Anda di halaman 1dari 16

CHILD MALTREATMENT

10.1177/1077559504273684
Knutson et al. / NEGLECT /AND
MAYHARSH
2005 PARENTING A REPLICATION

Care Neglect, Supervisory Neglect, and Harsh


Parenting in the Development of Children’s
Aggression: A Replication and Extension

John F. Knutson
University of Iowa
David DeGarmo
Oregon Social Learning Center
Gina Koeppl
University of Iowa
John B. Reid
Oregon Social Learning Center

To understand the effects of neglectful parenting, poor super- the health and well-being of children, possible distal
vision, and punitive parenting in the development of chil- consequences of maltreatment, such as problems in
dren’s aggression, 218 children ages 4 to 8 years who were the child’s social adjustment, have become an impor-
disadvantaged and their mothers were recruited from two tant focus of research. Because there is unequivocal
states to develop a sample that was diverse with respect to de- evidence from three National Incidence Studies
gree of urbanization and ethnicity. Multimethod and (NIS; Office of Human Development Services
multisource indices of the predictive constructs (Social Disad- [OHDS], 1981, 1988; Sedlak & Broadhurst, 1996) and
vantage, Denial of Care Neglect, Supervisory Neglect, and the National Child Abuse and Neglect Data System
Punitive Discipline) and the criterion construct (Aggression) (NCANDS; U.S. Department of Health and Human
were used in a test of a theoretical model using structural Services, Administration on Children, Youth, and
equation modeling. The results established the role of care ne-
glect, supervisory neglect, and punitive parenting as media- Authors’ Note: Supported, in part, by funding by National Institute
tors of the role of social disadvantage in the development of of Mental Health (NIMH) and Administration on Children, Youth,
children’s aggression, the importance of distinguishing be- and Families (ACYF; Research Grant MH61731), John F. Knutson,
tween two subtypes of neglect, and the need to consider the role principal investigator; in part by funding by National Institute on
of discipline in concert with neglect when attempting to Child Health and Development (Research Grant HD 42115), Da-
vid S. DeGarmo, principal investigator; and funding by National In-
understand the parenting in the development of aggression. stitute on Drug Abuse and NIMH (Grants P20 DA 017592 and P30
46690), respectively, John B. Reid, principal investigator. The facili-
tation of the research by Paul Spencer (Oneida County Depart-
Keywords: children’s aggression; care neglect; supervisory ment of Social Services), Barry Bennett, Cheryl Whitney, Marc
neglect; punitive discipline Batey, Mark Schmidt, and Wayne McCracken (Iowa Department of
Human Services) is gratefully acknowledged. The assistance of
Angela Anderson, Beth Boyer, Minka Dawson, Becky Fetrow,
Although there is little question that child maltreat- Deanna Heckenberg, Martha Early, Aubra Hoffman, Esther
Hoffman, Kathy Jordan, Robert Latzman, Theresa Mayne, Mary
ment can occasion an immediate adverse effect on McCarren, Donna Palmer, Nicole Shay, and Nizete-ly Valles is
greatly appreciated. Correspondence regarding this article should
CHILD MALTREATMENT, Vol. 10, No. 2, May 2005 92-107 be sent to John F. Knutson, Department of Psychology, E-11 Sea-
DOI: 10.1177/1077559504273684 shore Hall, The University of Iowa, Iowa City, IA 52240; e-mail:
© 2005 Sage Publications john-knutson@uiowa.edu

92
Knutson et al. / NEGLECT AND HARSH PARENTING A REPLICATION 93

Families [USDHHS, ACYF], 2001, 2002) that neglect ment of antisocial behavior. For example, several
is the most common form of maltreatment experi- studies have previously implicated poor supervision
enced by children in the United States, it would and monitoring in the development of aggression,
seem that the outcomes of neglect would be well- antisocial behavior, as well as affiliations with deviant
established from a research literature. Comprehen- peers (e.g., Bank & Burraston, 2001; Dishion,
sive literature reviews (e.g., Knutson & Schartz, 1997) Patterson, & Greisler, 1994; Dishion, Spracklen,
have noted, however, that there is little evidence of a Andrews, & Patterson, 1996; Pettit, Bates, Dodge, &
specific link between neglect and identified child out- Meece, 1999); however, the supervisory neglect
comes. This state of affairs is attributable, in part, to included in several taxonomies of neglectful
the wide range of parenting practices that are sub- parenting (cf. Giovannoni, 1985; Hegar & Yungman,
sumed under the rubric of neglect (Flango, 1988; 1989; Sedlak & Broadhurst, 1996; Trocmé, 1996) has
Giovannoni, 1985; Hegar & Yungman, 1989; Knutson yet to be included or distinguished in studies of
& Schartz, 1997; National Research Council, 1993; neglect and child outcomes.
Trocmé, 1996; Zuravin, 1999), the fact that neglect is The Knutson et al. (2004) model specified how
often associated with other forms of child maltreat- care neglect could also play a functional role in the
ment (Sullivan & Knutson, 1998, 2000; USDHHS, development of children’s aggression via several
ACYF, 2001), and the fact that neglect is importantly developmental pathways. Care neglect that is mani-
related to socioeconomic disadvantage (e.g., OHDS, fested in such child attributes as poor hygiene and
1981, 1988; Sedlak & Broadhurst, 1996) that occa- inadequate clothing could result in poor peer rela-
sions considerable risk to children (e.g., Duncan & tions and rejection (Bolger, Patterson, & Kupersmidt,
Brooks-Gunn, 2000; Duncan, Brooks-Gunn, & 1998) and the compromised social participation
Klebanov, 1994; McLoyd, 1998). Thus, to understand (e.g., solitary play) that can be associated with the
how neglect might be related to the social adjustment development of emotional dysregulation (e.g.,
of children, research needs to either control for or in- Spinrad et al., 2004), aggression (e.g., Shields &
clude a consideration of the co-occurrence of other Cicchetti, 2001), and the hostile attributional biases
forms of maltreatment and social disadvantage as well underlying some childhood aggression (e.g., Dodge,
as provide well-specified operational definitions of Laird, Lochman, & Zelli, 2002).
neglect. After specifying unique roles of supervisory and
Knutson, DeGarmo, and Reid (2004) provided an care neglect, the model advanced by Knutson et al.
empirical test of a theoretical model that was (2004) also included harsh discipline as an important
designed to understand how deficient parenting, construct. Harsh punitive discipline is the key compo-
reflected in neglect and harsh discipline, could nent of coercive parenting behaviors that reinforce
account for the well-established link between social- child aggression. Harsh discipline practices include a
economic disadvantage and the development of child continuum of coercive punishment involving yelling,
and adolescent behavior problems (cf. Duncan et al., scolding, and spanking all the way to abusive physical
1994; Hanson, McLanahan, & Thomson, 1997; punishment. Prior studies have shown a strong associ-
McLanahan & Sandefur, 1994; McLoyd, 1998; ation between physical abuse and the development of
Patterson & Yoerger, 1999). Because of the research children’s aggression (e.g., Jaffee, Caspi, Moffitt, &
literature that linked neglect to delinquency (e.g., Taylor, 2004).
Steinberg, Lamborn, Darling, Mounts, & Dornbusch, In an initial evaluation, Knutson et al. (2004) used
1994), diagnosed behavior problems (Manly, structural modeling to specify supervisory neglect,
Cicchetti, & Barnett, 1994), school disciplinary diffi- care neglect (i.e., inadequate food, clothing, hygiene,
culties (e.g., Eckenrode, Laird, & Doris, 1993; etc.), and harsh parenting constructs to predict anti-
Kendall-Tackett & Eckenrode, 1996) and later crimi- social behavior and long-term change in antisocial
nal behavior (e.g., Henry, Caspi, Moffitt, & Silva, behavior of two age-based cohorts of children
1996; Viermö, 1996; Widom, 1998); Knutson et al. enrolled in a community-based universal prevention
(2004) reasoned that neglectful parenting could be project. For the younger cohort of children, whose
an important mediator between disadvantage and families were enrolled when the children were in first
adverse child outcomes in the form of antisocial and grade, the model established that care neglect and
aggressive behavior. punitive discipline contributed uniquely to children’s
A major advance in the Knutson et al. (2004) antisocial behavior; that is, care neglect was associated
model was the attempt to distinguish the role of with initial antisocial status as well as increases in anti-
supervisory neglect from other forms of neglect in social behaviors from first grade to fifth grade. Puni-
understanding deficient parenting in the develop- tive discipline was marginally associated with antiso-

