Child Abuse & Neglect 31 (2007) 445–461

Attributing responsibility for child maltreatment when

domestic violence is present夽
Miriam J. Landsman ∗ , Carolyn Copps Hartley
University of Iowa School of Social Work, 308 North Hall, Iowa City, IA 52240, USA
Received 9 June 2005; received in revised form 2 August 2006; accepted 20 August 2006

Abstract

Objective: The purpose of this study was to examine factors that influence how child welfare workers attribute
responsibility for child maltreatment and child safety in cases involving domestic violence.
Methods: The study used a factorial survey approach, combining elements of survey research with an experimental
design. Case vignettes were constructed by randomly assigning characteristics to vignettes believed to be related to
assessments about responsibility for child maltreatment. Public child welfare workers were systematically sampled
and asked to rate vignettes on male and female caregivers’ responsibility for child maltreatment and concerns for
safety.
Results: The presence of domestic violence significantly affected workers’ assessments of the attribution of respon-
sibility and concern for child safety, more so than variables related to child maltreatment. Responsibility for exposing
a child to domestic violence differed for males and females, with more factors explaining female responsibility.
Substance use by either caregiver was significant in attributing responsibility for physical harm, not watching the
child closely enough, and concern for child safety, but not for exposure to domestic violence.
Conclusions: Domestic violence appeared to heighten workers’ assessments of responsibility for child maltreatment
and concerns about child safety, taking precedence over the characteristics of the child maltreatment itself. Battering
tends to work against the domestic violence victim in terms of the attribution of responsibility. A greater number of
factors affect female responsibility for exposing a child to domestic violence than male responsibility, even though
in every case the male was the designated domestic violence batterer.
© 2007 Elsevier Ltd. All rights reserved.

Keywords: Child maltreatment; Domestic violence; Child welfare

夽
This study was funded by the Social Science Funding Program, University of Iowa, Office of the Vice President for Research.
∗
Corresponding author.

0145-2134/$ – see front matter © 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.chiabu.2006.08.008
446 M.J. Landsman, C.C. Hartley / Child Abuse & Neglect 31 (2007) 445–461

Introduction

The child welfare field is increasingly aware of families coming to the attention of child protection
agencies that are also experiencing domestic violence. The field is slowly beginning to understand that
the fundamental assumption of child protection—the expectation that parents must protect their children
from harm—may be complicated in situations where parents are also being victimized by their partners.
A recent review of studies on the co-occurrence of domestic violence and child maltreatment found
rates between 30 and 60% (Edleson, 1999). Child protection responses to the co-occurrence of child
maltreatment and domestic violence have raised tensions between child protection workers and battered
women’s advocates regarding the rights and safety of children versus the rights and safety of battered
women (Fleck-Henderson, 2000; Magen, Conroy, & Del Tufo, 2000). Child protection systems have been
criticized for holding only mothers responsible for the protection and safety of children, while fathers
have remained “invisible” in the child protection system (Edleson, 1998).
In families where domestic violence is present, this “legally sanctioned gender bias” is even more
apparent (Davis, 1995). As Mills (2000) notes, child welfare workers have strong views about battered
women’s responsibilities; they view the mother as the primary caregiver and often hold her to a higher
level of responsibility than her husband or partner to protect her children. Domestic violence advocates
feel this higher standard finds battered women being held unfairly responsible for failing to protect their
children from exposure to the batterers’ violence against them and/or from the risk of the batterers’ direct
physical abuse of the children (Davidson, 1995; Davis, 1995; Magen, 1999).
When looking at the attribution of responsibility for child maltreatment when domestic violence is
present, evidence supporting advocates’ concerns that battered women are being held responsible for
failure to protect their children from the batterer’s violence comes from qualitative interviews with workers
(Beeman, Hagemeister, & Edleson, 1999) and anecdotal observations of cases (Hartley, 2004). One of the
few empirical studies available found that even after domestic violence training, a substantial proportion of
child protection workers (40.5%) held women responsible for the child’s safety and stopping the batterer’s
violence against them (Saunders & Anderson, 2000). This study did not identify which characteristics of
the cases affected workers’ attribution of responsibility.
More recently, Coohey (in press) found that among the small proportion of child maltreatment incidents
involving failure to protect a child from domestic violence, child protection workers were less likely to
substantiate an allegation of exposure to domestic violence if the domestic violence victim engaged in
protective behaviors that ended contact between the batterer and her children. This same study determined
that all the batterers investigated who were in the caregiver role at the time of exposure to domestic violence
were substantiated. These findings, while encouraging, are based on 31 cases in one mid-western county.
The purpose of the current study was to examine factors that influence how child welfare workers attribute
responsibility for child maltreatment and child safety when domestic violence is also occurring in a family,
using an experimental survey design methodology with a large sample of cases.
Attributing responsibility for domestic violence to the victim occurs frequently in the general popula-
tion. In addition, alcohol use and the combination of race (Black) and alcohol create more victim-blaming
(Harrison & Esqueda, 2000). Bias occurs even among professionals (police and social workers), with
police placing more blame on victims then social workers, however male social workers attributed more
responsibility to victims than female social workers (Home, 1994).
Attributing responsibility for child maltreatment appears to be affected by the presence of domestic
violence. A domestic abuse victim’s perception of the seriousness of the domestic violence can impact
 M.J. Landsman, C.C. Hartley / Child Abuse & Neglect 31 (2007) 445–461 447

how workers’ perceive her responsibility for exposure to domestic violence. Domestic violence victims
who come in contact with the child protection system often are not voluntarily seeking services for the
domestic violence. They may appear “resistant” or “reluctant” to discuss the batterer’s violence against
them, and child protection workers find women’s “denial” or “minimization” of the violence very troubling
(Fleck-Henderson, 2000). Surveying frontline child welfare workers about their assessments of domestic
violence, Shepard and Raschick (1999) found that workers’ assessments were influenced by the severity
of the injury as well as the domestic violence victim’s behaviors—such as efforts to seek help. Workers
decisions may also be influenced by police involvement in child maltreatment cases, which is more likely
in cases involving domestic violence (Jones, Gross, & Becker, 2002).

