuals who have contact with children, such as physical and behavioral health professionals, teachers, and
childcare workers, are mandated to report sus- pected maltreatment. The definitions for abuse and neglect in Illinois fall under Comp. Stat. Ch. 325, § 5/3. Acts are considered physical abuse if the caregiver inflicts, causes or allows to be inflicted, or creates a substantial risk of physical injury (by other than accidental means) that causes death, disfigurement, impairment of physical or emotional health, or loss or
Risk for Child Abuse and Neglect 291
impairment of any bodily function. Physical abuse relates to acts of torture, excessive cor- poral punishment, female genital mutilation, human trafficking, or involuntary servitude. Finally, presence of a child when metham- phetamine is being made is considered physical abuse. Acts are considered to be neglect if the child is not receiving adequate supervision, food, clothing, shelter, or medical treatment or if the child is born with controlled substances in his or her body. When reports are received, a CPS agency worker decides whether the information provided in the phone call meets statutory def- initions for child abuse and neglect, and when it does, the case is forwarded on for investigation.
Child maltreatment was operationalized using official CPS records of investigated child maltreatment reports in the 2 years after the first interview date. The outcome variable was categorized according to types of allegations: physical abuse only, neglect only, and inves- tigations of both physical abuse and neglect together. Investigated maltreatment reports were used in lieu of substantiated reports for two rea- sons: (a) the rate of substantiation in our sample is so low that we do not have the statistical power to detect these outcomes (1.50% for physical abuse, 1.01% for neglect, and 2.22% for both) and (b) prior research has found that families with unsubstantiated investigated reports have a similar rate of recidivism compared with families with substantiated reports (Drake, Jonson-Reid, Way, & Chung, 2003). Physical abuse and neglect were focused on exclusively because the rate of other forms of maltreatment was low among the sample (e.g., eight families had investigations of sexual abuse).
Demographic characteristics. The present study focuses on the variance of risk factors that tend to be proximal and changeable. Demographic controls were all categorical and included race or ethnicity (non-Hispanic White, non-Hispanic Black, Hispanic or other), marital status (mar- ried, unmarried but living with a partner, single), teen motherhood (mother was under 20 years of age at the time of her first birth), number of chil- dren in the household (caregiver had fewer than three vs. three or more children), and caregiver education (less than a high school education vs. at least a high school diploma or GED).
Risk factors. The selection of risk factors was based on prior investigations of this sample
and focused on changeable factors about a family (i.e., nondemographic characteristics; Yang, 2015; Yang & Maguire-Jack, 2016). The previous studies relied on a comprehensive meta-analysis conducted by Stith et al. (2009), which examined the major risk factors for child abuse and neglect. Thirteen risk factors were selected: four economic, two parenting, and seven physical and psychosocial well-being factors. The four economic risk factors included current TANF receipt, unemployment, any of six material hardships within the past 12 months (e.g., difficulty paying rent, utility shutoff, eviction), and perceived hardship (caregiver self-report that family financial situation is “much worse” than her neighbors’ financial situations). The two risk factors related to parenting were parenting stress (summary score of five items adapted from the Parenting Stress Index [Abidin, 1995], e.g., “child gets on nerves,” “child makes too many demands”; Cronbach’s 𝛼 = .73) and spanking (caregiver report of using spanking as a method of disci- pline). Seven risk factors viewed as affecting children’s and caregiver’s physical and psy- chosocial well-being were assessed: children’s health (caregiver report that at least one child had a health condition that limited regular activ- ity), caregiver depressive symptoms (summary score of 12 items adopted from the Center for Epidemiological Studies Depression scale [Radloff, 1977]; Cronbach’s 𝛼 = .92), frequent alcohol or drug use (caregiver report of consum- ing more than five drinks in 1 day on three or more occasions during the past year or use of marijuana, heroin, cocaine, or other illicit drug once or more in the past year), severe physical intimate partner violence (caregiver report of experiencing any of three physical violence incidents such as being hit, slapped, or kicked by partner or spouse in the past year), learning disability (caregiver has been told by a doctor, counselor, or teacher that she may have a learn- ing disability), lack of emotional social support (caregiver report of having few people to listen to her problems or encourage her in meeting her goals), and childcare concerns (caregiver report of difficulty being able to arrange emergency childcare when needed). Each risk factor vari- able was coded dichotomously (1 = presence of risk, 0 = absence of risk). To dichotomize con- tinuous measures (e.g., depressive symptoms and parenting stress), the 75th percentile cutoff was used (1 = score above the 75th percentile).
292 Family Relations
This cutoff point has been used by other risk researchers (Day, Ji, Dubois, Silverthorn, & Faly, 2016; Gerard & Buehler, 2004; Newsome, Vaske, Gehring, & Boisvert, 2016). We were unable to use the clinical cutpoint for the depres- sion measure because the IFS did not include all of the original items from the scale.
Cumulative risk index. The dichotomized scores on the 13 risk factors delineated in the preceding paragraphs were summed to create a cumulative risk score, which in these data ranged from 0 to 11 with a mean score of 3.5. In the first cumula- tive models, we relied on this linear measure of cumulative risk. In the next series of models, the cumulative risk score was then divided into three levels based on the proportion of the risk score distribution of the sample: low risk (no, one, or two risks; 34%, n = 402), moderate risk (three or four risks; 36%, n = 425), and high risk (five or more risks; 30%, n = 354), following the work of Appleyard et al. (2005), to test for threshold effects. The specific cutpoints were chosen to have a relatively equal proportion of families in each group.
