Elsevier

Child Abuse & Neglect

Volume 77, March 2018, Pages 198-210
Child Abuse & Neglect

Research article
Risk factors for child neglect: A meta-analytic review

https://doi.org/10.1016/j.chiabu.2018.01.006Get rights and content

Abstract

Knowledge of risk factors and their effects is vital for successfully preventing and reducing child neglect. This study provides a meta-analytic update of research on risk factors for child neglect. A total of 315 effect sizes were extracted from 36 primary studies and classified into 24 risk domains. Effects of 15 risk domains were significant and ranged from small (r = .110) to large (r = .372) in magnitude. Most risks were found at the parental level, such as having a history of antisocial behavior/criminal offending (r = .372); having a history of mental/psychiatric problems (r = . 259); having mental/physical problems (r = .207); and experiences of abuse in own childhood (r = .182). The effect of mother-related risk factors was not significantly different from the effect of father-related risk factors. It is concluded that child neglect is determined by multiple risk domains and that especially parent-related risk factors are important in preventing and reducing child neglect. Implications of the results for clinical practice are discussed.

Introduction

Child neglect has a relatively high prevalence rate, compared to other types of child maltreatment, such as physical and sexual abuse (Sedlak et al., 2010; Stoltenborgh, Bakermans-Kranenburg, Alink, & Van IJzendoorn, 2015). The impact of child neglect on the health and development of children is at least as negative as the impact of other types of child maltreatment (Norman et al., 2012), not to mention the high societal, medical, and personal costs associated with victimization of child neglect (Florence, Brown, Fang, & Thompson, 2013; Gilbert et al., 2009). Paradoxically, however, child neglect has received the least scientific and public attention compared to other forms of child maltreatment (Gilbert et al., 2009), which has been dubbed by researchers as the “neglect of neglect” (McSherry, 2007; Stoltenborgh, Bakermans-Kranenburg, & Van IJzendoorn, 2013).

Given the serious consequences of child neglect, scientific knowledge as well as clinical awareness of risk factors for child neglect is essential. From a scientific perspective, insight in risk factors and their effects may shed more light on the etiology of child neglect, whereas from a clinical perspective, risk and care needs assessment procedures may be improved. In this way, the proper care needs of high-risk children and their families can be better targeted to prevent (the recurrence of) child neglect. Although a variety of risk factors for child neglect were examined in the scientific literature, different populations, methods, and study designs were used in primary studies. This limits the possibility to be conclusive on not only the factors that can be designated as risk factors, but also on the strength of the effect of these risk factors. To overcome this problem, systematic and meta-analytical integrations of results of primary studies on the effects of risk factors are needed to estimate an “overall” (or mean) effect of risk factors as well as variables that may moderate these overall effects. The aim of the current study was to perform such a meta-analytic review.

Child neglect is a heterogeneous construct covering rather dissimilar negative child experiences, such as poor quality of supervision, inadequate or insufficient availability of food, lack of school attendance, and lack of required medical attention. In general, neglect refers to the omission of caretaking behavior that is necessary for a child's healthy development, whereas other forms of abuse most often have to do with harmful acts that are committed against a child (Mennen, Kim, Sang, & Trickett, 2010). However, two major issues contribute to difficulties in defining and operationalizing neglect (Dubowitz et al., 2005). First, it is debated by scholars whether neglect should only include actual harm or also potential harm. Second, there is an ongoing discussion on whether neglect should be seen as not meeting a child's basic needs from the child's perspective, or whether neglect should be seen as parental ommissions in care. Moreover, legal definitions of neglect vary by jurisdiction and between countries (Mennen, Kim, Sang, & Trickett, 2010). Throughout the years, multiple types and subtypes of child neglect have been proposed by different researchers. Examples of neglect categories are physical neglect, emotional neglect, medical neglect, mental health neglect, and educational neglect (Erickson & Egeland, 2002); cognitive neglect (Slack, Holl, Altenbernd, McDaniel, & Stevens, 2003); psychological and environmental neglect (Dubowitz, Pitts, & Black, 2004); lack of supervision (Kaufman Kantor et al., 2004); and denial of professional care and treatment (Knutson, DeGarmo, & Reid, 2004). Although consequences of child neglect can be very serious (Norman et al., 2012), there is limited societal agreement on whether parental ommissions in care are valid reasons for child welfare services to intrude in the life of a child and its family (Mennen, Kim, Sang, & Trickett, 2010).