Research questions

The current study was designed to address the overall research question: what are the net effects of
domestic violence factors, relative to other factors, on child welfare workers’ determination of parental
responsibility for child maltreatment? Specifically, what are the relative effects of domestic violence
factors, family demographics, and child maltreatment factors, on independent assessments of male and
female caregivers’ degree of responsibility for three types of maltreatment: causing physical injury to the
child, failing to provide adequate supervision, and exposing the child to domestic violence? The study
also examined the relative effects of these factors on child welfare workers’ assessment of child safety.
While most existing research on the study of the attribution of responsibility for child maltreat-
ment has been guided by correlational analysis rather than theoretical models, decision theory provides
a useful framework for studying this topic. Research on decision theory has been categorized into
three paradigms: formal-empiricist, rationalist, and naturalistic (Cohen, 1993; Fitch, 2001). The formal-
empiricist approach, which uses methods such as case vignettes, attempts to vary model parameters
systematically in order to develop normative explanations of decision-making behavior. This is the
approach in the current study. The rationalist perspective attempts to deconstruct the decision-making
process and to identify biases within this process, while the naturalistic perspective attempts to understand
decision-making as an adaptive process that individuals use in real-life situations (Cohen, 1993; Fitch,
2001). While rationalist and naturalistic perspectives are useful in field research, the formal-empiricist
approach is an important starting point for understanding factors that influence how child welfare workers
attribute responsibility for child maltreatment when domestic violence is present.

Methods

Design

The study methodology used a factorial survey approach to study human judgments through the
construction and evaluation of case vignettes (Rossi & Anderson, 1982). This method combined elements
of survey research with an experimental design. To test what factors are most important in assessing
parental responsibility for different types of child maltreatment, a sample of vignettes was developed.
Workers were asked to read a set these vignettes and assess the adult caregivers’ (male and female)
responsibility for three different types of child maltreatment.
448 M.J. Landsman, C.C. Hartley / Child Abuse & Neglect 31 (2007) 445–461

The first step in developing the vignettes was identifying a set of dimensions, the key variables that were
theoretically important to understanding judgments concerning the attribution of responsibility for child
maltreatment. Next the various levels of these dimensions, or a range of values that each dimension might
assume, were specified. Using a computer-generated program based on simple random sampling, one level
from each dimension was selected. For example, for the dimension of prior contact with child protection,
the computer-generated program selected one of three levels: no prior referrals; one prior referral; or
multiple reports to child protection. This selection method ensured that each level of each dimension had
an equal probability of being chosen. The computer-generated selection for all dimensions taken together
comprised a case vignette. Survey respondents were asked to assess the degree of responsibility of the
male and female caregiver for the child maltreatment in each case vignette. Since the purpose of this
study was to examine the attribution of responsibility for child maltreatment in the presence or absence
of domestic violence, the computer-generated program assigned each case vignette a child maltreatment
incident illustrating either a lack of supervision (neglect) or physical harm to the child (physical abuse).
In addition, the computer-generated program assigned either a co-occurring domestic violence incident
or no domestic violence.
To avoid overly legalistic interpretations of these fictitious case vignettes, the focus was on assessing
the degree of responsibility for three specific child maltreatment actions, not the substantiation of child
abuse or neglect. The three child maltreatment actions included: (1) not watching the child closely enough
(typically referred to as inadequate or lack of supervision), (2) causing physical harm to the child (or
physical abuse), and (3) exposing the child to domestic violence, a form of child maltreatment that is
increasingly being recognized in state statutes. Thus, when referring to exposure to domestic violence
throughout the paper, this means a type of maltreatment finding for which one or both caregivers could
be held responsible.
The survey accompanying each vignette asked respondents to read the vignette, check the type or types
of maltreatment they believed occurred, and rate the degree of responsibility, if any, each adult (male
and female) had for each relevant type of maltreatment (on a seven-point Likert scale, with 1 meaning
the individual was definitely not responsible and 7 indicating the individual definitely was responsible).
Although only one adult was specified as the lead actor for one type of child maltreatment in the vignette,
the workers could rate either or both adults as responsible for none or all three types of maltreatment
actions. In vignettes assigned domestic violence factors, domestic violence could have been present in
the case without respondents rating caregivers as responsible for exposing the child to domestic violence.
Respondents were also asked to rate their level of concern for the safety of the child involved in the
incident on a seven-point scale from no safety concerns to very serious safety concerns.

Case vignettes and measures

To develop the dimensions measured in this study, empirical literature was reviewed to identify the key
dimensions related to assessing responsibility for child maltreatment. This empirical literature includes
research on workers’ decisions to investigate and/or substantiate an allegation of child physical abuse
or neglect, as well as factors used to determine risk for subsequent maltreatment. Factors identified
include race (Eckenrode, Powers, Doris, Munsch, & Bolder, 1988; Fluke, Yuan, Hedderson, & Curtis,
2003; Fuller & Wells, 2003); family structure (Fuller & Wells, 2003; Radhakrisha, Bou-Saada, Hunter,
Catellier, & Kotch, 2001); parental substance abuse (Coohey, 2003; Craft, Epley, & Clarkson, 1980;
 M.J. Landsman, C.C. Hartley / Child Abuse & Neglect 31 (2007) 445–461 449

English, Marshall, Brummel, & Orme, 1999; English, Marshall, Coghlan, Brummel, & Orme, 2002;
Fuller & Wells, 2003; Hamilton & Browne, 1999; Wolock & Magura, 1996; Wolock, Sherman, Feldman,
& Metzger, 2001); severity of the injury (Coohey, 2003; English et al., 2002; Leiter, Myers, & Zingraff,
1994); prior involvement with child protective services (English et al., 1999; Fuller, Wells, & Cotton,
2001; English et al., 2002; Winefield & Bradley, 1992; Zuravin, Orme, & Hegar, 1995); direct evidence
identifying the abuser (Craft et al., 1980; Eckenrode et al., 1988; Jones, 1993); and parents’ response,
cooperation and assuming responsibility for the maltreatment (Coohey, 2003; English et al., 1999; English
et al., 2002).
Two expert consultants with experience in both child maltreatment and domestic violence reviewed the
list of dimensions and levels for content validity and clarity. They were also asked to complete the survey
for 10 vignettes. Debriefing sessions with the consultants were held to obtain feedback and make revisions
to the dimensions and their levels, as well as other survey items. The revised survey was subsequently pre-
tested with three graduate level social work students with experience in child welfare, domestic violence,
or both.
The final dimensions included in all vignettes were: race, legal relationship between adults, adults’
biological relationship to the child, prior child protective service contact, lead actor in the child maltreat-
ment incident, type and severity of child maltreatment, female and male caregiver’s alcohol use during
the maltreatment incident, and adults’ response to the child maltreatment incident. The lead actor in the
child maltreatment incident was the person described as initiating the child maltreatment act (i.e., leaving
a 3-year-old child unattended in the bathtub for lack of supervision or throwing a shoe that struck the
child for the physical abuse incident). Vignettes involving domestic violence contained an additional set
of dimensions: specific acts of the domestic violence incident based on the Conflict Tactics Scale (Straus,
1979), current and prior police involvement for domestic violence, female caregiver’s perception of the
seriousness of the domestic violence, and past protective actions taken by the female caregiver. Age of
the child was not randomized due to the need to limit the number of factors.
In each domestic violence vignette, the male was the identified batterer and the female was the victim
of the domestic abuse. This decision was made based on the predominance of male to female violence
and for methodological reasons in that including both male violence against females and female violence
against males would have required more dimensions and levels than was feasible given the sample size.
After generating the final sample of case vignettes, frequency distributions for the levels of each
dimension were calculated to verify that the random selection process resulted in equivalent distributions
of levels of the dimensions measured. For example, for the dimension on female caregiver intoxicated,
which had two levels (female caregiver intoxicated or female caregiver not intoxicated), roughly one-half
of the vignettes were assigned an intoxicated female caregiver and one-half a female caregiver who was
not intoxicated. The analysis of frequency distributions was repeated for the surveys that were returned,
and again, equivalent distributions of the levels of dimensions were found.