CHILD MALTREATMENT / MAY 2005


94 Knutson et al. / NEGLECT AND HARSH PARENTING A REPLICATION

cial behavior at enrollment but significantly


predicted increases in antisocial behavior on follow- Punitive
Supervisory
up 5 years later. From the standpoint of the theoreti- Neglect
Discipline
cal model that was advanced, supervisory neglect was
not shown to be important in the development of anti-
social behavior in the younger cohort. For the older
(Mediated Influence) Child
cohort, whose families were enrolled when the chil- Social Aggression
Disadvantage
dren were in fifth grade a somewhat different, but the-
oretically consistent, pattern emerged. Although the
care neglect construct was associated with higher lev- Care
els of initial antisocial behaviors and marginally Neglect
associated with increases in antisocial behavior over
time in the older cohort, the most notable difference FIGURE 1: Hypothesized Relationships for Parent Deficiency
between the two cohorts was the finding that the Model of Supervisory and Care Neglect
supervisory neglect construct was related to care
levels of neglectful parenting nor high levels of
neglect and punitive discipline in predicting child
severely punitive discipline. Second, and perhaps
antisocial behavior. A second notable difference
more important, the study was imbedded in a univer-
between the cohorts was the significant link between
sal prevention program that was not designed to fully
care neglect and punitive discipline. Finally, and
assess abusive and neglecting parenting. Thus, the
importantly, for the older cohort, punitive discipline
supervisory neglect construct and the care neglect
at enrollment was important in predicting antisocial
construct were developed using proxy variables and
status at enrollment, but less important as a predictor
derivative approximations of the target indicators. In
of change in antisocial behavior 5 years later when
particular, although the supervisory neglect construct
compared to the test of the model with the younger
was developed using indicators that worked well for
cohort. When the model tests were considered in con-
an older cohort developing into preadolescence, the
cert, two important findings emerged. First, the
indicators might not have been suitable for taping the
importance of punitive discipline in predicting anti-
essential features of supervision in the younger years.
social behavior was apparent, a finding that is consis-
Thus, it seems probable that the Knutson et al. (2004)
tent with the work by Jaffee et al. (2004). Second,
model would fit data in which a more disadvantaged
there was strong evidence of the importance of care
sample was tested and better indices of care neglect,
neglect in the development of children’s aggression
supervisory neglect, and harsh discipline were avail-
and some evidence of the possible importance of
able. Thus, the current research is designed to repli-
supervisory neglect in the development of children’s
cate the theoretical model that specifies a causal role
aggression.
for care neglect, supervisory neglect, and harsh puni-
Although the Knutson et al. (2004) study provided
tive discipline in the development of children’s anti-
some empirical support for a theoretical model in
social behavior advanced by Knutson et al. (2004) by
which three components of deficient parenting (care
using a more neglecting, abusive, and disadvantaged
neglect, punitive discipline, and supervisory neglect)
sample of families and stronger multimethod and
contributed uniquely to the development of chil-
multisource indicators of the core constructs. Based
dren’s aggression and antisocial behavior, the fact
on the model advanced by Knutson et al. (2004) and
that supervision only played a role in the older cohort
their findings, a figure depicting the hypothesized
compromised reaching unequivocal conclusions
paths in a structural model are shown in Figure 1.
about the viability of the model. A number of limita-
tions of the Knutson et al. (2004) study, however, sug-
gested that a systematic replication would be useful. METHOD
First, although the sample used in that research was
Participants
drawn from a community with elevated risk for child
antisocial behavior (high juvenile arrest rates and Two hundred and eighteen mothers and their
community indices of delinquency), the recruited child who had been consecutively recruited at
sample represented a relatively broad range of socio- research sites in southeastern Iowa and north central
economic strata and a relatively low base rate of fami- Wisconsin during the first 28-month period of an
lies who had been identified by child protective ser- ongoing study of neglectful parenting and children’s
vices (CPS) agencies as being either physically abusive aggression participated in the current study.
or neglectful. Thus, the sample did not reflect high Although some single-parent fathers and fathers in

CHILD MALTREATMENT / MAY 2005


Knutson et al. / NEGLECT AND HARSH PARENTING A REPLICATION 95

two-parent households had been recruited, the num- this target population are notoriously difficult to con-
ber of fathers was too small to include in the analyses tact and enroll in research. Focus groups with
for this article. Each recruited mother-child dyad had members of the target population were conducted in
to meet a number of eligibility criteria. First, the fam- the community to identify factors that could compro-
ily had to have a child between ages 4 and 8 years resid- mise or facilitate recruitment. Frequent address
ing in the home at the time of recruitment. Second, changes, phone disconnection or number changes,
the family had to have been a recipient of social or use of cell phones, unlisted numbers, caller ID, as well
economic services at the time of recruitment. Services as an acknowledged reluctance to open mail that
could have been occasioned by the family being iden- seems official or is from an unknown source were
tified as physically abusive or neglecting, or because noted by focus group participants. As a result, the
of economic disadvantage. Families with children recruitment strategy that was adopted reflected some
known to have been sexually abused were not of the recommendations of those focus groups.
included in the pool of potential participants. In addi- Eligible families from the lists first received a letter
tion, because the focus of the research was on parent- from an administrator of the social service agency
child relational function, children who had been in who was not in a service delivery role. The letter
an out-of-home placement (i.e., foster care) or were informed potential participants of their eligibility to
recipients of intensive in-home services were not participate in research on parenting and children’s
identified as eligible for participation. social development for which they would be compen-
The recruitment area of Iowa included two coun- sated at a rate of $50 per session, plus out-of-pocket
ties that encompassed two metropolitan areas classi- (i.e., transportation and babysitting) expenses. Par-
fied as small urban and suburban, as well as rural agri- ents who did not respond to the initial letter received
cultural areas and several small towns. The a follow-up letter and, later, telephone contact by a
recruitment area of Wisconsin included two rural clerk employed by the project. In the letters, and in all
counties characterized by several small towns, rural of the direct contacts, potential participants were
development, as well as isolated rural living. In total, informed that participation was voluntary, and enroll-
the sample represented a broad range of degree of ment or nonenrollment in the project would have no
urbanization; however, living arrangements influence on their receipt of any services from rele-
prototypical of inner-city environments (i.e., high-rise vant state or county agencies. When the eligible par-
apartments) were not represented. Many of the ent indicated an interest in participation, an initial
recruited families, however, had recently relocated appointment was scheduled in his or her home; dur-
from large urban areas of the Midwest (e.g., Chicago, ing that initial session the research protocol was
Illinois; Detroit, Michigan; Gary-Hammond, Indiana; explained to the participant, and informed consent
Milwaukee, Wisconsin), and many of the children had was obtained. If the parent chose to participate in the
considerable exposure to inner-city life. The children research, an initial structured interview was con-
were slightly more racially diverse than the mothers; ducted in the home immediately following the
64% of the mothers identified themselves as White, informed consent procedure. For families in which
23% African American, 5% Latino/a, 2% Native more than one child was eligible to participate, a
American, 1% Asian, and 4% as multiracial. Children single child was randomly selected to be the research
were identified by their mothers as 56% White, 23% participant.
African American, 5% Latino/a, 3% Native Ameri- Of the families on the eligibility lists, approxi-
can, 1% Asian, and 12% as multiracial.1 mately 50% were known to have been actually con-
At both of the recruitment sites, sentinel agencies tacted by mail or phone. Of those, approximately 55%
were used to identify potential research participants. scheduled an in-person, in-home appointment to
Because of the different structure of the provision of learn about the project. Of those who scheduled an
social and economic services in the two states, slightly in-home interview and participated in the informed
different procedures were followed in Iowa and Wis- consent process, all but one agreed to participate in
consin. Basically, the appropriate state or county the project. The mother who declined to enroll was
agency produced a list of potential research partici- the mother of twins, and she would not agree to enroll
pants at approximately quarterly intervals. Those if only one of the children could participate in the
agencies identified families who had recently been study. A few potential participants repeatedly sched-
identified as having had a substantiated physical uled in-home sessions but were not home when the
abuse or neglect allegation, or who were receiving research assistant arrived at the scheduled time. After
state or county support (i.e., Temporary Assistance one third missed scheduled appointment, no further
for Needy Families (TANF); Title XIX). Families in efforts to recruit were made. Participant recruitment