Several theoretical models have been developed to explain child neglect, and in general, child neglect is perceived as the result of a complex interplay of risk factors present in children and their rearing environment. For instance, in the theoretical model of Belsky (1980), which was based on the ecological perspective on development of Bronfenbrenner (1979, 2000), risk factors can be present at four different levels: (1) the ontogenetic development of parents, which refers to the phenomena that negative parental experiences from the past are brought into their parenting behavior; (2) characteristics of the child and the family (i.e., the microsystem); (3) characteristics of the living environment (i.e., the exosystem); and (4) the attitude of society on children and child maltreatment (i.e., the macrosystem). In this model, the occurrence of child maltreatment (including neglect) is explained by a disbalance between risk and protective factors. A second theoretical model is the transactional model of Cicchetti and Rizley (1981), in which the reciprocal interactions between a child, the caregiver(s), and their environment play a central role. This model does not only stress the importance of risk factors of which the presence can fluctuate over time, but also the importance of protective factors, which can decrease the risk for child neglect. A last theoretical model is the one of Wolfe (1991), who stated that child maltreatment is an escalating process, located at the maladaptive end of a continuum of parenting behaviors. Put differently, this model implies that inadequate parenting behavior is the most important risk factor for child neglect. In sum, each of the described theoretical models assumes that the accumulation of - and interactions between - multiple protective and risk factors either increases or decreases the likelihood of victimization of child neglect rather than the presence of a single factor (see also Cicchetti & Carlson, 1989). Focusing on the accumulation of and interaction between different risk factors is therefore more promising for understanding the etiology of child neglect (MacKenzie, Kotch, & Lee, 2011).

Knowledge on risk factors is not only important for improving insight in the etiology of child neglect, but also for improving clinical practice aimed at preventing (the recurrence of) neglect. A theoretical model often applied in forensic (youth) care and in which risk factors play a central role is the Risk-Need-Responsivity model (RNR model; first formulated by Andrews, Bonta, & Hoge, 1990). The RNR model is widely used in penal law as a method for assessing and treating criminal offenders with the aim to reduce recidivism (Andrews, Bonta, & Wormith, 2011; Ward, Melser, & Yates, 2007). The model has three core principles: 1) the ‘risk’ principle stating that an intervention’s intensity should match an offender’s risk for recidivism; 2) the ‘needs’ principle stating that dynamic (i.e., changeable) risk factors associated with recidivism should be targeted in an intervention; and 3) the ‘responsivity’ principle stating that an intervention should be matched to characteristics of the offender, so that its potential positive impact is maximized (Andrews & Bonta, 2010). Although the RNR-model was specifically designed for preventing recidivism of criminal offenders, it may be very promising to apply the RNR principles in child welfare services. After all, child neglect, just like criminal recidivism, can be explained by the balance between multiple risk and protective factors present in the child and different ecological systems surrounding the child. To successfully bring the risk and needs principles into clinical practice, it is important that valid and reliable risk and needs assessment instruments are available in which only variables are assessed that have empirically been found to be risk factors for child neglect. This underlines the importance of meta-analytic research on the effects of risk factors for child neglect.

To the best of our knowledge, Stith et al., 2009 were the first to meta-analytically examine the effect of risk factors for different forms of child maltreatment. As for child neglect, they showed that risk factors involving the parent-child relationship and parental perceptions of the child (e.g., perceiving the child as problematic) were the two strongest predictors of child neglect. Relatively strong effects were also found for low social competence of children, high levels of parental stress, high levels of parental anger, and low parental self-esteem. Although it was the first and rather comprehensive meta-analytic study on risk factors for child maltreatment (including neglect), the study had several important limitations. A first shortcoming is that no moderator analyses were performed, implying that possible moderating effects of risk factor, study or sample characteristics were not revealed. For example, the type of neglect could be a moderating factor, since it cannot just be assumed that effects of risk factors are equal for different forms of child neglect. A second potential moderator could be the source that is used in primary research to identify neglect victimization, which is often either official records (mostly retrieved from CPS [Child Protection Services]) or self-reports on experiences of victimization. Given that prevalences of child neglect derived from official records are consistently lower than those derived from self-reports (Euser, Van IJzendoorn, Prinzie, & Bakermans-Kranenburg, 2010; Hussey et al., Hussey, Chang, & Kotch, 2006), official records may underestimate true prevalence rates and therefore, effects of risk factors in studies in which CPS data are used may be different from effects found in studies using self-reported data.

A second limitation is that Stith et al. (2009) averaged effects of risk factors in cases where multiple effect sizes were reported in primary studies. This not only leads to loss of information and thus less adequate estimates of effects of risk factors, but also to less statistical power in analyses, since not all relevant effect sizes are part of the dataset (see also Assink et al., 2015; Assink & Wibbelink, 2016). Furthermore, reservations can be put forward regarding the completeness of the data derived from the included primary studies. This is due to the search being conducted in only a single literature database combined with the fact that it was not examined whether the study results were influenced by forms of bias, such as publication or selection bias (Rosenberg, 2005). Finally, Stith et al. (2009) only included studies published until the year 2002, which calls for an update of their meta-analysis in which results of recent primary research is also included.

In sum, the aim of the present study was to meta-analytically examine effects of risk factors for victimization of child neglect. A second aim was to examine several variables as potential moderators of the effects of risk factors. The present study is relevant from both a scientific and clinical perspective, as updated knowledge on (effects) of risk factors for victimization of child neglect may improve our understanding of the etiology of child neglect and clinical practice in terms of risk and needs assessment of children and their families.