Sample

The respondents for this case vignette study were selected from a list of current front line child welfare
employees from a mid-western state. Two hundred currently employed child welfare assessment workers
(previously called child protective workers) and case managers (those with ongoing responsibility for
services and permanency planning) were systematically sampled from a list of all child welfare workers
employed in the state. Each employee was sent a packet with five unique randomly generated case
450 M.J. Landsman, C.C. Hartley / Child Abuse & Neglect 31 (2007) 445–461

vignettes and a few questions related to the employee’s demographics, position, education and training.
The survey was limited to five vignettes per respondent in order to increase the likelihood of participation.
The cover letter accompanying the survey explained the purpose of the study, voluntary participation,
and how confidentiality of responses would be protected. The University of Iowa’s Institutional Review
Board approved the study protocol.
Twenty-five envelopes were returned indicating that the individual was no longer employed at DHS or
had moved to another division. These individuals were removed from the sample. Additional follow-up
mailings were done to obtain a total number of 87 survey respondents out of 175 potential respondents,
for a final response rate of 49%. This is slightly lower than the 50% generally considered an adequate
response rate (Rubin & Babbie, 2005). Since the study uses the vignette rather than the respondent as the
unit of analysis, the surveys provided a total sample size of 435 case vignettes.
Of the 87 respondents, 38% were assessment workers and 55% were ongoing service workers for
children and families. The remaining 8% designated an ‘other’ child protection position (i.e., supervisor).
Respondents had a mean of 11.5 (SD = 7.62) years of experience in the child protection field, and a mean of
7.53 (SD = 6.61) years in their current position. Eighty-five percent of respondents were female. Most held
a Bachelor’s degree (88.5%), while 11.5% held a Master’s degree; 54% had a social work degree, while
other prominent degrees included psychology (18.4%) and sociology (12.6%). The remaining had degrees
in counseling, education, criminal justice, business, or political science. Most respondents had specialized
training in substance abuse (93%), domestic violence (90%), child development (92%), mental health
(85%) and cultural competence (75%). Respondents’ race was not asked, primarily because the Iowa
child welfare workforce is approximately 94% Caucasian; thus race could not be used in any meaningful
analysis. In addition, the researchers were concerned that asking respondents to identify their race might
heighten their sensitivity to that dimension in the case vignettes and possibly produce biased assessments.

Data analysis

The primary method of data analysis used least squares regression in which each dependent variable
was regressed on all of the levels of each dimension randomly assigned to create the vignettes. Since
factors were assigned using simple random sampling, an additive model with orthogonal factors was
assumed. Each level of each dimension was modeled as a dummy variable. Thus, for dimensions that had
more than two levels, one level was omitted from the equation to avoid perfect collinearity. There were
seven dependent variables in the study, resulting in seven regression models: male responsibility for not
watching the child closely enough, female responsibility for not watching the child closely enough, male
responsibility for physical harm, female responsibility for physical harm, male responsibility for exposing
the child to domestic violence, female responsibility for exposing the child to domestic violence, and
concern for the child’s safety. The first six models attempted to explain the degree of responsibility that
respondents assigned to the male and female caregivers for each type of child maltreatment action, using
the complete set of dimensions and levels. The seventh model attempted to explain the degree of concern
that respondents had about the child’s safety in the home, again using the complete set of dimensions and
levels. Below are results of these ordinary least squares (OLS) regression models; weighted least squares
was not used because there was insufficient prior research on which to base weights and OSL is robust
against violations of assumptions. The conventional probability level of .05, using two-tailed tests, was
used to determine statistical significance.
 M.J. Landsman, C.C. Hartley / Child Abuse & Neglect 31 (2007) 445–461 451

Results

Tables 1 and 2 present the unstandardized regression coefficients, standard errors, and p-values for
each variable in the model, the direction of the coefficient (a positive value increases responsibility, a
negative value decreases responsibility) as well as the probability value for each coefficient and the total
explained variation in the dependent variable for each model.

Responsibility for not watching the child closely enough

The degree of male responsibility for not watching the child closely enough was predicted by five
variables (those with a p-value of less than .05): the male was named as the lead actor in the child
maltreatment incident; the male was intoxicated at the time of the incident; domestic violence was part
of the vignette; the female caregiver had tried to leave the relationship; and the female caregiver had used
domestic violence services (Table 1). The first three variables had positive coefficients, which means
that these variables increased male responsibility for not watching the child closely enough; the last two
variables (the female caregiver had tried to leave in the past and had used domestic violence services)
had negative coefficients, indicating that these variables decreased male responsibility. The R2 for this
model indicates that 34.4% of variation in male responsibility for not watching the child closely enough
is explained by the model.
The degree of female responsibility for not watching the child closely enough was predicted by six
variables: the female was named the lead actor in the child maltreatment incident; the female was intox-
icated at the time of the incident; domestic violence was part of the vignette; the female caregiver was
slapped by the batterer; she was slammed into the wall by the batterer; and she was choked by the
batterer. All of these factors increased female responsibility for not watching the child closely enough.
This model explained 36.5% of variation in female responsibility for not watching the child closely
enough.