CHILD MALTREATMENT / MAY 2005


96 Knutson et al. / NEGLECT AND HARSH PARENTING A REPLICATION

and all of the procedures were conducted under the members, amount of space available to the number of
aegis of The University of Iowa IRB-02. residents of the household, and any health or safety
hazards threatening children of the household could
Procedures be identified by the interviewer in a direct examina-
In developing indicators of the core constructs of tion of the household. Following the interview, the re-
care neglect, supervisory neglect, and harsh punitive search assistants completed an objective checklist
discipline, a number of factors were considered. Most while conducting a visual survey of the physical envi-
important, it was deemed appropriate that all of the ronment in the vicinity of the participant’s home to
indices of parenting be based on directly obtained identify environmental hazards and the general con-
information rather than the use of administrative dition of the immediate community in which the
data or agency records. In addition, for all of the con- child resides (e.g., broken glass, drug paraphernalia,
structs, a multimethod and multisource approach was houses in disrepair, boarded windows, trash). For par-
adopted. With the exception of the initial structured ticipants residing in a municipal area, an area within a
interview and in-home assessment, all of the measures 1-block radius was examined. For rural participants,
were obtained in three laboratory sessions requiring the area was generally comparable; however, the ar-
approximately 90 minutes each. Because of the longi- rangement of rural roads and sparse population
tudinal nature of the project in which the current sometimes required a greater distance to evaluate the
study was embedded, additional measures that are status of nearby houses to complete the rating of the
not the focus of this research were obtained but not community condition and accessible hazards.
described in full. Only procedures and derived mea- Laboratory Assessments
sures relevant to the scope of the current analyses are
described below. Within a median of 26 days of the home visit, the
first of a series of laboratory sessions was completed.
In-home interview and environmental assessment. Subsequent sessions were arranged as soon as possi-
When a parent agreed to participate in the research, a ble; however, scheduling difficulties and missed
structured interview designed to obtain background appointments occasionally resulted in protocol com-
and demographic information, as well as information pletion more than 60 days after enrollment. During
pertinent to the parenting constructs, was immedi- the first session, the parent was administered the
ately conducted by a research assistant. In most in- Reading and Spelling subtests of the Wide Range
stances, one research assistant conducted the Achievement Test (WRAT; Jastak Associates, 1993) to
interview and recorded participant responses while a establish literacy. For mothers who were not suffi-
second research assistant recorded the participant’s ciently literate to complete the paper-and-pencil self-
responses on a parallel form. This structured inter- report instruments of the protocol, those instruments
view was based, in part, on a modification of the were administered in an interview format by a
Home Observation for Measurement of the Environ- research assistant.
ment (Caldwell & Bradley, 1978) recently developed
for use in the Project on Human Development in Chi- Parent-child interaction task. Microsocial indices of
cago Neighborhoods (PHDCN; Leventhal, Selner- parent-child interaction were coded from videotaped
O’Hagan, Brooks-Gunn, Bingenheimer, & Earls, interactions between the mother and child while they
2004) and, in part, on the information framework participated in a structured laboratory task during the
that emerged from the recommendations of the Re- first session. This 45-minute parent-child laboratory
search Subcommittee of the Interagency Task Force task was conducted in a room outfitted to approxi-
on Child Abuse and Neglect (see Sternberg et al., mate a living room setting, with comfortable stuffed
2004). The interview also included questions regard- furniture, an end table, a coffee table, and appropri-
ing supervision related to injury prevention, which ate appointments. So coders are uninformed with re-
were derived from research by Peterson, Ewigman, spect to the source of the families; the laboratory
and Kivlahan (1993) and a home safety inventory rooms at the Wisconsin laboratory and the Iowa labo-
(Tymchuk, Lang, Dolyniuk, Berney-Ficklin, & Spitz, ratory are identical with respect to size, furnishings,
1999). Because this interview was conducted in the carpeting, and other appointments. The task begins
home of the child, it was also possible to obtain direct with a communication test in which the child first
information regarding evidence of neglect manifest plays an unfamiliar (i.e., commercially unpopular)
in the condition of the home, following the strategy age-appropriate board game with an experimenter.
developed by PHDCN; that is, sleeping arrangements, At that time, the parent completes a questionnaire on
cleanliness, plumbing, personal hygiene of family child behavior issues in an adjacent room. After 5

CHILD MALTREATMENT / MAY 2005


Knutson et al. / NEGLECT AND HARSH PARENTING A REPLICATION 97

minutes playing the game, the parent is brought to neously shifting steadily away from an established
the room and left alone with the child with the in- standard (DeMaster, Reid, & Twentyman, 1976). The
struction to learn from the child what had just tran- effective agreement for occurrence was .89 for Con-
spired. The parent’s task is to learn as much as tent codes with a kappa coefficient of .80. The
possible about what the child did with the experi- effective agreement score for Affect codes was .93,
menter and the details of the game. After 5 minutes, with a kappa of .76. The Activity code agreement was
the child is removed from the room, and the experi- .93.
menter interviews the parent to determine how suc-
Confidentiality and Child Safety
cessfully she was able to obtain information from her
child. The second component of the laboratory task is The entire project was conducted under a Certifi-
a social problem-solving task involving role-playing cate of Confidentiality issued by the office of the
and pretend activities by the parent and child based National Institutes of Mental Health. The Certificate
on instructions given to the parent (e.g., preparing was obtained with the provision in the informed con-
for the visit of another family with a child who is dis- sent documents that the investigators would report to
liked by the participant child). The third component relevant CPS agencies any circumstances that were
involves the parent leading a discussion regarding the deemed to occasion risk to the child. Such circum-
most important of the child issues identified in the stances could be identified from child interviews, par-
questionnaire completed while the child played the ent interviews, or direct observations. It should be
game with the experimenter. Fourth, the parent and noted that the submission of such reports to CPS
child engage in free play for 10 minutes. Finally, the agencies did not result in any loss of participants in
parent is to induce the child to straighten the room the protocol described in this article.
and put the toys away. All of the parent-child interac-
tions are recorded using a remote control dome
OPERATIONAL INDICATORS OF
camera.
THEORETICAL CONSTRUCTS
Coding of parent-child interactions. The videotapes of
Social Status
the structured parent-child interactions were coded
by a professional team of four behavior coders at Ore- Four indicators constituted the social status con-
gon Social Learning Center (OSLC) using the Inter- struct. Two scales were from the Home Environment
personal Process Code (IPC; Rusby, Estes, & Dishion, Questionnaire (HEQ; Laing & Sines 1982; Sines,
1991). The IPC provides a real-time assessment of vir- Clarke, & Lauer, 1984). The HEQ is a true-false ques-
tually all verbal and nonverbal interactions among tionnaire that uses objectively phrased items to obtain
family members by recording the Activity (the global information about the child’s environment from the
context or setting in which the interactions occur), child’s parent. Developed using the rational-statistical
Content (a description of each verbal, nonverbal, and approach (Loevinger, 1956), the HEQ has eight
physical behavior), and Affect (the emotional tone ac- empirically derived scales measuring specific dimen-
companying each content code) of the mother-child sions of a child’s psychosocial environment, with each
interactions. The IPC was used to index the frequency of the scales assessing the presence of environmental
of maternal aversive and physically negative behaviors attributes and events that are theoretically important
(e.g., contempt, anger, hitting with hand, hitting with for the expression of deviant and nondeviant child
an object, pinching, ear flicking, kicking, grabbing, behaviors (Murray & Sines, 1996). For the purpose of
restraining, spitting, shoving) as well as those of the the current research, two scales were selected as indi-
child. To enhance reliability of the coding, the family cators of the Social Status construct. The Achieve-
interaction tapes were aggregated and coded in a ran- ment scale measures family conditions that model or
dom order. Of the sessions, 15% were randomly se- provide support for achievement on the part of the
lected for reliability checks throughout the time of child, and the Socioeconomic Status scale comprises
coding. Contrasts are made between coders who had items that refer to activities and attitudes related to
been most recently established as reliable and those academic and intellectual pursuits, as well as partici-
for whom the longest time had elapsed since the last pation in community affairs. Two additional indica-
reliability assessment. This procedure permits the de- tors were mothers’ education and occupation. Educa-
tection of the potential problem of so-called observer tion ranged from 1 (never reached high school) to 8
drift, a phenomenon in which coders have acceptable (graduate or professional degree). Occupation ranged
levels of agreement across time but are simulta- from 1 (unskilled laborer) to 7 (professional). Students,

CHILD MALTREATMENT / MAY 2005


98 Knutson et al. / NEGLECT AND HARSH PARENTING A REPLICATION

homemakers, and unemployed working status were Harsh Punitive Discipline


recoded to 1 to reflect economic contribution to the
social status indicator of occupational status. The punitive discipline construct was measured
with five indicators from multiple sources of data
Supervision and Tracking involving direct observation, analog procedures, and
As noted by Dishion and McMahon (1998), aware- self-report. The self-report measures included a scale
ness of child activities is a critical component of of inconsistent discipline consisting of 11 items used
parental monitoring. Thus, for the young children of in Knutson et al. (2004) indicating a pattern of incon-
the current study, it was hypothesized that parental sistent or erratic discipline. Sample items included
awareness measured by congruence between parent “get angry when punishing child, how often parent
report and child report would serve as the optimal does not follow through on punishment, threaten
index of supervision. Two congruence or concor- punishment to get child to do something, and so on.”
dance scores were derived to measure effective super- Cronbach’s alpha was .98. The second reported scale
vision and tracking skills of the parent. As one mea- was an index of abusive discipline resulting in injuries
sure, the parent and the child independently to the child as a result of discipline. Ten items from
completed the Children’s Experience and Excite- the home interview were scored as 0 (did not occur)
ment Scale (CEES: Selner, 1992; Selner & Knutson, or 1 (has occurred) and then summed. Sample items
1990) during the first laboratory session. The CEES include “red marks that lasted more than 24 hours
consists of 44 slides depicting children engaging in a from being spanked, bruises after being disciplined,
range of activities. To minimize sex-role responding, broken bones from being disciplined, child required
there is one form for boys, with male actors, and stitches after discipline, child spanked with object
another form for girls, with female actors. In an inter- other than a hand (e.g., belt, paddle, tree branch, hair
view format, child participants are asked whether they brush, other), unusual punishment (e.g., child tied
have ever engaged in the depicted activity. If they had up, lock in closet, physically restrained, push ups,
not engaged in the activity, they were asked whether etc.). The observed score ranged from 0 to 6 (M =
they have had the opportunity to do so. Without hav- 1.56, SD = 1.18).
ing any knowledge of their child’s response to the The third indicator was a factor score of physical
CEES slides, parents complete the CEES in a self- responses and escalation in angry discipline in
report format by indicating whether their child has response to Analog Parenting Task (APT) first
ever engaged in the pictured activity. If they indicate described by Zaidi, Knutson, and Mehm (1989). The
their child has not engaged in the activity, they indi- APT, consists of 28 slides, each depicting a child
cate whether the child ever had the opportunity to do engaging in a developmentally appropriate or devel-
so. Concordance between child and parent reports of opmentally inappropriate activity that could be irri-
experiences and opportunity provide an index of tating or concerning to a child’s caretaker. Seven
supervision by adding mother-child affirmative scenes depict a child engaged in destructive acts (e.g.,
matched pairs and mother-child negative pairs across stepping on a calculator, tearing pages from a book),
all slides. The index ranged from 12 to 39 agreements seven scenes depict dangerous activities (e.g., loading
(M = 30.94, SD = 4.75). a revolver, hanging out the window of a moving car,
The second index of supervision was the corre- sitting on the edge of a roof), and seven depict rule-
spondence between child report and parent report violating behaviors (e.g., theft, drinking an alcoholic
on the Children’s Reinforcement Survey Schedule beverage, smoking). The remaining scenes include
(RSS; Clement & Richard, 1976). The RSS, adminis- age-appropriate acts (e.g., spilling a jar of salsa, messy
tered by interview, asks the children to identify the play with toys). In response to each scene, the mother
people with whom they spend the most time, their is asked to imagine that she is charged with the
favorite foods, the toys they use most often, and the responsibility of caring for the depicted child, to indi-
places they spend the most time. They are also asked cate her emotional reaction to the depicted child
to identify toys they do not have but would like, places (e.g., anger, worry, annoyance, amusement), and to
they would like to spend more time, and people with classify the depicted behavior (e.g., sloppy, destruc-
whom they would like to spend more time. The par- tive, dangerous, fine). After rating and classifying the
ent completes a paper-and-pencil RSS form. Based on depicted behavior the participant is asked to select, in
the work of Hall (1986), the effective agreement for a closed-set format, the disciplinary response she
occurrence statistic between the child report and the would use if she were attempting to alter the child’s
parent report is used an index of supervision (M = .35, behavior. Disciplinary choices included such acts as
SD = .09). ignoring, verbal reprimands, restricting of privileges,