Data Analysis
Stata version 15 (StataCorp, 2017) was used to complete all analyses. We first ran a series of bivariate multinomial logistic regressions to estimate the bivariate associations between the risk factors and the three types of inves- tigated maltreatment reports (physical abuse only, neglect only, and both physical abuse and neglect). We then ran a series of multivariate multinomial logistic regression models to iden- tify the risk factors for the three types of child maltreatment and to examine the effect of cumu- lative risk on the odds of the occurrence of these three outcomes, compared with no maltreat- ment (the reference category) while controlling demographic variables. Specifically, we first ran multivariate multinomial logistic regression models using the 13 individual risk factors for maltreatment separately (receipt of TANF, not working, material hardship, perceived hardship, child health problems, depression, alcohol and drug abuse, intimate partner violence, learning disability, lack of social support, childcare concerns, parenting stress, and spanking) to identify how the individual risks relate to the three outcomes. Second, we ran another multi- variate multinomial logistic regression models
using a linear measure of cumulative risk, which used a count measure of the number of risks. Finally, we ran the last set of multivariate multi- nomial logistic regression models using the three cutpoints of risk, comparing moderate-risk (three or four risks) and high-risk (five or more risks) to the low-risk group (no, one, or two risks). In all multivariate multinomial logistic regression, demographic variables such as race and ethnicity, marital status, teen motherhood, number of children, and high school degree were controlled. To adjust for the overrepresentation of small counties (a function of the stratified sample design) and nonresponse, statistical weights were used in all analyses.
Results
In the bivariate logistic regressions (see Table 2), we found that material hardship, intimate part- ner violence, and whether the caregiver had a learning disability were all correlated with being investigated for physical abuse only. We also found that unemployment, perceived hardship, depression, childcare concerns, parenting stress, and spanking were associated with being inves- tigated for neglect only. Regarding investiga- tions for both physical abuse and neglect, we found correlations with children’s health prob- lems, caregiver depression, alcohol and drug use, intimate partner violence, learning disabilities, lack of social support, and parenting stress.
The first objective of the multivariate analysis was to estimate the net effect of each risk factor on the three types of CPS involvement. Among other parameter estimates, Table 3 presents the adjusted odds ratio (OR) and p–value of each individual risk factor for the three outcomes along with the coefficient (B), standard error (SE), and 95% confidence intervals (CI). The adjusted OR reflects the risk of being inves- tigated for each type of child maltreatment category relative to the reference category (no investigated physical abuse or neglect). Eleven of the 13 individual risk factors were statistically associated with at least one type of child maltreatment. Four risk factors were associated with being investigated for physical abuse only: TANF receipt (OR = 1.9, p = .046); experience of any material hardship (OR = 2.2, p = .027); severe physical intimate partner violence (OR = 2.9, p = .042); and learning disability (OR = 3.9, p = .014). Four risk factors were associated with being investigated for
Risk for Child Abuse and Neglect 293
T ab
le 2.
B iv
a ri
a te
M u lt
in o m
ia l
L og
is ti
c R
eg re
ss io
n s
B et
w ee
n In
d iv
id u a l
R is
k F
a ct
o rs
a n d
O u tc
o m
es
P hy
si ca
l ab
us e
N eg
le ct
B ot
h
B S
E p
O R
95 %
C I
B S
E p
O R
95 %
C I
B S
E p
O R
95 %
C I
E co
no m
ic ch
ar ac
te ri
st ic
s T
A N
F 0.
47 0.
30 .1
16 1.
60 [0
.8 9,
2. 89
] 0.
32 0.
27 .2
47 1.
37 [0
.8 0,
2. 35
] 0.
30 0.
30 .3
26 1.
34 [0
.7 4,
2. 43
] N
o w
or k
0. 22
0. 29
.4 35
1. 25
[0 .7
1, 2.
20 ]
1. 07
0. 30
< .0
01 2.
91 [1
.6 2,
5. 20
] 0.
59 0.
30 .0
51 1.
80 [1
.0 0,
3. 26
] A
ny m
at er
ia l
ha rd
sh ip
0. 96
0. 33
.0 04
2. 61
[1 .3
6, 5.
01 ]
0. 16
0. 27
.5 61
1. 17
[0 .6
9, 1.
99 ]
0. 61
0. 32
.0 53
1. 84
[0 .9
9, 3.
42 ]
P er
ce iv
ed ha
rd sh
ip 0.
68 0.
33 .0
41 1.
97 [1
.0 3,
3. 75
] 0.
96 0.
29 .0
01 2.
60 [1
.4 7,
4. 61
] 0.
76 0.
40 .8
49 1.
08 [0
.4 9,
2. 35
] C
hi ld
an d
m ot
he r’
s w
el l-
be in
g C
hi ld
’s he
al th
co nd
it io
n 0.
16 0.
34 .6
27 1.
18 [0
.6 1,
2. 28
] 0.
53 0.
29 .0
68 1.
70 [0
.9 6,
3. 01
] 1.
01 0.
30 .0
01 2.
75 [1
.5 3,
4. 96
] D
ep re
ss iv
e sy
m pt
om s
–0 .1
6 0.
37 .6
64 0.
85 [0
.4 2,
1. 75
] 0.
57 0.
29 .0
49 1.
76 [1
.0 0,
3. 10
] 1.
67 0.
30 < .0
01 5.
31 [2
.9 4,
9. 60
] A
lc oh
ol an
d dr
ug ab
us e
0. 04
0. 67
.9 47
1. 05
[0 .2
8, 3.
91 ]
0. 00
0. 67
.9 97
1. 00
[0 .2
9, 3.
51 ]
1. 00
0. 47
.0 35
2. 72
[1 .0
7, 6.
88 ]
S ev
er e
IP V
1. 33
0. 47
.0 04
3. 80
[1 .5
2, 9.
47 ]
0. 88
0. 51
.0 84
2. 42
[0 .8
9, 6.
60 ]
2. 00
0. 39
< .0
01 7.
39 [3
.4 1,
15 .9
8] L
ea rn
in g
di sa
bi li
ty 0.
97 0.
49 .0
47 2.
64 [1
.0 2,
6. 85
] 0.