Responsibility for physically harming the child

The degree of male responsibility was predicted by four variables: the male was named the lead
actor in the child maltreatment incident; the male was intoxicated at the time of the incident; domestic
violence was part of the vignette; and the female caregiver was slammed into the wall during the domes-
tic violence incident (Table 1). The first three variables increased male responsibility for physically
harming the child; the last variable (the female was slammed into the wall) decreased male respon-
sibility. This model explained 49.3% of variation in male responsibility for physically harming the
child.
The degree of female responsibility for physically harming the child was predicted by six variables:
the female was named the lead actor in the child maltreatment incident; the incident was physical mal-
treatment; the female was intoxicated at the time of the incident; the female caregiver minimized the
seriousness of the maltreatment; there was no mention of police involvement for past domestic violence;
and the female caregiver was beaten up by the batterer. The first five factors increased female responsi-
bility for physically harming the child; the last factor decreased the female’s responsibility for physically
harming the child. This model explained 46% of variation in female responsibility for physically harming
the child.
452 M.J. Landsman, C.C. Hartley / Child Abuse & Neglect 31 (2007) 445–461

Table 1
Regression of male and female responsibility for dimensions and levels
Dimension and level Not watching child closely enough Physically harming child
Male responsibility Female responsibility Male responsibility Female responsibility
β (SE) p β (SE) p β (SE) p β (SE) p
Parents’ racea
African American .075 (.200) .708 −.241 (.250) .337 .423 (.250) .092 .050 (.265) .849
Hispanic −.128 (.200) .523 −.128 (.252) .611 .227 (.260) .385 −.085 (.277) .759
Parents’ legal relationshipb
Married .310 (.160) .053 .037 (.200) .853 .275 (.205) .180 .202 (.216) .351
Biological relationshipc
Female bio only .246 (.200) .220 .249 (.252) .325 −.316 (.257) .221.2 .309 (.277) .266
Male bio only −.034 (.201) .865 −.359 (.252) .154 −.270 (.258) .296 −.156 (.270) .565
Prior CPS contactd
More than one prior −.168 (.206) .415 .037 (.258) .886 .011 (.260) .967 .021 (.277) .941
referral
Multiple reports −.079 (.200) .694 .316 (.252) .211 .304 (.258) .240 .234 (.279) .402
Lead actor in incidente
Male caregiver 1.07 (.163) .000 −1.57 (.204) .000 2.324 (.211) .000 −2.44 (.223) .000
Type of maltreatmentf
Physical abuse −.280 (.169) .099 .225 (.212) .289 −.146 (.222) .511 .533 (.238) .026
Outcome of maltreatmentg
Laceration .181 (.202) .370 .170 (.254) .505 .433 (.256) .093 −.368 (.275) .183
Fracture .189 (.192) .325 .346 (.241) .152 .220 (.253) .386 −.162 (.269) .549
Male caregiver’s alcohol useh
Intoxicated .600 (.170) .000 −.153 (.214) .475 .443 (.213) .039 .357 (.227) .118
i
Female caregiver’s alcohol use
Intoxicated −.116 (.160) .471 .845 (.201) .000 −.070 (.206) .734 .937 (.218) .000
j
Adults’ response to child maltreatment
Female minimized −.220 (.254) .387 .329 (.315) .297 .067 (.315) .833 .719 (.331) .031
Female some .022 (.264) .932 −.246 (.330) .458 −.080 (.320) .802 .413 (.334) .218
responsibility
Male minimized .155 (.225) .492 −.385 (.283) .174 .308 (.297) .301.5 .382 (.320) .233
Male some .107 (.252) .672 −.094 (.314) .764 .196 (.307) .524 .558 (.330) .092
responsibility
Co-occurring domestic violencek
Domestic violence 1.71 (.166) .000 .887 (.207) .000 1.291 (.213) .000 .355 (.225) .166
incident
Severity of domestic violencel
Threatened knife −.026 (.375) .945 .700 (.466) .134 −.738 (.527) .162 −.730 (.543) .180
Kicked .060 (.394) .878 .944 (.501) .061 −.955 (.532) .074 −.784 (.550) .155
Slapped .073 (.434) .867 1.222 (.538) .024 −.714 (.574) .215 −.541 (.589) .359
Punched .226 (.364) .536 .600 (.455) .188 −.646 (.513) .209 −1.03 (.528) .053
Slammed into wall −.409 (.399) .306 .574 (.493) .246 −1.309 (.574) .023 −1.04 (.585) .077
Grabbed by arm .229 (.393) .561 1.024 (.485) .035 −.721 (.538) .181 −.424 (.551) .442
Choked .138 (.413) .739 1.48 (.513) .004 0.12 (.558) .983 .373 (.570) .513
Beat up .038 (.413) .928 .678 (.515) .189 −.327 (.548) .551 −1.177 (.559) .036
Shoved .175 (.411) .670 .762 (.509) .135 −.539 (.552) .330 −.609 (.558) .276
Threw vase at −.070 (.379) .855 .456 (.470) .333 .020 (.549) .971 .223 (.559) .690
Police involvement for current domestic violencem
Police called .145 (.201) .279 −.196 (.254) .442 .350 (.255) .171 −.370 (.269) .171
 M.J. Landsman, C.C. Hartley / Child Abuse & Neglect 31 (2007) 445–461 453

Table 1 ( Continued )
Dimension and level Not watching child closely enough Physically harming child
Male responsibility Female responsibility Male responsibility Female responsibility

β (SE) p β (SE) p β (SE) p β (SE) p

No mention .003 (.197) .473 −.247 (.253) .330 .392 (.265) .141 −.514 (.277) .064
Police involvement for past domestic violencen
Police called past .003 (.197) .986 −.056 (.246) .820 −.376 (.250) .134 .075 (.263) .775
No mention .005 (.201) .982 −.122 (.253) .630 −.131 (.268) .627 .669 (.284) .019
Female caregiver’s perception of seriousness of domestic violenceo
Female minimized −.281 (.226) .216 −.201 (.283) .478 .003 (.289) .992 .045 (.302) .882
Female worried −.433 (.229) .060 −.071 (.286) .803 .010 (.299) .972 .066 (.312) .833
about finances
Female caregiver −.165 (.240) .494 −.334 (.301) .269 −.221 (.291) .449 −.343 (.307) .264
exploring options
Protective actions taken by non-battering parentp
Called the police −.211 (.350) .546 −.733 (.441) .098 .840 (.480) .081 .249 (.503) .621
No-contact order −.120 (.338) .722 −.112 (.421) .790 .334 (.430) .437 .120 (.457) .792
Tried to end the −.566 (.346) .103 −.599 (.435) .170 .369 (.452) .415 .213 (.467) .649
relationship
Told him to leave −.396 (.338) .242 −.643 (.423) .129 −.213 (.443) .631 −.169 (.463) .716
Tried to leave herself −.929 (.348) .008 −.191 (.436) .662 .240 (.442) .588 .340 (.456) .456
Gone to shelter −.541 (.366) .140 .557 (.451) .218 .167 (.438) .703 .505 (.456) .269
Used domestic −.718 (.345) .038 −.066 (.439) .880 .437 (.430) .311 .273 (.472) .564
violence services
Safety plan −.194 (.336) .565 −.222 (.421) .598 −.112 (.458) .808 .701 (.473) .140
Explained variation and R2 = .389, N = 344 R2 = .363, N = 342 R2 = .493, N = 286 R2 = .460, N = 288
sample size
Reference Group.
a
Caucasian.
b
Not married.
c
Both bio related.
d
First contact.
e
Female caregiver.
f
Lack of supervision.
g
Bruise.
h
No mention male caregiver’s alcohol use.
i
No mention female caregiver’s alcohol use.
j
Female caregiver took seriously, male caregiver took seriously.
k
No domestic violence.
l
Twisted arm.
m
Police not called for current incident.
n
Police not called for past domestic violence incidents.
o
Female seeking help.
p
No protective actions noted.