CHILD MALTREATMENT / MAY 2005


Knutson et al. / NEGLECT AND HARSH PARENTING A REPLICATION 99

spanking, striking other than spanking, and striking etc.), and the log-transformed count of total negative
with objects. Although the more severe acts were physicals (e.g., hit, pinch, slap, etc.) of the mother
potentially injurious, the possible injurious conse- directed to the child.
quences of the acts were not specified in the response
Care Neglect and Environmental Neglect Index
choices. After selecting a disciplinary strategy, the par-
ticipant was asked to indicate how many times she In the Knutson et al. (2004) study, the assessment
would permit the child to engage in the depicted of parental care neglect was relatively limited and
behavior before changing her disciplinary response. comprised a summative index using six items from a
If the participant indicated that she would change her home visitor and from teacher observations (e.g.,
disciplinary tactic, she was asked to indicate what that child’s hygiene). One focus of the current study was to
next disciplinary alternative would be. develop a multisource measure of care neglect. Thus,
There are three primary dependent measures the assessment battery was expanded to include par-
obtained from the APT. The first was the frequency ent report and objective observer ratings of care
with which the participant’s initial disciplinary choice neglect (e.g., child does not have a toothbrush), and
involved the use of physical discipline. The second household environmental conditions that would
dependent measure was the use of escalated disci- occasion social (household is overly crowded, inade-
pline as described by Knutson and Bower (1994). quate illumination) and physical risks to a child
Escalated discipline is a circumstance where the par- (unsafe stairs, inadequate plumbing, animal feces
ticipant shifts from a nonphysical form of discipline to present, accessible pharmaceuticals). Items were all
physical discipline if the depicted child were to persist scored in a direction to indicate neglect and then
in the displayed behavior. Escalated discipline can summed. Items were chosen based on their inclusion
also occur when the participant shifts from minor in the research literature or recommendations from
physical discipline (e.g., spanking) to potentially inju- the Interagency Task Force on Defining Child Mal-
rious discipline (e.g., striking with an object) within a treatment (see Sternberg et al., 2004) and were
scene. The computed analog discipline indicator was obtained from parent interviews and also from
a factor score of total physical responses ranging from observer ratings of the home environment. Proximal
0 to 15 (M = .48, SD = 1.46) and total escalated circumstances outside the home that could occasion
responses ranging from 0 to 13 (M = 1.42, SD = 2.70), risk to the child that were observed during the home
because of the skewed distributions before combin- visit (e.g., broken glass, drug paraphernalia) were also
ing, each score was log transformed (plus a constant included in the index. In selecting this summative risk
of 0.5 for zero values). index, we considered item relevance or face validity,
Based on the work of Greenwald, Bank, Reid, and avoided operational confounding with other con-
Knutson (1997), a third measure from the APT con- structs, and considered the importance of compre-
tributed to the anger indicator in the punitive disci- hensiveness as suggested by Turner and Wheaton
pline construct. This APT anger score was the number (1995). The obtained total neglect index scores
of scenes that evoke an anger response from the par- ranged from 1 to 20 (M = 9.56, SD = 3.67).
ent. Using the framework of Averill’s (1982) work on
Child Aggression
normative and nonnormative anger experiences,
mothers completing the APT were asked to indicate Five indicators comprised the child aggression
their emotional reaction to the depicted child scene, construct. The first was the Child Behavior Checklist
with anger being among the possible responses. The long form (CBCL; Achenbach, 1992) completed by
number of scenes evoking an anger response ranged the mothers during the first or second laboratory ses-
from 1.0 to 3.83 (M = 1.98, SD = .58, α = .91). The sec- sion. The 1992 CBCL T scores were converted to 2001
ond anger indicator was the State-Trait Anger Scale norms using the Achenbach System of Educational
(STAS: Spielberger, Jacobs, Russell, & Crane, 1983) Based Assessment (ASEBA; Achenbach & Rescorla,
that was administered during the second or third lab- 2000, 2001). Similarly, teachers directly working with
oratory session. The STAS score ranged from 19 to 65 each child completed the Teacher Report Form
(M = 29, SD = 6.15). The final anger composite indica- (TRF; Achenbach, 1991) behavior rating scales dur-
tor was a factor score of the APT pictorial anger ing the spring semester following participant enroll-
response and the STAS. ment, at time when they should be most familiar with
The final indicator was from the IPC observational the child’s behavior. The 1991 TRF was also converted
coding system described above. Two scores consti- to 2001 ASEBA norms.
tuted the indicator, the log-transformed frequency As a third indicator, the child vignette was a mea-
count of total aversive behaviors (contempt, anger, sure of the child’s aggressive responses to four social

CHILD MALTREATMENT / MAY 2005


100 Knutson et al. / NEGLECT AND HARSH PARENTING A REPLICATION

CBC-L T e5

e1 Mother

4
Education .44 Child
** Scenarios

.5
D1 e6
0
Social Status .80* .4
e2
HEQSS
** .39* Interviewer
Mothers' -.22* Child Rating
e7

*** Social .31*


Achievement .73 Status Aggression
log IPC
e3 .4
HEQACH
*** 8* Child to e8
.36 r 2 .13 * Parent
8*

.08
e4
Mother
Occupation -.2
TRF T e9

Age Sex
of Child of Child

FIGURE 2: Child Aggression Factor Predicted by Mothers’ Social Status


NOTE: HEQ SS = Home Environment Questionnaire - Social Status scale; HEQ ACH = Home Environment Questionnaire - Achievement
scale; CBCL-T = Child Behavior Checklist Aggression Scale T-Score; IPC = Interpersonal Process Code; TRF T = Teacher Report Form Ag-
gression Scale T-Score.
χ2 = 43.30(46), p = .59, χ2/df = .94, CFI = 1.00.
*p < .05. **p < .01. ***p < .001.

situational vignettes. For example, in one vignette a sured in numerical years and sex of the child coded 1
child is hit in the head with a ball during play. The for female and 2 for male.
index score summed the child’s endorsement of We estimated the SEM path models using Full
either behavioral retaliation or hostile intent of the Information Maximum Likelihood (FIML). FIML uti-
child in the scenario, or both, thus the score could lizes all portions of data in a covariance matrix to esti-
range from 0 to 8 (M = 1.68, SD = 1.76). As a fourth mate parameters in the model. SEM parameters are
indicator, the interviewer rating of aggression was a typically estimated with list-wise deletion or mean-
summative index of dichotomous items rated by the substitution matrices. However, under assumptions
interviewer in the home and at each laboratory visit. of random missing data, FIML produces optimally
Seven items constituted the checklist of items efficient estimates of standard errors (Arbuckle,
rescored to indicate aggression: child struck parent, 1996; Wothke, 2000). Therefore, in the first step, we
positive to interviewer, cooperative, angry-irritable, conducted a missing value analysis of the variables
screamed or yelled, noncompliant, friendly to par- modeled in the SEM covariance matrix, including
ents. The final aggression indicator was from the IPC controls. Little’s test of missingness uses the EM algo-
observational coding system. The indicator was the rithm to compare estimated mean and variance val-
factor score of log-transformed negative physical ues of partial data cases with complete data cases. The
behaviors directed to the parent and the log- analysis indicated that the data were missing com-
transformed measure of aversive behaviors directed pletely at random (MCAR), χ2 = 237.94(261), p = .84,
to the parent. and that FIML was the optimal estimation procedure
for SEM.
Analytic Strategy
We employed structural equation path modeling RESULTS
(SEM) to test a replication of the hypothesized associ-
ations in the theoretical model. SEM is a latent vari- Results of the first model specifying the association
able regression technique that simultaneously com- of mothers’ social status affecting levels of child
bines factor analyses with path analyses under the aggression is presented in Figure 2. Standardized
assumptions of multivariate normality. SEM is partic- coefficients are presented for the regression paths of
ularly suited for multimethod multisource covariance the latent variable factor loadings and regression
matrices used in the current study because of its abil- paths for prediction between the theoretical latent
ity to partial measurement error of constructs and, constructs. The model obtained excellent fit to the
more specifically, to specify error between same- data with no significant difference between the speci-
source indicators, thus controlling for certain types of fied theoretical model and the observed covariance
mono-method, mono-agent biases. Two control vari- matrix, χ2 = 43.30(46), p = .59, χ2/df = .94, CFI = 1.00.
ables were used in the analyses. Age of the child mea- For estimation, the first indicator of social status (i.e.,