49 0.
55 .3
69 1.
63 [0
.5 6,
4. 76
] 1.
15 0.
47 .0
14 3.
15 [1
.2 6,
7. 89
] L
ac k
of so
ci al
su pp
or t
0. 58
0. 29
.0 45
1. 78
[1 .0
1, 3.
12 ]
0. 03
0. 27
.9 16
1. 03
[0 .6
0, 1.
76 ]
1. 14
0. 31
< .0
01 3.
11 [1
.7 0,
5. 71
] C
hi ld
ca re
co nc
er n
0. 41
0. 36
.2 52
1. 51
[0 .7
5, 3.
06 ]
0. 91
0. 30
.0 02
2. 49
[1 .3
9, 4.
48 ]
0. 36
0. 37
.3 30
1. 44
[0 .6
9, 3.
00 ]
P ar
en ti
ng ch
ar ac
te ri
st ic
s P
ar en
ti ng
st re
ss –0 .0
2 0.
38 .9
63 1.
01 [0
.5 1,
2. 01
] 0.
68 0.
30 .0
21 1.
87 [1
.0 7,
3. 28
] 0.
72 0.
32 .0
26 2.
66 [1
.4 8,
4. 81
] S
pa nk
in g
0. 51
0. 29
.0 80
1. 67
[0 .9
4, 2.
95 ]
0. 85
0. 28
.0 03
2. 33
[1 .3
4, 4.
07 ]
0. 35
0. 30
.2 41
1. 41
[0 .7
9, 2.
52 ]
N o
te .C
I =
co nfi
de nc
e in
te rv
al fo
r od
ds ra
ti o
(O R
); IP
V =
in ti
m at
e pa
rt ne
r vi
ol en
ce ;
T A
N F =
T em
po ra
ry A
ss is
ta nc
e fo
r N
ee dy
F am
il ie
s.
294 Family Relations
T ab
le 3.
M u lt
iv a ri
a te
M u lt
in o m
ia l
L og
is ti
c R
eg re
ss io
n s
B et
w ee
n In
d iv
id u a l
R is
k F
a ct
o rs
a n d
O u tc
o m
es
P hy
si ca
l ab
us e
N eg
le ct
B ot
h
B S
E p
O R
95 %
C I
B S
E p
O R
95 %
C I
B S
E p
O R
95 %
C I
E co
no m
ic ch
ar ac
te ri
st ic
s T
A N
F 0.
64 0.
32 .0
46 1.
91 [1
.0 1,
3. 60
] 0.
04 0.
30 .8
85 1.
04 [0
.5 8,
1. 88
] 0.
35 0.
34 .3
04 1.
42 [0
.7 3,
2. 76
] N
o w
or k
0. 00
0. 31
.9 99
1. 00
[0 .5
4, 1.
85 ]
0. 91
0. 32
.0 04
2. 49
[1 .3
3, 4.
68 ]
–0 .0
1 0.
35 .9
84 0.
99 [0
.5 0,
1. 95
] A
ny m
at er
ia l
ha rd
sh ip
0. 79
0. 36
.0 27
2. 20
[1 .1
0, 4.
41 ]
–0 .2
4 0.
30 .4
30 0.
79 [0
.4 4,
1. 42
] –0 .0
3 0.
36 .9
24 0.
97 [0
.4 8,
1. 96
] P
er ce
iv ed
ha rd
sh ip
0. 40
0. 37
.2 77
1. 49
[0 .7
3, 3.
05 ]
0. 77
0. 32
.0 17
2. 15
[1 .1
5, 4.
05 ]
–0 .6
8 0.
46 .1
40 0.
51 [0
.2 1,
1. 25
] C
hi ld
an d
m ot
he r’
s w
el l-
be in
g C
hi ld
’s he
al th
co nd
it io
n –0
.1 8
0. 37
.6 25
0. 83
[0 .4
0, 1.
72 ]
0. 08
0. 33
.8 03
1. 08
[0 .5
7, 2.
06 ]
0. 77
0. 35
.0 29
2. 15
[1 .0
8, 4.
27 ]
D ep
re ss
iv e
sy m
pt om
s –0
.7 7
0. 40
.0 54
0. 46
[0 .2
1, 1.
01 ]
0. 14
0. 33
.6 69
1. 15
[0 .6
0, 2.
20 ]
1. 14
0. 36
.0 01
3. 13
[1 .5
5, 6.
33 ]
A lc
oh ol
an d
dr ug
ab us
e –0
.3 4
0. 72
.6 34
0. 71
[0 .1
7, 2.
90 ]
–0 .3
6 0.
68 .6
02 0.
70 [0
.1 8,
2. 67
] 0.
82 0.
56 .1
45 2.
28 [0
.7 5,
6. 88
] S
ev er
e IP
V 1.
06 0.
52 .0
42 2.
89 [1
.0 4,
8. 05
] 0.
55 0.
55 .3
19 1.
73 [0
.5 9,
5. 05
] 1.
53 0.
46 .0
01 4.
64 [1
.9 0,
11 .3
2] L
ea rn
in g
di sa
bi li
ty 1.
35 0.
55 .0
14 3.
87 [1
.3 2,
11 .3
6] 0.
22 0.
58 .7
05 1.
25 [0
.4 0,
3. 89
] 1.
06 0.
55 .0
51 2.
90 [0
.9 9,
8. 45
] L
ac k
of so
ci al
su pp
or t
0. 45
0. 32
.1 63
1. 57
[0 .8
3, 2.
96 ]
–0 .3
4 0.
31 .2
75 0.
71 [0
.3 8,
1. 31
] 0.
81 0.
36 .0
23 2.
25 [1
.1 2,
4. 55
] C
hi ld
ca re
co nc
er n
0. 34
0. 38
.3 69
1. 41
[0 .6
7, 2.