Responsibility for exposing the child to domestic violence

The degree of male responsibility for exposing the child to domestic violence was predicted by only
two variables: the male was African-American, and domestic violence was part of the vignette (Table 2).
Both variables increased male responsibility for exposing the child to domestic violence, and the model
explained 39.2% of variation.
The degree of female responsibility for exposing the child to domestic violence was predicted by eight
variables: the female was the only caregiver biologically related to the child; the female was named the
454 M.J. Landsman, C.C. Hartley / Child Abuse & Neglect 31 (2007) 445–461

Table 2
Regression of male and female responsibility for dimensions and levels
Dimension and level Exposing child to domestic violence Concern for child safety

Male responsibility Female responsibility

β (SE) p β (SE) p β (SE) p
Parents’ racea
African American .321 (.147) .030 .275 (.257) .248 −.332 (.141) .023
Hispanic .124 (.147) .400 .130 (.257) .614 −.123 (.141) .392
Parents’ legal relationshipb
Married .027 (.116) .818 .288 (.203) .156 .008 (.113) .944
c
Biological relationship
Female bio only .128 (.147) .387 .802 (.258) .002 .163 (.141) .248
Male bio only .214 (.147) .147 .259 (.257) .314 −.146 (.141) .302
Prior CPS contactd
More than one prior referral −.045 .766 −.115 (.261) .661 −.300 (.145) .039
Multiple reports .102 .499 −.151 (.262) .566 .231 (.143) .107
Lead actor in incidente
Male caregiver .045 (.119) .708 −.886 (.208) .000 −.013 (.116) .914
f
Type of maltreatment
Physical abuse .112 (.124) .368 .989 (.217) .000 .206 (.115) .076
g
Outcome of maltreatment
Laceration .196 (.144) .175 .090 (.251) .721 .067 (.144) .640
Fracture .092 (.147) .532 .405 (.256) .116 .149 (.138) .279
Male caregiver’s alcohol useh
Intoxicated .097 (.125) .438 −.352 (.218) .108 .355 (.117) .003
i
Female caregiver’s alcohol use
Intoxicated .020 (.120) .869 .393 (.209) .061 .331 (.113) .004
j
Adults’ response to child maltreatment
Female minimized −.001 (.179) .995 .210 (.312) .501 .032 (.174) .856
Female some responsibility −.002 (.189) .992 −.161 (.334) .629 .103 (.186) .579
Male minimized −.019 (.169) .913 .006 (.295) .983 .045 (.164) .784
Male some responsibility −.046 (.180) .799 .103 (.314) .744 .027 (.176) .876
Co-occurring domestic violencek
Domestic violence incident 1.74 (.174) .000 1.34 (.304) .000 1.23 (.115) .000
l
Severity of domestic violence
Threatened knife .115 (.298) .699 .725 (.519) .164 .001 (.271) .998
Kicked .112 (.302) .711 1.059 (.528) .046 −.068 (.280) .809
Slapped −.265 (.332) .426 .599 (.579) .302 −.130 (.303) .669
Punched .203 (.281) .471 1.055 (.490) .033 .045 (.264) .865
Slammed into wall .185 (.318) .560 .595 (.554) .284 −.010 (.285) .971
Grabbed by arm .016 (.331) .961 .225 (.577) .697 .156 (.279) .577
Choked .471 (.301) .118 .169 (.524) .748 −.052 (.297) .862
Beat up .274 (.301) .363 .433 (.528) .413 .197 (.297) .507
Shoved .014 (.309) .963 .100 (.539) .853 .026 (.293) .928
 M.J. Landsman, C.C. Hartley / Child Abuse & Neglect 31 (2007) 445–461 455

Table 2 ( Continued )
Dimension and level Exposing child to domestic violence Concern for child safety
Male responsibility Female responsibility

β (SE) p β (SE) p β (SE) p

Threw vase at .382 (.296) .198 .933 (.516) .072 .264 (.277) .342
m
Police involvement for current domestic violence
Police called .204 (.145) .161 −.171 (.253) .500 .288 (.141) .042
No mention .175 (.150) .246 .156 (.263) .553 .197 (.143) .169
Police involvement for past domestic violencen
Police called past −.070 (.145) .628 .256 (.254) .314 −.123 (.139) .375
No mention .042 (.153) .781 .354 (.267) .186 −.251 (.145) .084
Female caregiver’s perception of seriousness of domestic violenceo
Female minimized −.108 (.167) .520 −.170 (.292) .562 .103 (.159) .517
Female worried about finances −.142 (.168) .398 .408 (.295) .167 .211 (.161) .189
Female caregiver exploring options −.230 (.180) .203 .134 (.316) .671 −.022 (.165) .896
Protective actions taken by non-battering parentp
Called the police .369 (.257) .153 −.424 (.448) .345 −.192 (.250) .442
No-contact order .056 (.245) .819 −.524 (.428) .222 −.287 (.239) .231
Tried to end the relationship .016 (.256) .951 −1.237 (.446) .006 −.335 (.244) .171
Told him to leave .002 (.249) .995 −.914 (.436) .037 −.485 (.243) .047
Tried to leave herself .290 (.251) .248 .151 (.438) .730 .041 (.244) .868
Gone to shelter .117 (.252) .643 −.177 (.440) .688 .148 (.258) .566
Used domestic violence services −.035 (.264) .895 .065 (.461) .889 −.041 (.247) .869
Safety plan −.051 (.250) .838 .043 (.436) .921 −.035 (.244) .885
Explained variation and sample size R2 = .392, N = 267 R2 = .347, N = 266 R2 = .352, N = 426
Reference Group.
a
Caucasian.
b
Not married.
c
Both bio related.
d
First contact.
e
Female caregiver.
f
Lack of supervision.
g
Bruise.
h
No mention male caregiver’s alcohol use.
i
No mention female caregiver’s alcohol use.
j
Female caregiver took seriously, male caregiver took seriously.
k
No domestic violence.
l
Twisted arm.
m
Police not called for current incident.
n
Police not called for past domestic violence incidents.
o
Female seeking help.
p
No protective actions noted.
456 M.J. Landsman, C.C. Hartley / Child Abuse & Neglect 31 (2007) 445–461

lead actor in the child maltreatment incident; the incident was a physical maltreatment incident; domestic
violence was part of the vignette; the female caregiver was kicked by the batterer; the female was punched
by the batterer; the female caregiver had tried to end the relationship with the batterer; and the female had
told the batterer to leave. The first six factors increased female responsibility for exposing the child to
domestic violence; the last two factors decreased female responsibility for exposing the child to domestic
violence. This model explained 34.7% of variation in responsibility for exposing the child to domestic
violence.