CHILD MALTREATMENT / MAY 2005


Knutson et al. / NEGLECT AND HARSH PARENTING A REPLICATION 101

e5 e6

log Abusive e8
CEES CRSS Index
Congruence Congruence

***
Analog

.2 .46
Physicals -

.38
.71
e9
Escalation

*
D3

9*
D1
r 2 .16 r 2.29
Supervision- * Anger
Tracking
-.53* .63** STAS - APT
e10
Punitive
Discipline .36
Mother *

.46
e1
Education .4 Inconsistent

.03
4 e11

**
Discipline

.2
5
-.0
Social Status .80

.33
*
*** †
-.21
e2

**
HEQSS log IPC
Social

-.29*
Mother to e12
.73*** Status -.06

.46*
Achieve- Child
e3
***
HEQACH

.01
.36
-.1

Mother

.22
CBCL T
2

e4 e13
Occupation D4 0

*
.7
** Child
.33 Scenarios e14
Care & .24* Child
Environmental Aggression .32*
2
r .12 Neglect .2 Interviewer e15
1* Rating
0
1.0

r 2.43
D2

.48
log IPC
Neglect Child to e16

**
e7 Index Parent

TRF T e17

FIGURE 3: Child Aggression Predicted by Mediating Parenting Constructs Controlling for Age and Sex of the Child
NOTE: CEES = Children’s Experience and Excitement Scale; CRSS = Concordance Scores from the Reinforcement Survey Schedule; HEQ
SS = Home Environment Questionnaire - Social Status scale; HEQ ACH = Home Environment Questionnaire - Achievement scale; STAS -
APT = Spielberger State-Trait Anger Scale Analog Parenting Task; IPC = Interpersonal Process Code; CBCL-T = Child Behavior Checklist Ag-
gression Scale T-Score; TRF T = Teacher Report Form Aggression Scale T-Score.
χ2 = 166.64(138), p = .05, χ2/df = 1.21, CFI = .90.
*p < .05. **p < .01. ***p < .001.

education) and child aggression (i.e., CBCL-T) were ables, older children were estimated to be lower on
fixed at 1.0 in the unstandardized parameters to scale the multimethod construct of aggression compared
the rest of the factor loadings. The remaining freely to younger children (β = –.28, p < .05). Overall, 13% of
estimated factor loadings for each construct were sig- the variance in child aggression was explained by
nificant and ranged from 0.31 to 0.80; meaning each social status, and the findings in Figure 2 were consis-
indicator significantly contributed to the common tent with prior developmental research on harsh con-
variance underlying the theoretical construct. textual factors of social disadvantage and protective
Although this sample is rather homogeneous in environmental factors of socioeconomic status and
terms of being from lower socioeconomic strata, the consistent with the theoretical model that was used to
indicators operationalizing the social status construct structure the current analyses (Knutson et al., 2004).
obtained factor variance that was significantly differ-
We next modeled the behavioral parenting mecha-
ent from zero (VAR = .44, SE = .07, p < .001); meaning
nism hypothesized to link social status to child aggres-
that although a relatively homogeneous sample was
sion. These factors included multimethod indicators
recruited by design, there was a range in the mothers’
of care neglect, supervision tracking, and harsh puni-
achieved and ascribed social status variables. Like-
wise, the factor measuring child aggression using five tive discipline. To evaluate generalizability of the the-
independent sources of data (i.e., observer, teacher, oretical model, paths among the constructs were
parent, child, and interviewer), obtained good con- specified in accordance to the associations modeled
vergence and significant factor variance (VAR = 2.42, in Knutson et al. (2004) using the longitudinal pre-
SE = .79, p < .01). Regarding the effect of social status ventive intervention sample from Oregon. Results for
on child aggression, the findings supported the theo- the effects of parenting mechanisms in the current
retical model; higher levels of maternal social status sample are displayed in Figure 3 using standardized
was associated with predicted lower levels of child path coefficients. For visual clarity, the paths for the
aggression (β = –.22, p < .05). Among the control vari- control variables were suppressed in the final figure.

CHILD MALTREATMENT / MAY 2005


102 Knutson et al. / NEGLECT AND HARSH PARENTING A REPLICATION

The fully specified path model obtained adequate a b l e s that ar e the o r e ti c a l l y c o nsi s te nt b u t
fit to the data, χ2 = 166.64(138), p = .05, χ2/df = 1.21, methodologically different, the theoretical model is
CFI = .90, the χ2 minimization coefficient comparing submitted to a more demanding test (see Meehl,
the theoretical model and the observed covariance 1978). When the obtained results are similar, we gain
was not significantly different, the comparative fit was greater confidence in our ability to generalize about
.90 or higher and the χ2 ratio χ2/df was below the con- the underlying theoretical mechanisms (Conger,
servative value of 2.00 (Byrne, 1989). Each of the Patterson, & Ge, 1995; DeGarmo & Forgatch, 2004;
respective factor loadings obtained significant contri- Dishion, Patterson, & Kavanagh, 1992), particularly
butions to the communality of factor variance. when results replicate across samples with appreciably
The theoretical model was supported by the data. different characteristics. With those considerations in
Starting from left to right with social status, the direct mind, the most significant finding of the current
path from social status to aggression was rendered study was the systematic replication of the earlier test
nonsignificant on entering the intervening mecha- of the theoretical model advanced by Knutson et al.
nisms. Further, social status was marginally associated (2004). In the earlier test of the model, care neglect,
with lower levels of punitive discipline (β = –.21, p < supervisory neglect, and harsh discipline were shown
.10), meaning that punitive discipline, in combina- to play distinguishable roles in determining the
tion with supervision and neglect, mediated or impact of social and economic disadvantage on the
accounted for the effect that social status had on child development of children’s antisocial behavior.
aggression. Although the current findings replicated the pattern
Among the parenting constructs, supervision of the fifth-grade cohort more closely than the first-
tracking measured by the congruence indicators pre- grade cohort in that supervision and tracking pre-
dicted lower levels of neglect (β = –.29, p < = .05) and dicted neglect, and neglect was directly associated
lower levels of punitive discipline (β= –.53, p < .05). with child aggression in the current study, it is impor-
Harsh and angry punitive discipline, in turn, pre- tant to note that Knutson et al. (2004) had hypothe-
dicted higher levels of child aggression (β = .46, p < sized a significant role for supervisory neglect in both
.05). In addition, care and environmental neglect sig- samples in the earlier research. In view of the fact that
nificantly predicted child aggression (β = .24, p < .05). the current sample is more suitable for identifying the
The addition of the intervening parenting mecha- circumstances of care neglect, supervisory neglect,
nisms accounted for a 30% increase in explained vari- and harsh punitive parenting than the high risk, but
ance of child aggression (r 2 = .43). Among the control appreciably more normative, sample of the Knutson
variables (paths not shown), boys were associated with et al. (2004) study, the parallels between the two
lower supervision (β = –.34, p < .01), and older chil- studies argue even more for the viability of the
dren were associated with greater supervision (β = .20, theoretical model.
p < .05). All other paths were not significantly differ- The difference between the earlier sample and the
ent from zero, including the direct effect of age on current sample is, perhaps, most telling in the differ-
aggression shown in Figure 2 (p = .11 in Figure 3). ence in the parental aversives coded from the parent-
Three correlation paths were estimated among resid- child interactions in the two samples. The mean num-
ual error terms in the final model to control for ber of aversive behaviors per minute of interaction
method or agent bias contributing to shared variance (i.e., rate per minute) was significantly higher for the
in the factor regression paths; these included the current sample compared to the sample from Oregon
covariance between the IPC observational scores used in the Knutson et al. (2004) study (M = 1.00 and
(re12,e16 = .22, p < .05) and the covariance among the .47, p < .001, respectively, for aversives and M = .14 and
parent reported scales rated on Likert-type scales .09, p = .06, respectively, for negative physicals). Thus,
(re10,e13 = .46 and re11,e13 = .33, p < .01). Thus, any poten- using a much more at-risk and socially disadvantaged
tial coder bias or response-set bias by the parents was sample the theoretical model advanced by Knutson
controlled in the model test. Substantively identical et al. (2004) was closely replicated in distinguishing
results are obtained constraining all residual dimensions of maltreatment.
covariances or freeing those shown. Another factor that underscores the viability of the
theoretical model is the fact that the indicators of the
constructs were obtained with theoretically consistent
DISCUSSION
but operationally different approaches in the two
When theoretical models are replicated with inde- studies. In the current study, the indices of deficient
pendent samples that are significantly different on parenting are based on much more direct
important dimensions, and when the test uses vari- multimethod indicators than was possible in the