Section snippets

Inclusion criteria

Several inclusion criteria were formulated for the selection of primary studies. First, studies that were published between January 1st, 1990 and April 30th, 2016 were included. Studies performed prior to 1990 were excluded, primarily because earlier attitudes on and definitions of types of child maltreatment differ substantially from contemporary notions of types of child maltreatment (Gelles, 1980; Goode, 1971). In addition, early research on child maltreatment was, in general, conducted with

Descriptives

In the present review, a total of K = 36 primary studies were included. Most studies were conducted in the USA (k = 33), and single studies were conducted in South Korea (k = 1), Vietnam (k = 1), and the Netherlands (k = 1). The total sample size (N) was 729,840 children, of which n = 19,851 were victims of neglect, and n = 706,936 were not a victim of neglect. The victimization status could not be determined for n = 3053 children due to insufficient data. The sample size of the included

Discussion

Using a multilevel meta-analytic approach, we examined the effects of multiple domains of risk factors for child neglect. A total of 315 effect sizes of (potential) risk factors were retrieved from 36 primary studies, and clustered into 24 risk domains. The overall effect was significant for 15 risk domains, with magnitudes ranging from small (r = .110 for the domain of problematic family behavior and cognitions) to medium (r = .372 for the domain of parental history of antisocial behavior or

Conclusion

The present review contributes to the literature on risk factors for child neglect by examining the effect of multiple risk domains. The largest effects were found for parent related risks, such as having a history of antisocial behavior/criminal offending, having a history of mental/psychiatric problems, having mental or physical problems, and having a low educational level. Mental, physical, and/or behavioral problems of children were also found to increase the risk for child neglect.

References (80)

  • P. Sidebotham et al.

    Child maltreatment in the “children of the nineties”: A cohort study of risk factors

    Child Abuse and Neglect

    (2006)
  • T. Ward et al.

    Reconstructing the risk – need – responsivity model: A theoretical elaboration and evaluation

    Aggression and Violent Behavior

    (2007)
  • S.S. Wu et al.

    Risk factors for infant maltreatment: A population-based study

    Child Abuse & Neglect

    (2004)
  • D.A. Andrews et al.

    Classification for effective rehabilitation: rediscovering psychology

    Criminal Justice and Behavior

    (1990)
  • D.A. Andrews et al.

    The risk-need-responsivity (RNR) model: does adding the good lives model contribute to effective crime prevention?

    Criminal Justice and Behavior

    (2011)
  • C.M. Arata et al.

    Single versus multi-type maltreatment: an examination of the long-term effects of child abuse

    Journal of Aggression, Maltreatment & Trauma

    (2005)
  • M. Assink et al.

    The intergenerational transmission of child maltreatment: A three-level meta-analysis

    Manuscript

    (2017)
  • M. Assink et al.

    Fitting three-level meta-analytic models in R: A step-by-step tutorial

    The Quantitative Methods for Psychology

    (2016)
  • J. Belsky

    Child maltreatment: an ecological integration

    American Psychologist

    (1980)
  • J. Belsky

    Etiology of child maltreatment: A developmental ecological analysis

    Psychological Bulletin

    (1993)
  • M. Bittler et al.

    Child abuse and abortion availability

    The American Economic Review

    (2002)
  • U. Bronfenbrenner

    The ecology of human development

    (1979)
  • U. Bronfenbrenner

    Ecological theory

  • M.W.L. Cheung

    Modeling dependent effect sizes with three-level meta-analyses: A structural equation modeling approach

    Psychological Methods

    (2014)
  • D. Cicchetti et al.

    Developmental perspectives on the etiology, intergenerational transmission, and sequelae of child maltreatment

    New Direction for Child and Adolescent Development

    (1981)
  • P. Cohen et al.

    Child abuse and neglect and the development of mental disorders in the general population

    Development and Psychopathology

    (2001)
  • K. Connell-Carrick

    A critical review of the empirical literature: identifying correlates of child neglect

    Child and Adolescent Social Work Journal

    (2003)
  • H. Cooper

    Research synthesis and meta-analysis

    (2010)
  • K.K. Davison et al.

    Fathers' representation in observational studies on parenting and childhood obesity: A systematic review and content analysis

    American Journal of Public Health

    (2016)
  • H. Dubowitz et al.

    Measurement of three major subtypes of child neglect

    Child Maltreatment

    (2004)
  • S.J. Duval et al.

    Trim and fill: A simple funnel-plot-based method of testing and adjusting for publication bias in meta-analysis

    Biometrics

    (2000)
  • S.J. Duval et al.

    A nonparametric "trim and fill" method of accounting for publication bias in meta-analysis

    Journal of the American Statistical Association

    (2000)
  • S.J. Duval

    The trim and fill method

  • M.F. Erickson et al.

    Child neglect

  • E.M. Euser et al.

    Prevalence of child maltreatment in the Netherlands

    Child Maltreatment

    (2010)
  • G.A. Ferguson

    Statistical analysis in psychology & education

    (1966)
  • D.M. Fergusson et al.

    The stability of child abuse reports: A longitudinal study of the reporting behaviour of young adults

    Psychological Medicine

    (2000)
  • A. Field

    London, United Kingdom: Sage Publications

    Discovering statistics using SPSS

    (2005)
  • Cited by (0)

    View full text