96 ]
0. 72
0. 33
.0 28
2. 05
[1 .0
8, 3.
89 ]
0. 03
0. 42
.9 50
1. 03
[0 .4
5, 2.
35 ]
P ar
en ti
ng ch
ar ac
te ri
st ic
s P
ar en
ti ng
st re
ss –0
.4 1
0. 41
.3 17
0. 67
[0 .3
0, 1.
48 ]
0. 39
0. 33
.2 29
1. 48
[0 .7
8, 2.
82 ]
0. 18
0. 37
.6 17
1. 20
[0 .5
9, 2.
46 ]
S pa
nk in
g 0.
37 0.
31 .2
43 1.
44 [0
.7 8,
2. 66
] 0.
78 0.
31 .0
11 2.
17 [1
.1 9,
3. 96
] –0 .1
0 0.
35 .7
73 0.
91 [0
.4 6,
1. 78
]
N o
te .T
he m
od el
s co
nt ro
lf or
ra ce
/e th
ni ci
ty ,m
ar it
al st
at us
,t ee
n m
ot he
rh oo
d, nu
m be
r of
ch il
dr en
,a nd
hi gh
sc ho
ol di
pl om
a. C
I= co
nfi de
nc e
in te
rv al
fo r
od ds
ra ti
o (O
R );
IP V =
In ti
m at
e P
ar tn
er V
io le
nc e;
T A
N F =
T em
po ra
ry A
ss is
ta nc
e fo
r N
ee dy
F am
il ie
s.
Risk for Child Abuse and Neglect 295
Table 4. Multivariate Multinomial Logistic Regression Models Estimating Cumulative Risk and Outcomes
Physical abuse Neglect Both
B SE p OR 95% CI B SE p OR 95% CI B SE p OR 95% CI
Model 1
Number of risks 0.23 0.81 .005 1.25 [1.07, 1.47] 0.37 0.08 <.001 1.40 [1.20, 1.64] 0.48 0.09 <.001 1.62 [1.35, 1.94]
Model 2
Moderate
risks(low risks) –0.15 0.41 .711 0.86 [0.38, 1.92] 0.61 0.47 .190 1.84 [0.74, 4.60] 1.07 0.57 .062 2.92 [0.95, 9.01]
High risks(low risks) 0.76 0.36 .033 2.15 [1.06, 4.34] 1.63 0.42 <.001 5.09 [2.22, 11.68] 2.11 0.54 <.001 8.24 [2.86, 23.71]
Note. Reference category in parentheses. CI = confidence interval for odds ratio (OR); IPV = intimate partner violence. The models control for race/ethnicity, marital status, teen motherhood, number of children, and high school diploma.
neglect only: not working (OR = 2.5, p = .004); perceived hardship (OR = 2.2, p = .017); diffi- culty arranging emergency childcare (OR = 2.1, p = .028); and spanking (OR = 2.2, p = .011). Finally, four risk factors for being investigated for both physical abuse and neglect were identi- fied: child’s limiting health condition (OR = 2.2, p = .029), caregiver depressive symptoms (OR = 3.1, p = .001), severe physical intimate partner violence (OR = 4.6, p = .001), and lack of social support (OR = 2.3, p = .023). These results show that the risk factors for each type of child maltreatment were different except for severe physical intimate partner violence, which was associated with investigations of physical abuse only as well as investigations of both types of maltreatment.
The models examining the effects of cumula- tive risks on child maltreatment are presented in Table 4. Model 1 presents the continuous cumu- lative risk score and Model 2 presents results for the three levels of cumulative risk. Demo- graphic characteristics were controlled in both models. As Model 1 illustrates, as the number of risk factors increased, the risk of child mal- treatment investigation increased, regardless of the type of child maltreatment (physical abuse only OR = 1.3, p = .005; neglect only OR = 1.40, p < .001; being investigated for both types of maltreatment OR = 1.6, p < .001).
To check for a threshold effect of cumulative risk, the continuous cumulative risk score was recoded into three levels (i.e., low risk, moderate risk, and high risk). Model 2 in Table 4 shows that the odds of families with a moderate num- ber (three or four) of risks being investigated for child maltreatment was not statistically different from families with low risk (no, one, or two risks) across all types of child maltreatment. However, the odds of high-risk families (five or more risks) being investigated for child
maltreatment were statistically greater than the odds of families with low risk for physical abuse (p = .033), neglect (p < .001), and both (p < .001). Compared with families with low risks, families with five or more risks have two times higher odds of being investigated for physical abuse, five times higher odds of being investigated for neglect, and eight times higher odds of being investigated for both types of maltreatment.
To examine the relationship between number of risks and types of child maltreatment more closely, the marginal probability of being inves- tigated for child maltreatment for each number of risks is plotted in Figure 1. As the number of risk factors increases, the probability of being investigated for child maltreatment increases across all child maltreatment outcomes. Although the probability of being investi- gated due to physical abuse increases more linearly as the number of risks increases, the probability of being investigated due to neglect only or both types of maltreatment increases dramatically as the number of risks increases.
Discussion
The present study was designed to examine three types of models for assessing child maltreatment risk. First, we examined whether 13 individual risk factors were associated with child physical abuse investigations, child neglect investiga- tions, or investigations of both. Second, we examined whether linear cumulative risk was associated with these three outcomes. Finally, we examined whether there were threshold effects related to the cumulative risk and the three outcomes.
In terms of economic characteristics, we found interesting differences by type of mal- treatment. TANF receipt and experiencing
296 Family Relations
FIGURE 1. The marginal probability of each type of Child Protective Services report by the number of
risks.