CHILD MALTREATMENT / MAY 2005


Knutson et al. / NEGLECT AND HARSH PARENTING A REPLICATION 103

Knutson et al. (2004) study. In many ways, it is the rep- cial and aggressive behavior in adolescents (e.g.,
lication of the model using these more specific and Patterson, 1982; Patterson, Reid, & Dishion, 1992);
direct indicators that provides the strongest support and, at present, inadequate monitoring could be seen
for the theoretical model. Thus with significantly dif- as a developmental extension of supervisory neglect.
ferent samples and different indicators of the core However, deficient monitoring generally has not
constructs, Knutson et al. (2004) and the current been extended to children as young as those enrolled
study indicate that the three distinguishable attrib- in the current study because of the developmental
utes of deficient parenting play a role in the develop- appropriateness of self-reported measures of chil-
ment of children’s aggression and antisocial behav- dren’s activities outside of the home (e.g., time spent
ior. Even with the truncated range that comes with a at friend’s home, time arriving home, reports of
sample that is recruited entirely from a disadvantaged unobserved activities with peers, etc.). In most
population, the current study replicated the well- respects, the measure of supervision and tracking in
established link between socioeconomic disadvan- the current study is an index of awareness of child
tage and poor child outcome (e.g., Bornstein & activity regardless of developmental independence of
Bradley, 2003; Brooks-Gunn, Duncan, & Maritato, the child. In adolescence, such an awareness can be
1997). More important, however, was the finding that seen as either a component of effective monitoring
the link between socioeconomic disadvantage and (Dishion & McMahon, 1998) or as a reflection of
the development of children’s aggression was medi- child-to-parent disclosure (e.g., Kerr & Stattin, 2000;
ated by the parenting variables of care neglect, poor Kerr, Stattin, & Trost, 1999). It is interesting to note,
supervision, and harsh punitive discipline. Evans whether parental awareness of adolescent behavior is
(2004) called for an assessment of the cumulative a prerequisite for effective monitoring or even
risks that confront children reared in poverty and the whether parental solicitation of information from an
need to directly study the myriad of suboptimal physi- adolescent could have adverse effects has occasioned
cal and social conditions that are associated with pov- considerable debate in analyses of adolescent con-
erty and how those variables play important process duct problems (see Brody, 2003; Capaldi, 2003;
roles in the development of children. When the cur- Hayes, Hudson, & Matthews, 2003, 2004; Kerr &
rent findings are considered in concert with other Stattin, 2003). Although the methodology of the cur-
studies that involve observations of parent-child inter- rent study does not offer definitive information as to
actions in disadvantaged groups (e.g., Conger, Con- whether the parental awareness is attributable to the
ger, & Elder, 1997; DeGarmo, Forgatch, & Martinez, child’s willingness to disclose or the parent’s willing-
1999), an argument can be advanced that multi- ness to solicit information, it does suggest that the
method investigations of parenting include assess- awareness is an important factor in determining
ments of the unique contributions of care neglect, childhood aggression long before adolescence, when
ineffective supervision, and punitive discipline as monitoring is most often studied.
essential components in studies of disadvantageous More important, in social learning models of chil-
social and physical environments and the develop- dren’s conduct problems, poor monitoring has been
ment of children’s aggression. Moreover, the findings conceptualized as a consequence of the coercive pro-
underscore the view that research on the outcomes of cess, with parent-adolescent conflict seen as compro-
child maltreatment, and clinicians working with mising the monitoring process. In the current study,
families, need to examine the multiple forms of as well as in the Knutson et al. (2004) study, it would
maltreatment in concert rather than focusing on a seem that poor supervision (inadequate awareness) is
single form of maltreatment (Dong et al., 2004) a precursor to harsh discipline in the ontogeny of
It is impressive that social disadvantage operated child and, perhaps, adolescent aggression. The cur-
the same with respect to supervisory neglect in three rent replication of supervisory neglect, predicting
samples across two studies, and that supervision, mea- care neglect independent of social status, suggests
sured in very different ways, operated the same in the that awareness and tracking of basic needs and aware-
fifth-grade cohort in the Knutson et al. (2004) study ness of the child’s wants and desires are important
and in the current sample. In much research with prerequisites of skillful parenting and providing care.
young children, supervisory neglect is inferred when Somewhat related developmental studies have exam-
apparently preventable tragic events occur, such as ined how maternal responsiveness, attention, and
house fires (playing with a cigarette lighter) or drown- rejection of irritable infants can predict growth in
ing (not wearing a personal flotation device [PFD]). externalizing behavior (cf., Owens, Shaw, & Vondra,
Inadequate monitoring has played a central role in 1998; Shaw et al., 1998). In relation to other aspects of
social learning models of the development of antiso- parenting, the current data also suggest that aware-

CHILD MALTREATMENT / MAY 2005


104 Knutson et al. / NEGLECT AND HARSH PARENTING A REPLICATION

ness and tracking could be an important skill set for model does not include a consideration of possible
identifying and applying appropriate discipline. temperamental factors operating within the children
Unequivocal support for that hypothesis would or at a family level, with parents and children sharing
require cross-lagged evaluations within a prospective a biological substrate (see Rutter, Pickles, Murray, &
longitudinal design employing measures of tracking, Eaves, 2001, for a critical discussion of such designs).
neglect, and discipline that are developmentally In short, these replicated data suggest the addition of
appropriate across early and late childhood. supervisory and care neglect are important independ-
In addition to the improved understanding of the ent effects within developmental models; however,
specific role of parenting in the development of chil- much more work on neglectful parenting, harsh
dren’s aggression, the current study also makes a discipline, and social disadvantage in the develop-
methodological contribution to the measurement of ment of aggression is required.
neglectful parenting and harsh punitive discipline.
One of the chronic problems in maltreatment NOTE
research has been the reliance on administrative data
to establish the critical independent variables. The 1. Although this was a diverse sample, racial or ethnic
use of such administrative data has occasioned criti- variation was not a theoretical focus of the study design. In-
cism by the National Research Council (1993) and sufficient power at this point is available to conduct a mean-
ingful examination of between- or within-group variation in
spawned a number of efforts by federal agencies to
the theoretical process. At a reviewer’s request, tests of
consider strategies for operationally defining mal-
mean differences were conducted on the latent variable fac-
treatment (see Sternberg et al., 2004). The current tor scores; no statistically significant differences were found
findings document that it is possible to establish between minority and nonminority participants.
investigator-based multimethod and multisource
indicators of the core constructs of care neglect,
REFERENCES
supervisory neglect, and harsh punitive discipline,
and to link those constructs in a meaningful theoreti- Achenbach, T. M. (1991). Manual for the Teacher’s Report Form and
1991 Profile. Burlington, VT: Department of Psychology, Univer-
cal framework. In addition to providing a flexible and sity of Vermont.
robust tool for evaluating a theoretical model, the Achenbach, T. M. (1992). Revised Child Behavior Checklist.
SEM analyses provide a means for assessing the utility Burlington: University of Vermont.
Achenbach, T. M., & Rescorla, L. A. (2000). Manual for the ASEBA
and the validity of multimethod and multisource preschool forms and profiles. Burlington, VT: University of Ver-
strategies for measuring conditions of child maltreat- mont, Research Center for Children, Youth, and Families.
ment. Rather than treating either care neglect, super- Achenbach, T. M., & Rescorla, L. A. (2001). Manual for the ASEBA
school-age forms & profiles. Burlington: University of Vermont,
visory neglect, or harsh punitive discipline as a cate- Research Center for Children, Youth, and Families.
gorical variable, the SEM approach makes it possible Arbuckle, J. L. (1996). Full information estimation in the presence
to develop indices of deficient parenting that are dis- of incomplete data. In G. A. Marcoulides & R. E. Schumaker
(Eds.), Advanced structural equation modeling: Issues and techniques
tributed continuously, thus offering analytic advan- (pp. 243-277). Mahwah, NJ: Lawrence Erlbaum.
tage as well as being consistent with theoretical frame- Averill, J. R. (1982). Anger and aggression: An essay on emotion. New
works that treat child maltreatment as a continuous York: Springer-Verlag.
Bank, L., & Burraston, B. (2001). Abusive home environments as
variable rather than a categorical one (e.g., Green- predictors of poor adjustment during adolescence and early
wald et al., 1997; Knutson & Bower, 1994). In short, adulthood. Journal of Community Psychology, 29, 195-217.
the current findings establish the feasibility of Bolger, K. E., Patterson, C. J., & Kupersmidt, J. B. (1998). Peer rela-
tionships and self-esteem among children who have been mal-
studying circumstances of child maltreatment with- treated. Child Development, 69, 1171-1197.
out relying on administrative data to operationally Bornstein, M. H., & Bradley, R. H. (2003). Socioeconomic status,
define the critical variables. parenting, and child development. Mahwah, NJ: Lawrence
Erlbaum.
Although the current findings provide consider- Brody, G. H. (2003). Parental monitoring: Action and reaction. In
able support for a model that integrates the putative A. C. Crouter & A. Booth (Eds.), Children’s influence on family
effects of supervisory neglect, care neglect, and puni- dynamics: The neglected side of family relationships (pp. 163-169).
Mahwah, NJ: Lawrence Erlbaum.
tive discipline in the development of children’s Brooks-Gunn, J., Duncan, G. J., & Maritato, N. (1997). Poor fami-
aggressive behavior, the findings do not necessarily lies, poor outcomes: The well-being of children and youth. In
provide a complete picture. As suggested above, it is G. J. Duncan & J. Brooks-Gunn (Eds.), Consequences of growing up
poor (pp. 1-17). New York: Russell Sage Foundation.
critical to examine the role of these parenting factors Byrne, B. M. (1989). A primer for LISREL: Basic applications and pro-
in the context of a prospective longitudinal design; gramming for confirmatory factor analytic models. New York:
that is, it is important to show how these parental fac- Springer-Verlag.
Caldwell, B. M., & Bradley, R. H. (1978). Home observation and mea-
tors influence the developmental course of children’s surement of the environment. Little Rock: University of Arkansas
antisocial and aggressive behavior. In addition, the Press.