Concern for child’s safety

Respondents’ assessment of the degree of concern for the child’s safety was predicted by seven vari-
ables: the family was African-American; the family had had more than one prior referral to CPS; and the
female caregiver had told the batterer to leave decreased concerns about child safety. The male caregiver
was intoxicated at the time of the incident; female caregiver was intoxicated at the time of the incident;
domestic violence was part of the vignette; and police were called during the current domestic violence
incident increased concerns about child safety (Table 2). This model explained 35.2% of variation in
workers’ concerns about the child’s safety.

Discussion

This study examined factors that influence how child welfare workers attribute responsibility for child
maltreatment and child safety when domestic violence is present. This was accomplished using a formal-
empiricist approach that systematically varied aspects of case vignettes in order to develop normative
explanations of decision-making behavior. Several important conclusions can be drawn from the results.
Most notably, the findings suggest that the very presence of co-occurring domestic violence significantly
affects workers’ assessments of the degree of responsibility for male and female caregivers for failing
to watch a child closely enough, for exposing a child to domestic violence as a specific type of child
maltreatment, and for males, for physically harming a child, as well on workers’ assessments regarding
the child’s safety in the home. The only situation in which the presence of domestic violence did not have
a significant effect was in the assessment of the female caregiver’s responsibility for physically harming
a child. In this instance alone, co-occurring domestic violence did not influence the workers’ rating of
the female caregiver’s degree of responsibility.
A second and related conclusion from these findings is that neither the severity of injury to the child
nor the caregiver’s response to the maltreatment (whether minimizing the incident or taking it seriously)
has an effect on workers’ assessments of the degree of responsibility of either male or female caregivers
for any type of maltreatment. These findings differ from previous studies which have found that severity
of injury and the caretaker’s attitude to be related to the substantiation of child maltreatment (English et
al., 2002; Leiter et al., 1994) or risk of subsequent maltreatment (Coohey, 2003). Furthermore, while the
number of prior referrals to child protective services had an effect on the assessment of child safety, it
did not affect workers’ ratings of caregivers’ responsibility for child maltreatment. This is contrary to the
findings of some previous studies that have found prior referrals to be related to substantiation (English
et al., 2002; Winefield & Bradley, 1992; Zuravin et al., 1995) or reoccurring maltreatment (English et al.,
1999; Fuller et al., 2001).
 M.J. Landsman, C.C. Hartley / Child Abuse & Neglect 31 (2007) 445–461 457

The differences in findings between the studies referenced above and the current one may be partially
due to differences in explanatory variables and the samples used. With the exception of Coohey (2003),
none of the studies discussed here included domestic violence as an explanatory factor. Coohey (2003),
who did include domestic violence, focused exclusively on cases of supervisory neglect and found that
mothers who were battered were substantially more likely to be placed on the state child abuse registry
than mothers who were not. In the current study, which includes cases of both physical abuse and neglect,
the presence of domestic violence co-occurring with child maltreatment appears to overshadow the
net effects of maltreatment characteristics when attributing responsibility. The only child maltreatment
factor predictive of degree of responsibility was the lead actor named in the incident description. Not
surprisingly, the caregiver randomly assigned to this lead actor role was assessed as being responsible for
the maltreatment incident.
With respect to assessing level of responsibility for exposing a child to domestic violence, the differ-
ences between the models for male and female caregivers are striking. In all vignette scenarios involving
domestic violence, the male caregiver was the batterer. Yet, the regression models found many more
factors that had significant effects on female responsibility for exposing the child to domestic violence
than male responsibility. This is consistent with Coohey’s (in press) finding that more factors explained
female responsibility than male responsibility for failure to protect a child from domestic violence. The
only factors that significantly explained male responsibility was whether domestic violence was part of
the vignette and whether the family was African-American—both of which increased male responsibility
for the child maltreatment action of exposing the child to domestic violence. On the other hand, eight
factors were significantly related to female responsibility for exposing the child domestic violence, six
of which increased her responsibility (domestic violence was part of the vignette, the female was kicked
or punched during the incident having the largest effects). Only two factors, both of which involved the
female taking protective actions (having tried to end the relationship and having told the male to leave in
the past), decreased the female’s responsibility for exposing the child to domestic violence. These models
are consistent with Coohey’s (in press) study and provide empirical support for the heretofore anecdotal
suggestion that child welfare workers may be attributing responsibility for exposing a child to domestic
violence to the victim of domestic violence.
The notion that the domestic violence victim may be seen by child welfare workers as having some
responsibility for exposing her child to domestic violence is further supported by the non-intuitive effects
of specific acts of domestic violence on the attribution of responsibility for child maltreatment. When
such actions on the part of the male caregiver toward the female caregiver were significant—such as kick-
ing, slapping, punching, and choking—these almost always had the effect of increasing the child welfare
worker’s rating of female responsibility or decreasing the rating of male responsibility for child maltreat-
ment. The exception to this rule was when the female caregiver was “beaten up”—this act decreased her
responsibility for physically harming her child.
In addition to the domestic violence findings, the results regarding parental intoxication affirm the
seriousness with which child welfare workers view substance abuse—whether excessive use of alcohol
or other drugs—in their substantiation decisions (Craft et al., 1980; English et al., 2002), and in the risk
and recurrence of maltreatment (Coohey, 2003; English et al., 1999; Fuller & Wells, 2003; Hamilton
& Browne, 1999; Wolock & Magura, 1996; Wolock et al., 2001). Intoxication significantly increased
concerns about the safety of the child in the home and significantly increased male and female caregivers’
responsibility for physically harming the child and for not watching the child closely enough, though not
for exposing the child to domestic violence.
458 M.J. Landsman, C.C. Hartley / Child Abuse & Neglect 31 (2007) 445–461