0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
1 2 3 4 5 6 7
M ar
gi na
l P
ro ba
bi li
y of
C P
S R
ep or
t
Number of Risk Factors
Physical Abuse Neglect Both
Note. The model controlled for race/ethnicity, marital status, teen motherhood, number of children, and high school diploma.
any material hardship were associated with physical abuse investigations; not working and level of perceived material hardship were associated with neglect investigations. Con- sidering that the sample for this study was economically disadvantaged, those who were unemployed or who perceived their financial situation to be worse than their neighbors’ may be indicative of caregivers in deep poverty. Thus, these results also support the supposition that neglect is more strongly associated with deep poverty. Although previous studies have shown inconsistent findings regarding the role of poverty in child maltreatment, it is generally accepted that neglect is strongly associated with poverty (Brown et al., 1998; Chaffin, Kelleher, & Hollenberg, 1996).
Despite certain economic stress variables being associated with physical abuse or neglect, these variables were not associated with the risk of being investigated for both types of maltreatment. Although we do not know for sure why the variables were not statistically significant for having both physical abuse and neglect investigations, it is possible that families experiencing the most severe forms of poverty are being identified for neglect only, rather than both, because neglect may be the more pressing or obvious concern. Additionally, it is possible that for families that are experiencing
the greatest levels of dysfunction and having both physical abuse and neglect, other types of challenges are more important.
Regarding maternal and child well-being and individual types of maltreatment, severe physi- cal intimate partner violence was strongly asso- ciated with an increased risk for investigations of physical abuse only but not neglect alone, which is consistent with prior research (Brown et al., 1998; Finkelhor & Jones, 2006). A body of prior work has supported the relationship between use of emergency care in times of need for respite and reduced stress (Cole, Wehrmann, Dewar, & Swinford, 2005; Haynes-Lawrence, 2009; Sub- ramanian, 1985). Additionally, Ha, Collins, and Martino (2015) found that childcare disrup- tions, including having to make special backup arrangements when care fell through and having to miss work or school because arrangements fell through, were associated with child abuse and neglect. The results of the present study reveal that caregivers who lack access to emergency childcare are more likely to be investigated for child neglect. Supervisory neglect is a key com- ponent of neglect, and prior work has found a connection between childcare problems and neglect (Ha et al., 2015; Yang & Maguire-Jack, 2016). Improving access to high-quality child- care and providing quality backup childcare opportunities to reduce disruptions to childcare may aid in the prevention of child neglect.
Several of the maternal and child well-being variables were associated with the likelihood of being investigated for both abuse and neglect. Depressive symptoms, lack of social support, child’s chronic illness, and intimate partner vio- lence were associated with investigations for both types of maltreatment. This finding sug- gests that physical and socioemotional health issues have a more serious impact on parenting for those parents who are committing multiple forms of maltreatment compared with the eco- nomic stress factors.
In terms of parenting variables, we unex- pectedly found that spanking was associated with child neglect investigations, but not abuse investigations. Given the connection between spanking and physical abuse (Zolotor, Theodore, Chang, Berkoff, & Runyan, 2008), we had expected to see a relationship between spank- ing and physical child abuse. However, prior research has found a link between child neglect and spanking (Slack, Holl, McDaniel, Yoo, & Bolger, 2004).
Risk for Child Abuse and Neglect 297
In addition to the association between individ- ual risk factors and CPS involvement, we found that cumulative risk was statistically associated with child maltreatment investigation, irrespec- tive of the subtype. Families reporting more adverse experiences had substantially greater risk for CPS reports. This supports that fami- lies exposed to multiple risks may experience greater stress and have fewer resources which, in turn, may increase the likelihood of harsh par- enting or the disregard for children’s basic needs. Furthermore, the associations between the level of risk and child maltreatment outcomes have varying degrees of magnitude in the odds of being investigated for child maltreatment. Fami- lies with more than five risk factors are substan- tially more at risk of being investigated for child maltreatment than families who have fewer than three risk factors. These results are similar to the findings of Nair et al. (2003), which showed that child abuse potential was higher for caregivers with five or more risk factors than it was for care- givers with fewer than two risks. Furthermore, there was no difference between caregivers with two or fewer risks and caregivers with three or four risks. These results support the risk accu- mulation theory, which suggests that as the num- ber of risks increases and reaches a threshold, the likelihood of an adverse outcome on chil- dren and families increases (Gutman et al., 2002; Popp et al., 2008). The threshold effect of cumu- lative risks demonstrates that families may be able to tolerate a moderate number of risk fac- tors; however, beyond a certain number of risks, families may not be able to cope adequately with the stress, and the likelihood of maltreating their child(ren) dramatically increases.
The effect of the number of risks varies by child maltreatment subtype. Compared with the risk of being investigated due to a single type of child maltreatment (either physical abuse only or neglect only), the risk of being investi- gated for both physical abuse and neglect was dramatically higher for families with more risk factors than for families with fewer risk factors. As Figure 1 shows, when families experience fewer risks, the probability of being investigated for a single type of child maltreatment is larger than the probability of being investigated for both physical abuse and neglect. However, if a family experiences more than six risks, the probability of being investigated for both types of maltreatment exceeds the probability of being investigated for a single type of child
maltreatment. These findings suggest that fam- ily disorganization or disrupted family function indicated by a high level of cumulative risks may increase both acts of physical abuse and a failure to provide for a child’s basic needs. In addition, the association between high levels of cumula- tive risks and being investigated for both types of maltreatment may provide an explanation for why children experiencing multiple types of child maltreatment are more likely to expe- rience detrimental outcomes than children who experience a single type of child maltreatment.