CHILD MALTREATMENT / MAY 2005


Knutson et al. / NEGLECT AND HARSH PARENTING A REPLICATION 105

Capaldi, D. M. (2003). Parental monitoring: Action and reaction. Greenwald, R. L., Bank, L., Reid, J. B., & Knutson, J. F. (1997). A
In A. C. Crouter & A. Booth (Eds.), Children’s influence on family discipline-mediated model of excessively punitive parenting.
dynamics: The neglected side of family relationships (pp. 171-179). Aggressive Behavior, 23, 259-280.
Mahwah, NJ: Lawrence Erlbaum. Hall, L. (1986). Parent awareness of children’s activities and preferences.
Clement, P., & Richard, R. (1976). Identifying reinforcers for chil- Unpublished doctoral dissertation, University of Iowa, Iowa
dren: A children’s reinforcement survey. In E. Mash & R. Terdal City.
(Eds.), Behavior therapy assessment: Diagnosis, design and evaluation Hanson, T. L., McLanahan, S., & Thomson, E. (1997). Economic
(pp. 207-216). New York: Springer. resources, parental practices, and children’s well-being. In G. J.
Conger, R., Patterson, G. R., & Ge, X. (1995). It takes two to repli- Duncan & J. Brooks-Gunn (Eds.), Consequences of growing up poor
cate: A mediational model for the impact of parents’ stress on (pp. 190-238). New York: Russell Sage Foundation.
adolescent adjustment. Child Development, 66, 80-97. Hayes, L., Hudson, A., & Matthews, J. (2003). Parental monitoring:
Conger, R. D., Conger, K. J., & Elder, G. H. J. (1997). Family eco- A process model of the parent-adolescent interaction. Behavior
nomic hardship and adolescent adjustment: Mediating and Change, 20, 13-24.
moderating processes. In G. J. Duncan & J. Brooks-Gunn (Eds.), Hayes, L., Hudson, A., & Matthews, J. (2004). Parental monitoring
Consequences of growing up poor (pp. 288-310). New York: Russell behaviors: A model of rules, supervision, and conflict. Behavior
Sage Foundation. Therapy, 35, 587-604.
DeGarmo, D. S., & Forgatch, M. S. (2004). Putting problem solving Hegar, R. L., & Yungman, J. J. (1989). Toward a causal typology of
to the test: Replicating experimental interventions for prevent- child neglect. Child and Youth Services Review, 11, 203-220.
ing youngsters’ problem behaviors. In R. D. Conger, F. O. Henry, B., Caspi, A., Moffitt, T. E., & Silva, P. A. (1996). Tempera-
Lorenz, & K. A. S. Wickrama (Eds.), Continuity and change in fam- mental and familial predictors of violent and nonviolent crimi-
ily relations: Theory, methods, and empirical findings (pp. 267-290). nal convictions: Age 3 to age 18. Developmental Psychology, 32,
Mahwah, NJ: Lawrence Erlbaum. 614-623.
DeGarmo, D. S., Forgatch, M. S., & Martinez, C. R. J. (1999). Jaffee, S. R., Caspi, A., Moffitt, T. E., & Taylor, A. (2004). Physical
Parenting of divorced mothers as a link between social status maltreatment victim to antisocial child: Evidence of an environ-
and boys’ academic outcomes: Unpacking the effects of SES. mentally mediated process. Journal of Abnormal Psychology, 113,
Child Development, 70(5), 1231-1245. 44-55.
DeMaster, B. R., Reid, J. B., & Twentyman, C. (1976). The effects of Jastak Associates. (1993). Wide Range Achievement Test. Wilmington,
different amounts of feedback on observers’ reliability. Behavior DE: Wide Range.
Therapy, 8, 317-329. Kendall-Tackett, K. A., & Eckenrode, J. (1996). The effects of
Dishion, T. J., & McMahon, R. J. (1998). Parental monitoring and neglect on academic achievement and disciplinary problems: A
the prevention of child and adolescent problem behavior: A developmental perspective. Child Abuse and Neglect, 20, 161-169.
conceptual and empirical formulation. Clinical Child and Family Kerr, M., & Stattin, H. (2000). What parents know, how they know
Psychology Review, 1, 61-75. it, and several forms of adolescent adjustment: Further support
Dishion, T. J., Patterson, G. R., & Greisler, P. C. (1994). Peer adap- for a reinterpretation of monitoring. Developmental Psychology,
tation in the development of antisocial behavior. In. L. R. 36, 366-380.
Huesmann (Ed.), Aggressive behavior: Current perspectives (pp. 61- Kerr, M., & Stattin, H. (2003). Parenting of adolescents: Action or
95). New York: Plenum. reaction? In A. C. Crouter & A. Booth (Eds.), Children’s influence
Dishion, T. J., Patterson, G. R., & Kavanagh, K. A. (1992). An exper- on family dynamics: The neglected side of family relationships (pp. 121-
imental test of the coercion model: Linking theory, measure- 151) Mahwah, NJ: Lawrence Erlbaum.
ment, and intervention. In J. McCord & R. Tremblay (Eds.), The Kerr, M., Stattin, H., & Trost, K. (1999). To know you is to trust you:
interaction of theory and practice: Experimental studies of interventions Parents’ trust is rooted in child disclosure of information. Jour-
(pp. 253-282). New York: Guilford. nal of Adolescence, 22, 737-752.
Dishion, T. J., Spracklen, K. M., Andrews, D. W., & Patterson, G. R. Knutson, J. F., & Bower, M. E. (1994). Physically abusive parenting
(1996). Deviancy training in male adolescent friendships. as an escalated aggressive response. In M. Potegal & J. F.
Behavior Therapy, 27, 373-390. Knutson (Eds.), The escalation of aggression in dyads and groups:
Dodge, K. A., Laird, R., Lochman, J. E., & Zelli, A. (2002). Multidi- Biological and social processes (pp. 195-225). Hillsdale, NJ: Law-
mensional latent-construct analysis of children’s social informa- rence Erlbaum.
tion processing patterns: Correlations with aggressive behavior Knutson, J. F., DeGarmo, D. S., & Reid, J. B. (2004). Social disadvan-
problems. Psychological Assessment, 14, 60-73. tage and neglectful parenting as precursors to the development
Dong, M., Anda, R. F., Felitti, V. J., Dube, S. R., Williamson, D. F., of antisocial and aggressive child behavior: Testing a theoretical
Thompson, T. J., et al. (2004). The interrelatedness of multiple model. Aggressive Behavior, 30, 187-205.
forms of childhood abuse, neglect, and household dysfunction. Knutson, J. F., & Schartz, H. A. (1997). Physical abuse and neglect
Child Abuse and Neglect, 28, 771-784. of children. In T. A. Widiger, A. J. Frances, H. A. Pincus, R. Ross,
Duncan, G. J., & Brooks-Gunn, J. (2000). Family poverty, welfare M. B. First, & W. Davis (Eds.), DSM-IV sourcebook (Vol. 3, pp. 713-
reform, and child development. Child Development, 71, 188-196. 804) Washington, DC: American Psychiatric Association Press.
Duncan, G. J., Brooks-Gunn, J., & Klebanov, P. K. (1994). Eco- Laing, J. A., & Sines, J. O. (1982). The Home Environment Ques-
nomic deprivation and early childhood development. Child tionnaire: An instrument for assessing several behavioral rele-
Development, 65, 296-319. vant dimensions of children’s environments. Journal of Pediatric
Eckenrode, J., Laird, M., & Doris, J. (1993). School performance Psychology, 7, 424-449.
and disciplinary problems among abused and neglected chil- Leventhal, T., Selner-O’Hagan, M. B., Brooks-Gunn, J.,
dren. Developmental Psychology, 29, 53-62. Bingenheimer, J. B., & Earls, F. J. (2004). The homelife inter-
Evans, G. W. (2004). The environment of childhood poverty. Ameri- view from the project on human development in Chicago
can Psychologist, 59, 77-92. neighborhoods: Assessment of parenting and home environ-
Flango, C. R. (1988). State courts’ jurisdiction and terminology for child ment for 3- to 15-year-olds. Parenting: Science & Practice, 4, 211-
abuse and neglect cases. Williamsburg, VA, National Center for 241.
State Courts. Loevinger, J. (1956). Objective tests as instruments of psychological
Giovannoni, J. (1985, November 10). Overview of issues on child theory. Psychological Reports, 3, 635-694.
neglect. In Child Neglect Monograph: Proceedings from a symposium Manly, J. T., Cicchetti, D., & Barnett, D. (1994). The impact of sub-
(pp. 1-6). Washington, DC: Department of Health and Human type, frequency, chronicity, and severity of child maltreatment
Services, Clearinghouse on Child Abuse and Neglect on social competence and behavior problems. Development and
Information. Psychopathology, 6, 121-143.