Strengths and limitations

This particular method offers important and unique advantages for studying the attribution of respon-
sibility. First, the sample size is based on the case vignette as the unit of analysis, not the individual
respondent reading and rating vignettes. Thus this method enables us to study the attribution of respon-
sibility for child maltreatment in 435 cases, rather than 87 respondents’ assessments of responsibility
for child maltreatment. This permits a substantial sample size of vignettes while not requiring inordinate
numbers of respondents, thus maximizing statistical power.
A second advantage of this approach is that it directly addresses some of the key problems in case
vignette research. One problem is that case vignettes cannot replicate real world situations without being
excessively complicated and introducing various sources of bias. The factorial survey approach simplifies
the case vignettes to a series of statements that appear in various combinations. By keeping the wording
of the same factors constant, there is a reduced potential for bias in interpreting these statements. Another
problem is that in real world situations there is multicollinearity among variables of interest. That is,
certain variables tend to be correlated with each other, making it hard to separate out individual effects;
in research the goal is to identify the unique effect of each variable. The factorial survey method creates
a set of all unique objects formed by all possible combinations of one level from every dimension. In
other words, the method creates vignettes composed of as many different combinations of factors as are
possible, in attempt to then identify which specific attributes hold the strongest explanatory power over
the dependent variables. By using simple random sampling to select one level from each dimension,
ensuring that each level has an equal probability of being chosen, judgments that reveal the respondents’
preferences in decision-making, not confounded by problems of multicollinearity, can be evaluated. This
also means that the vignettes themselves need not necessarily look realistic, though they cannot contain
absurd combinations of variables.
There are several limitations to the internal and external validity of this study which warrant mention.
First, the sample size, though respectable, is not large relative to the number of parameters in each model.
Statisticians vary in their recommendations regarding the ratio of parameters to sample size. This study is
closer to the minimum ratio of 5 to 1 required by Tabachnick and Fidell (1989), while a general rule of 10
to 1 is more common (Green, 1991). There is considerable debate among statisticians on the usefulness of
these ratios relative to concerns about sample size and statistical power (Maxwell, 2000). Moreover, there
are other factors identified in the literature as relevant to the assessment of child maltreatment which were
not randomized in this study due to concerns about statistical power and sample size (e.g., mental health,
age, number of children in the family). For practical purposes, the study limited the number of randomized
factors to those most prominent in the literature, though future studies might choose to broaden this range.
Additionally, in most of the models, a relatively small number of factors had significant effects on the
dependent variable. Explained variation ranged from a low of 35% for concern for the child’s safety to
a high of 49% in male responsibility for physically harming the child. Thus, factors other than those
randomized in the study account for the majority of explained variation in the attribution of responsibility
for child maltreatment and concern for child safety. The findings of this study are intended to be suggestive,
not conclusive.
An additional limitation is the response rate of 49%, which is just under the standard 50% for survey
research. While lower than desired, this does not seem unduly low for the population of public child
welfare workers who are typically overburdened with paperwork and considering that this survey required
respondents to complete a task rather than to respond to an attitudinal questionnaire. Nonetheless the
 M.J. Landsman, C.C. Hartley / Child Abuse & Neglect 31 (2007) 445–461 459

responses analyzed in this study represent slightly less than half of those surveyed; how those non-
responding workers would have attributed responsibility for child maltreatment in these case vignettes
remains unknown.
This study was conducted in a mid-western state, with respondents drawn from both metropolitan and
non-metropolitan areas. Due to variation across states in how child maltreatment is defined and assessed,
and how domestic violence is handled within child protection, the findings may not be readily generalized
to other state systems.
A caveat to interpreting the results is that in those case vignettes involving domestic violence, both
the child maltreatment and domestic violence incident were occurring simultaneously. This was a delib-
erate strategy to keep the vignettes brief and to make it clear that the two types of family violence were
indeed co-occurring. However, in practice this is not always the case. When a family is being inves-
tigated for child maltreatment, the domestic violence may not be immediately apparent. The result is
that the effect of the domestic violence incident on the attribution of responsibility for child maltreat-
ment and concerns about child safety may be heightened due to its proximity to the child maltreatment
incident.

Implications for practice

This study was explicitly designed to be artificial, first by randomizing each factor to create case
scenarios; and second, by using the language of attributing responsibility for various actions rather than
the more legalistic language of substantiation of child maltreatment. The artificial nature of the study is
not a limitation, but a strength of the study in terms of its stated purpose. By systematically controlling
the ways that case circumstances come together in a child protection case, the method reduces the
uncontrolled biases that would otherwise emerge in a more realistic field environment. This formal-
empiricist approach within the broader paradigm of decision theory is a first step in understanding
how the presence of domestic violence influences workers’ judgments about responsibility for child
maltreatment.
Therefore, despite being somewhat removed from the level of real world practice, the findings from
this study do have some implications for practice. The fact that public child welfare workers’ assessments
appear to be influenced by the presence of domestic violence suggests that greater attention should
be paid to assessing the dynamics of domestic violence during child protective investigations and in
ongoing case management and permanency planning with families involved in the child welfare system.
Thus, child welfare workers need training on the dynamics of domestic violence and preparing safety
plans with family members that address a broad range of safety issues. Furthermore, child protection
and child welfare practice should directly address the issue of differential expectations of male and
female caregivers. Child welfare services most frequently focus interventions on the female caregiver.
However, great attention should be paid to male caregivers’ responsibility and involvement in service
planning.
To move closer to the practice level, the “how” and “why” of the effects of these factors must be
examined in actual field research. The current study has deconstructed the decision-making process to
identify factors at work in assessing a set of hypothetical cases. By examining whether similar factors
emerge in the process of decision-making in real case situations, a greater understanding of how to
improve practice with families experiencing co-occurring domestic violence and child maltreatment
might be achieved.
460 M.J. Landsman, C.C. Hartley / Child Abuse & Neglect 31 (2007) 445–461

Acknowledgments

The authors wish to acknowledge Carol Coohey and the late Susan Schechter for their feedback on
the development of the study vignettes and survey and the Obermann Center for Advanced Studies for
providing logistical support. We would also like to thank research assistants Molly Kellor and Jamie
Kuhne for their work in creating the case vignettes.