Limitations
A number of limitations in the present study should be addressed. First, the outcome measure is based on investigated CPS reports rather than actual events of child maltreatment. The risk of being investigated by CPS may differ from the risk of actual child maltreatment (Miller-Perrin & Perrin, 2007). The drawback of a CPS mea- sure is that there are undoubtedly some cases of inappropriate reports that should not be con- sidered maltreatment. Indeed, only about 20% of children reported annually to CPS are found to be victims of maltreatment (U.S. Department of Health and Human Services, 2016). It is not clear how much of the disparities in these num- ber are a result of decision-making processes (e.g., whether maltreatment had occurred but there was not sufficient proof or workers chose not to substantiate because parents were com- plying with case plans) versus actual differences in maltreatment (Maguire-Jack & Byers, 2014). Additionally, individuals with risk factors, espe- cially those related to physical and behavioral health problems, receipt of public benefits, and experiencing intimate partner violence, probably have a higher likelihood of coming into contact with mandated reporters. As a result, some of the relations found in the present study may have been driven by surveillance bias, which in this context relates to parents being more likely to be reported to CPS if they are being helped by other service systems who employ mandated reporters due to having more “eyes” on the family com- pared with families that have less contact with mandated reporters.
Second, we did not differentiate between single and multiple CPS reports—that is, for example, between a one-time neglect report versus four neglect reports on the same family. Additionally, when measuring the physical
298 Family Relations
abuse and neglect together subtype, one-time co-occurring physical abuse and neglect investi- gated on a single occasion was not distinguished from multiple types of physical abuse and neglect investigated on multiple occasions. Relatedly, because we only followed families for 2 years, any maltreatment occurring outside of that time frame would not be included.
Third, although we examined physical abuse only, neglect only, and investigations of both types of maltreatment, we do not offer these as a continuum of severity. It is possible that a case with both abuse and neglect could be less severe than an extreme case that involved child torture or extreme deprivation, which may be investi- gated as physical abuse only or neglect only.
Fourth, because the sample was drawn during the late 1990s and was recruited from a sin- gle state and the majority of sample members were non-Hispanic Black, unmarried, TANF recipients, the results of the present study have limited generalizability. Also, to construct the cumulative risk score, continuous variables were dichotomized according to the sample distribution (e.g., depression, parenting stress). The depression and parenting stress levels of this sample might be higher than other popu- lation groups. Furthermore, risk groups were categorized according to the sample distri- bution. Those included in the moderate risk groups in this sample could be considered high risk in other samples. Thus, caution should be exercised when generalizing the results to other populations.
Finally, there are a number of macrolevel con- textual factors that likely influence the propen- sity to experience both the risk factors and child maltreatment reports investigated in this study. One such factor is institutional racism, which undoubtedly influences a number of negative outcomes for Black and Hispanic families within the sample. Although we are unable to account for this in this secondary data analysis, future research should investigate these relationships.
Conclusion and Implications
In terms of future research, the present study points to a number of possible directions. First, assessing the chronicity and severity of abuse and neglect reports will be essential for understanding whether the accumulation of risks is associated with more chronic and severe abuse, or whether, after the initial onset
of maltreatment, the number of risks is irrel- evant for the severity. Second, examining the pathways through which these risks relate to physical abuse and neglect will be essential to help inform the specific interventions provided to families. Specifically, if depression mediates the relationship between poverty and maltreat- ment, interventions could target both reducing poverty and improving mental health among those who are economically disadvantaged. Third, although we investigated the accumu- lation of risk, another approach for examining the relations between risks and maltreatment would be to use a latent class analysis approach to determine whether certain clusters of risks matter more for maltreatment compared with other risks. Such an approach would also allow for more nuanced interventions to be provided to families. Specifically, if poverty has a small impact on maltreatment, but poverty coupled with intimate partner violence has a large impact, then interventions should be targeted to individuals experiencing both.
Although no single approach will prevent maltreatment for all children, using a variety of methods to understand the complex ways in which parents experience risk is essential for informing intervention efforts. The findings of the present study suggest that both individual risk models and cumulative risk models pro- vide useful insights in the effort to prevent child maltreatment. The individual risk models show which specific risk factors have indepen- dent effects on different types of maltreatment and provide information about potential risk fac- tors to target for prevention efforts. Specifically, in this study, economic hardships were prob- lematic for abuse and neglect reports, as was intimate partner violence and caregiver depres- sion. Focusing efforts on these individual prob- lems may provide meaningful reductions in child abuse and neglect.
Additionally, the cumulative risk index can be an effective assessment tool for identifying the families at a high risk for child maltreat- ment. The cumulative risk perspective suggests that regardless of the specific risk, understand- ing the detrimental result of a high number of risks on child maltreatment is important. The Adverse Childhood Experiences Study has led to a proliferation of the use of screeners ask- ing about risk factors that occurred in childhood (e.g., Injury Prevention Center, n.d.; Wisconsin Children’s Trust, 2014). A similar tool could
Risk for Child Abuse and Neglect 299
be used to assess the number of risks currently being experienced by parents to assess the likeli- hood that problematic parenting may occur, and to provide an opportunity for prevention efforts, if warranted.
An intervention program that addresses only specific risk factors may not be effective in preventing child maltreatment for those who experience multiple risk factors. Thus, it may be more effective to offer comprehensive pro- grams that provide emergency money for food or housing, assistance with securing employ- ment, and access to affordable childcare service as well as psychological interventions in an effort to reduce depression. In addition, child development studies have shown that a family’s multiple risk factors have detrimental effects on child development, so alleviating the number of risks a family faces can promote both child well-being and safety.
Author Note
Administrative data linkages were developed by the Chapin Hall Center for Children, and survey data were collected by the Metro Chicago Infor- mation Center. We are grateful to Kristen Slack for her insights and comments.
References
Abidin, R. (1995).The parenting stress index (3rd ed.). Odessa, FL: Psychological Assessment Resources.