CHILD MALTREATMENT / MAY 2005


106 Knutson et al. / NEGLECT AND HARSH PARENTING A REPLICATION

Meehl, P. E. (1978). Theoretical risks and tabular asterisks: Sir Karl, Spinrad, T. L., Eisenberg, N., Harris, E., Hanish, L., Fabes, R. A.,
Sir Ronald, and the slow progress of soft psychology. Journal of Kupanoff, K., et al. (2004). The relation of children’s everyday
Consulting and Clinical Psychology, 46, 806-834. nonsocial peer play behavior to their emotionality, regulation,
McLanahan, S. S., & Sandefur, G. (1994). Growing up with a single and social functioning. Developmental Psychology¸ 40, 67-80.
parent. Cambridge, MA: Harvard University Press. Steinberg, L., Lamborn, S. D., Darling, N., Mounts, N. S., &
McLoyd, V. (1998). Socioeconomic disadvantage and child devel- Dornbusch, S. M. (1994). Over-time changes in adjustment and
opment. American Psychologist, 53, 185-204. competence among adolescents from authoritative, authoritar-
Murray, K. T., & Sines, J. O. (1996). Parsing the genetic and ian, indulgent, and neglectful families. Child Development, 65,
nongenetic variance in children’s depressive behavior. Journal 754-770.
of Affective Disorders, 38, 23-34. Sternberg, K. J., Knutson, J. F., Lamb, M. E., Bradaran, L. P., Nolan,
National Research Council. (1993). Understanding child abuse and C., & Flanzer, S. (2004). The Child Maltreatment Log: A PC-
neglect. Washington, DC: National Academy Press. based program for describing research samples. Child Maltreat-
Office of Human Development Services. (1981). National Study of ment, 9, 30-48.
the Incidence and Severity of Child Abuse and Neglect. Washington, Sullivan, P. M., & Knutson, J. F. (1998). The association between
DC: U.S. Department of Health and Human Services. child maltreatment and disabilities in a hospital-based pediatric
Office of Human Development Services. (1988). Study findings: sample. Child Abuse and Neglect, 22, 271-288.
Study of national incidence and prevalence of child abuse and neglect: Sullivan, P. M., & Knutson, J. F. (2000). Maltreatment and disabili-
1988. Washington, DC: U.S. Department of Health and Human ties: A population-based epidemiological study. Child Abuse and
Services. Neglect, 24, 1257-1274.
Owens, E. B., Shaw, D. S., & Vondra, J. I. (1998). Relations between Trocmé, N. (1996). Development and preliminary evaluation of
infant irritability and maternal responsiveness in low-income the Ontario Child Neglect Index. Child Maltreatment, 1, 145-155.
families. Infant Behavior and Development, 21(4), 761-777. Turner, R. J., & Wheaton, B. (1995). Checklist measurement of
Patterson, G. R. (1982). Coercive family process. Eugene, OR: Castalia. stressful life events. In S. Cohen, R. C. Kessler, & L. U. Gordon
Patterson, G. R., Reid, J. B., & Dishion, T. J. (1992). Antisocial boys. (Eds.), Measuring stress: A guide for health and social scientists
Eugene, OR: Castalia. (pp. 29-57). New York: Oxford University Press.
Patterson, G. R., & Yoerger, K. (1999). Intraindividual growth in Tymchuk, A. J., Lang, C. M., Dolyniuk, C. A., Berney-Ficklin, K., &
covert antisocial behavior: A necessary precursor to chronic and Spitz, R. (1999). The Home Inventory of Dangers and Safety
adult arrests? Criminal Behavior and Mental Health, 9, 24-38. Precautions-2: Addressing critical needs for prescriptive assess-
Peterson, L., Ewigman, B., & Kivlahan, C. (1993). Judgments ment devices in child maltreatment and in healthcare. Child
regarding appropriate child supervision to prevent injury: The Abuse and Neglect, 23, 1-14.
role of environmental risk and child age. Child Development, 64, U.S. Department of Health and Human Services, Administration
934-950. on Children Youth, and Families. (2001). Child maltreatment
Pettit, G. S., Bates, J. E., Dodge, K. A., & Meece, D. W. (1999). The 1999. Washington DC: U.S. Government Printing Office.
impact of after-school peer contact on early adolescent U.S. Department of Health and Human Services, Administration
externalizing problems is moderated by parental monitoring, on Children Youth, and Families. (2002). Child maltreatment
perceived neighborhood safety, and prior adjustment. Child 2000. Washington DC: U.S. Government Printing Office.
Development, 70, 768-778. Viermö, V. (1996). Factors in childhood that predict later criminal
Rusby, J. C., Estes, A., & Dishion, T. (1991). The Interpersonal Process behavior. Aggressive Behavior, 22, 87-97.
Code (IPC). Eugene: Oregon Social Learning Center. Widom, C. S. (1998). Child victims: Searching for opportunities to
Rutter, M., Pickles, A., Murray, R., & Eaves, L. (2001). Testing break the cycle of violence. Applied and Preventive Psychology, 7,
hypotheses on specific environmental causal effects on behav- 225-234.
ior. Psychological Bulletin, 127(3), 291-324. Wothke, W. (2000). Longitudinal and multi-group modeling with
Sedlak, A. J., & Broadhurst, D. D. (1996). Third national incidence missing data. In T. D. Little, K. U. Schnabel, & J. Baumert
study of child abuse and neglect: Final report. Washington, DC: U.S. (Eds.), Modeling longitudinal and multiple group data: Practical
Department of Health and Human Services, Administration for issues, applied approaches, and specific examples (pp. 269-281).
Children and Families, Administration on Children, Youth, and Mahwah, NJ: Lawrence Erlbaum.
Families, National Center on Child Abuse and Neglect. Zaidi, L. Y., Knutson, J. F., & Mehm, J. G. (1989). Transgenerational
Selner, M. E. (1992). The role of childhood sensation seeking and parental patterns of abusive parenting: Analog and clinical tests. Aggres-
supervision in the emergence of childhood psychopathology. Unpub- sive Behavior, 15, 137-152.
lished doctoral dissertation, University of Iowa, Iowa City. Zuravin, S. J. (1999). Child neglect: A review of definitions and
Selner, M. E., & Knutson, J. F. (1990, May 3). The development of the measurement research. In H. Dubowitz (Ed.), Neglected children:
Children’s Experience and Excitement Scale. Paper presented at the Research, practice, and policy (pp. 24-46). Thousand Oaks, CA:
Midwestern Psychological Association Meeting, Chicago. Sage.
Shaw, D. S., Winslow, E. B., Owens, E. B., Vondra, J. I., Cohn, J. F., &
Bell, R. Q. (1998). The development of early externalizing
John F. Knutson received his Ph.D. in psychology from Washing-
problems among children from low-income families: A
transformational perspective. Journal of Abnormal Child Psychol- ton State University. After a postdoctoral fellowship in medical psy-
ogy, 26(2), 95-107. chology at the University of Oregon Medical School, he joined the
Shields, A., & Cicchetti, D. (2001). Parental maltreatment and emo- faculty at The University of Iowa, where he has remained for 35
tional dysregulation as risk factors for bullying and victimization years. His current research focuses on the relational aspects of
in middle childhood. Journal of Clinical Child Psychology, 30, 349- neglectful parenting and punitive discipline, exposure to domestic
363. violence, the impact of maltreatment on children’s social competence
Sines, J. O., Clarke, W. M., & Lauer, R. M. (1984). Home Environ- and antisocial behavior, as well as the association between
ment Questionnaire. Journal of Abnormal Child Psychology, 12, maltreatment and disabilities.
519-529.
Spielberger, C. D., Jacobs, G. A., Russell, S., & Crane, R. (1983).
Assessment of anger: The State-Trait Anger Scale. In J. N.
David DeGarmo earned a doctorate in sociology from the Univer-
Butcher & C. D. Spielberger (Eds.), Advances in personality assess- sity of Akron in 1993 and was a postdoctoral fellow of the National
ment (Vol. 2, pp. 52-71). Hillsdale, NJ: Lawrence Erlbaum. Institute of Mental Health Family Research Consortium. His

CHILD MALTREATMENT / MAY 2005


Knutson et al. / NEGLECT AND HARSH PARENTING A REPLICATION 107

theoretical focus is on integrating social learning models and social Ministry Behavioral Health–Northern Region. Her areas of exper-
structure/social causation models to understand longitudinal fam- tise include parenting interventions, child abuse and neglect, and
ily adjustment within social class, family structure, support net- forensic psychology for children’s issues.
works, and ethnic contexts. Currently, he is conducting a 4-year
study of divorced father families focusing on the social and behav- John B. Reid received a doctorate in psychology (University of
ioral adjustment of divorced fathers, their children, and the impact Oregon, 1967) and held positions at the University of Wisconsin
of supportive confidants. (1967 to 1971) and the Oregon Research Institute (1971 to 1978).
With Gerald Patterson, he helped found the Oregon Social Learning
Gina Koeppl earned her Ph.D. in psychology from Duke Univer- Center in 1978, where he has worked on the development of life
sity in 1983 and then completed a postdoctoral fellowship in pediat- course models of externalizing behavior, positive and coercive
ric psychology at the University of Iowa. Since 2000, she has been an parenting, and using those models to inform the development of pre-
associate research scientist in the Department of Psychology associ- ventive and intervention trials. Currently, he is working on develop-
ated with the Rhinelander, Wisconsin, parenting laboratory, where mental intervention models for children in the child welfare system.
she is also the clinical director of outpatient behavioral services for

CHILD MALTREATMENT / MAY 2005

Anda mungkin juga menyukai