References

Beeman, S. K., Hagemeister, A. K., & Edleson, J. L. (1999). Child protection and battered women’s services: From conflict to
collaboration. Child Maltreatment, 4, 116–126.
Cohen, M. (1993). Three paradigms for viewing decision biases. In G. Klein, J. Orasanu, R. Calderwood, & C. Zsambok (Eds.),
Decision making in action: Models and methods (pp. 36–50). Norwood, NJ: Ablex.
Coohey, C. (in press). What criteria do child protective services investigators use to substantiate exposure to domestic violence?
Child Welfare.
Coohey, C. (2003). Making judgments about risk in substantiated cases of supervisory neglect. Child Abuse & Neglect, 27(7),
821–840.
Craft, J. L., Epley, S. W., & Clarkson, C. D. (1980). Factors influencing legal disposition in child abuse investigations. Journal
of Social Service Research, 4, 31–46.
Davidson, H. A. (1995). Child abuse and domestic violence: Legal connections and controversies. Family Law Quarterly, 29,
357–373.
Davis, J. (1995). Failure to protect and its impact on battered mothers. Courts and Communities Confronting Violence in the
Family, 1, 6–7.
Eckenrode, J., Powers, J., Doris, J., Munsch, J., & Bolder, N. (1988). Substantiation of child abuse and neglect reports. Journal
of Consulting and Clinical Psychology, 38(1), 9–15.
Edleson, J. L. (1999). The overlap between child maltreatment and woman battering. Violence Against Women, 5, 134–
154.
Edleson, J. L. (1998). Responsible mothers and invisible men: Child protection in the case of adult domestic violence. Journal
of Interpersonal Violence, 13, 294–298.
English, D. J., Marshall, D. B., Brummel, S. B., & Orme, M. (1999). Characteristics of repeated referrals to child protective
services in Washington State. Child Maltreatment, 4, 297–307.
English, D. J., Marshall, D. B., Coghlan, L., Brummel, S., & Orme, M. (2002). Causes and consequences of the substantiation
decision in Washington State child protective services. Children and Youth Services Review, 24(1), 817–851.
Fitch, D. K. (2001). A description of the child protective services investigative decision-making context with implications for
decision support system design. Unpublished doctoral dissertation, University of Texas, Arlington.
Fleck-Henderson, A. (2000). Domestic violence in the child protection system: Seeing double. Children and Youth Services
Review, 22, 333–354.
Fluke, J. D., Yuan, Y.-Y. T., Hedderson, J., & Curtis, P. A. (2003). Disproportionate representation of race and ethnicity in child
maltreatment: Investigation and victimization. Children and Youth Services Review, 25(5/6), 359–373.
Fuller, T. L., & Wells, S. J. (2003). Predicting maltreatment recurrence among CPS cases with alcohol and other drug involvement.
Children and Youth Services Review, 25(7), 553–569.
Fuller, T. L., Wells, S. J., & Cotton, E. E. (2001). Predictors of maltreatment recurrence at two milestones in the life of a case.
Children and Youth Services Review, 23, 49–78.
Green, S. B. (1991). How many subjects does it take to do a regression analysis? Multivariate Behavioral Research, 26(3),
499–510.
Hamilton, C. E., & Browne, K. D. (1999). Recurrent maltreatment during childhood: A survey of referrals to police protection
units in England. Child Maltreatment, 4, 275–286.
Harrison, L. A., & Esqueda, C. W. (2000). Effects of race and victim drinking on domestic violence attributions. Sex Roles, 42,
1043–1057.
 M.J. Landsman, C.C. Hartley / Child Abuse & Neglect 31 (2007) 445–461 461

Hartley, C. C. (2004). Severe domestic violence and child maltreatment: Considering child physical abuse, neglect, and failure
to protect. Children & Youth Services Review, 23, 373–392.
Home, A. M. (1994). Attributing responsibility and assessing gravity in wife abuse situations: A comparative study of police
and social workers. Journal of Social Service Research, 19, 67–84.
Jones, L. (1993). Decision making in child welfare: A critical review of the literature. Child and Adolescent Social Work Journal,
19(3), 241–262.
Jones, L. P., Gross, E., & Becker, I. (2002). The characteristics of domestic violence victims in a child protective service caseload.
Families in Society, 83, 405–415.
Leiter, J., Myers, K. A., & Zingraff, M. T. (1994). Substantiated and unsubstantiated cases of child maltreatment: Do they differ?
Social Work Research, 18(2), 67–78.
Magen, R. H. (1999). In the best interests of battered women: Reconceptualizing allegations of failure to protect. Child
Maltreatment, 4, 127–135.
Magen, R. H., Conroy, K., & Del Tufo, A. (2000). Domestic violence in child welfare preventative services: Results from an
intake screening questionnaire. Children and Youth Services Review, 22, 251–274.
Maxwell, S. E. (2000). Sample size and multiple regression analysis. Psychological Methods, 5(4), 434–458.
Mills, L. G. (2000). Woman abuse and child protection: A tumultuous marriage. Children and Youth Services Review, 22,
199–205.
Radhakriskha, A., Bou-Saada, I. E., Hunter, W. M., Catellier, D. J., & Kotch, J. B. (2001). Are father surrogates a risk factor for
child maltreatment? Child Maltreatment, 6(4), 281–289.
Rossi, P. H., & Anderson, A. B. (1982). The factorial survey approach: An introduction. In P. H. Rossi & S. L. Nock (Eds.),
Measuring social judgments: The factorial survey approach (pp. 15–67). Beverly Hills: Sage Publications.
Rubin, A., & Babbie, E. (2005). Research methods for social work. Belmont, CA: Brooks/Cole-Thomson Learning.
Saunders, D. G., & Anderson, D. (2000). Evaluation of a domestic violence training for child protection workers and supervisors:
Initial results. Children and Youth Services Review, 22, 373–395.
Shepard, M., & Raschick, M. (1999). How child welfare workers assess and intervene around issues of domestic violence. Child
Maltreatment, 4(2), 148–156.
Straus, M. A. (1979). Measuring intrafamily conflict and violence: The conflict tactics scales. Journal of Marriage and the
Family, 41, 75–88.
Tabachnick, B. G., & Fidell, L. S. (1989). Using multivariate statistics (2nd ed.). Cambridge, MA: Harper & Row.
Winefield, H. R., & Bradley, P. W. (1992). Substantiation of reported child abuse or neglect: Predictors and implications. Child
Abuse & Neglect, 16, 661–671.
Wolock, I., & Magura, S. (1996). Parental substance abuse as a predictor of child maltreatment re-reports. Child Abuse & Neglect,
20, 1183–1193.
Wolock, I., Sherman, P., Feldman, L. H., & Metzger, B. (2001). Child abuse and neglect referral patterns: A longitudinal study.
Children and Youth Services Review, 23, 21–47.
Zuravin, S. J., Orme, J. G., & Hegar, R. L. (1995). Disposition of child physical abuse reports: Review of the literature and test
of a predictive model. Children and Youth Services Review, 17(4), 547–566.