Appleyard, K., Egeland, B., van Dulmen, M. H., & Sroufe, L.A. (2005). When more is not better: The role of cumulative risk in child behavior out- comes. Journal of Child Psychology and Psychi- atry, 46, 235 – 245. https://doi.org/10.1111/j.1469- 7610.2004.00351.x
Arata, C. M., Langhinrichsen-Rohling, J., Bowers, D., & O’Farrill-Swails, L. (2005). Single ver- sus multi-type maltreatment: An examination of the long-term effects of child abuse. Journal of Aggression, Maltreatment & Trauma, 11, 29 – 52. https://doi.org/10.1300/J146v11n0402
Begle, A. M., Dumas, J. E., & Hanson, R. F. (2010). Predicting child abuse potential: An empirical investigation of two theoretical frameworks. Journal of Clinical Child & Adolescent Psy- chology, 39, 208 – 219. https://doi.org/10.1080/ 15374410903532650
Brown, J., Cohen, P., Johnson, J., & Salzinger, S. (1998). A longitudinal analysis of risk factors for child maltreatment: Findings of a 17-year prospec- tive study of officially recorded and self-reported child abuse and neglect. Child Abuse & Neglect,
22, 1065 – 1078. https://doi.org/10.1016/S0145- 2135(98)00087-8
Brown, J., Cohen, P., Johnson, J. G., & Smailes, E. M. (1999). Childhood abuse and neglect: Specificity of effects on adolescent and young adult depres- sion and suicidality. Journal of the American Academy of Child & Adolescent Psychiatry, 38, 1490 – 1496. https://doi.org/10.1097/00004583- 199912000-00009
Chaffin, M., Kelleher, K., & Hollenberg, J. (1996). Onset of physical abuse and neglect: Psychiatric, substance abuse, and social risk factors from prospective community data. Child Abuse & Neglect, 20, 191 – 203. https://doi.org/10.1016/ S0145-2134(95)00144-1
Cole, S. A., Wehrmann, K. C., Dewar, G., & Swinford, L. (2005). Crisis nurseries: Important services in a system of care for families and children. Children and Youth Services Review, 27, 995 – 1010. https:// doi.org/10.1016/j.childyouth.2004.12.023
Corso, P. S., Edwards, V. J., Fang, X., & Mercy, J. A. (2008). Health-related quality of life among adults who experienced maltreatment during childhood. American Journal of Public Health, 98, 1094 – 1100. https://doi.org/10.2105/AJPH .2007.119826
Day, J., Ji, P., Dubois, D. L., Silverthorn, N., & Flay, B. (2016). Cumulative social-environmental adversity exposure as predictor of psychological distress and risk behavior in urban youth. Child and Adolescent Social Work Journal. 33, 219 – 235. https://doi.org/10.1007/s10560-015-0421-5
Drake, B., Jonson-Reid, M., Way, I., & Chung, S. (2003). Substantiation and recidivism. Child Mal- treatment, 8, 248 – 260. https://doi.org/10.1177/ 1077559503258930
Dube, S., Felitti, V., Dong, M., Chapman, D., Giles, W., & Anda, R. (2003). Childhood abuse, neglect, and household dysfunction and the risk of illicit drug use: The adverse childhood experiences study. Pediatrics, 111, 564 – 572. https://doi.org/ 10.1542/peds.111.3.564
Edwards, V., Holden, G., Felitti, V., & Anda, R. (2003). Relationship between multiple forms of childhood maltreatment and adult mental health in community respondents: Results from the adverse childhood experiences study. American Journal of Psychiatry, 160, 1453 – 1460. https://doi.org/10 .1176/appi.ajp.160.8.1453
Felitti, V. J., Anda, R. F., Nordenberg, D., Williamson, D. F., Spitz, A. M., Edwards, V. . . . Marks, J. S. (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, 245 – 258. https://doi.org/ 10.1016/S0749-3797(98)00017-8
300 Family Relations
Finkelhor, D. & Jones, L. (2006). Why have child mal- treatment and child victimization declined? Jour- nal of Social Issues, 62, 685 – 716. https://doi.org/ 10.1111/j.1540-4560.2006.00483.x
Gerard, J., & Buehler, C. (2004). Cumulative envi- ronmental risk and youth maladjustment: The role of youth attributes. Child Development, 75, 1832 – 1849. https://doi.org/10.1111/j.1467-8624 .2004.00820.x
Gutman, L., Sameroff, A., & Eccles, J. (2002). The academic achievement of African American students during early adolescence: An examina- tion of multiple risk, promotive, and protective factors. American Journal of Community Psy- chology, 30, 367 – 399. https://doi.org/10.1023/A: 1015389103911
Ha, Y., Collins, M. E., & Martino, D. (2015). Child care burden and the risk of child maltreatment among low-income working families. Children and Youth Services Review, 59, 19-27. https://doi .org/10.1016/j.childyouth.2015.10.008
Hahm, H., Lee, Y., Ozonoff, A., & Van Wert, M. (2010). The impact of multiple types of child maltreatment on subsequent risk behaviors among women during the transition from adolescence to young adulthood. Journal of Youth and Ado- lescence, 39, 528 – 540. https://doi.org/10.1007/ s10964-009-9490-0
Haynes-Lawrence, D. (2009). Crisis nurseries: Emer- gency services for children and families in need. Exchange: The Early Childhood Leaders’ Maga- zine, 185, 16 – 20.
Higgins, D., & McCabe, M. (2001). Multiple forms of child abuse and neglect: Adult retrospec- tive reports. Aggression and Violent Behavior, 6, 547 – 578. https://doi.org/10.1016/S1359- 1789(00)00030-6
Injury Prevention Center, Phoenix Children’s Hos- pital. (n.d.). Overcoming adverse childhood experiences: Creating hope for a healthier Ari- zona. Retrieved from http://childhealthdata.org/ docs/drc/arizona-aces-efinal.pdf?Status=Master
Lewis, D. A., Shook, K., Lee, B. J., Kleppner P., Lewis, J., Riger, S., & Goerge, R. (2002). Wel- fare reform in Illinois: Is the moderate approach working? Second Annual Report from the Illinois Family Study. Evanston, IL: Institute for Policy Research, Northwestern University.
Lewis, D. A., Shook, K., Stevens A. B., Klepp- ner P., Lewis, J., & Riger, S. (2000). Work, wel- fare, and well-being: An independent look at wel- fare reform in Illinois: Project description and first-year report. Evanston, IL: Institute for Policy Research, Northwestern University.
MacKenzie, M. J., Kotch, J. B., & Lee, L. C. (2011). Toward a cumulative ecological risk model for the etiology of child maltreatment. Children and Youth Services Review, 35,1638 – 1647. https://doi.org/10 .1016/j.childyouth.2011.04.018
Mcguire-Jack, K., & Byers, K. (2014). The impact of prevention programs on decisions in child protec- tive services. Child Welfare, 92, 59 – 86.
Miller-Perrin, C., & Perrin, R. (2007). Child maltreat- ment: An introduction: Thousand Oaks, CA: Sage.
Nair, P., Schuler, M. E., Black, M. M., Kettinger, L., & Harrington, D. (2003). Cumulative environ- mental risk in substance abusing women: Early intervention, parenting stress, child abuse potential and child development. Child Abuse & Neglect, 27, 997 – 1017. https://doi.org/10.1016/S0145- 2134(03)00169-8
Newsome, J., Vaske, J. C., Gehring, K. S., & Boisvert, D. L. (2016). Sex differences in sources of resilience and vulnerability to risk for delin- quency. Journal of Youth and Adolescence. 45, 730 – 745. https://doi.org/10.1007/s10964-015- 0381-2
Popp, T. K., Spinrad, T. L., & Smith, C. L. (2008). The relation of cumulative demo- graphic risk to mother’s responsivity and control: Examining the role of toddler temperament. Infancy, 13, 496 – 518. https://doi.org/10.1080/ 15250000802329446
Radloff, L. (1977). The CES-D Scale: A self-report depression scale for research in the general. Applied Psychological Measurement, 1, 385 – 401. https://doi.org/10.1177/014662167700100306
Renner, L., & Slack, K. (2006). Intimate partner violence and child maltreatment: Under- standing intra- and intergenerational connec- tions. Child Abuse & Neglect, 30, 599-617. https://doi.org/10.1016/j.chiabu.2005. 12.005
Rosinsky, K., & Conelly, D. (2016). Child welfare financing SFY 2014: A survey of federal, state, and local expenditures. Washington, DC: Child Trends.
Sameroff, A., Seifer, R., Barocas, R., Zax, M., & Greenspan, S. (1987). Intelligence quotient scores of 4-year-old children: Social-environmental risk factors. Pediatrics, 79, 343 – 350.
Sedlak, A., Metternburg, J., Basena, M., Petta, I., McPherson, K., Green, A., & Li, S. Fourth National Incidence Study of Child Abuse and Neglect (NIS-4): Report to congress, executive summary. Washington, DC: Department of Health and Human Services, Administration for Children and Families.
Slack, K. S., Font, S. A., Maguire-Jack, K., & Berger, L. M. (2017). Predicting child protective services (CPS) involvement among low-income U.S. fam- ilies with young children receiving nutritional assistance. International Journal of Environmen- tal Research and Public Health, 14, 1197 – 1208. https://doi.org/10.3390/ijerph14101197
Slack, K. S., Holl, J. L., McDaniel, M., Yoo, J., & Bolger, K. (2004). Understanding the risks of child neglect: An exploration of poverty and parenting characteristics. Child Maltreatment, 9, 395 – 408. https://doi.org/10.1177/1077559504269193
Risk for Child Abuse and Neglect 301
StataCorp. 2017. Stata Statistical Software: Release 15. College Station, TX: StataCorp LLC.
Stith, S. M., Liy, T., Daview, L. C., Boykin, E. L., Alder, M. C., Harries, J. M., . . . Dee, J. E. M. E. G. (2009). Risk factors in child maltreatment: A meta-analytic review of the literature. Aggression and Violent Behavior, 14, 13 – 29. https://doi.org/ 10.1016/j.avb.2006.03.006
Subramanian, K. (1985). Reducing child abuse through respite center intervention. Child Welfare, 54, 501 – 509.
U.S. Department of Health and Human Services, Administration for Children and Families, Admin- istration on Children, Youth, and Families, Children’s Bureau. (2016). Child Maltreatment 2014. Washington, DC. Retrieved from https:// www.acf.hhs.gov/sites/default/files/cb/cm2014 .pdf#page=10
Whitfield, C., Anda, R., Dube, S., & Felitti, V. (2003). Violent childhood experiences and the risk of inti- mate partner violence in adults: Assessment in a large health maintenance organization. Journal of
Interpersonal Violence, 18, 166 – 185. https://doi .org/10.1177/0886260502238733
Wisconsin Children’s Trust Fund. (2014). Wisconsin ACE brief: 2011 and 2012 data. Child Abuse & Neglect Prevention Board. Retrieved from https://preventionboard.wi.gov/Pages/OurWork/ Publications.aspx
Yang, M. (2015). The effect of material hardship on child protective service involvement. Child Abuse & Neglect, 41, 113 – 125. https://doi.org/10.1016/j .chiabu.2014.05.009
Yang, M., & Maguire-Jack, K. (2016). Predictors of basic needs and supervisory neglect: Evidence from the Illinois Families Study. Children and Youth Services Review, 67, 20 – 26. https://doi.org/ 10.1016/j.childyouth.2016.05.017
Zolotor, A., Theodore, A., Chang, J. J., Berkoff, M. C., & Runyan, D. K. (2008). Speak softly—and forget the stick: Corporal punishment and child physical abuse. American Journal of Preventive Medicine, 35, 364 – 369. https://doi.org/10.1016/j .amepre.2008.06.031.
Copyright of Family Relations is the property of Wiley-Blackwell and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.