Intergenerational cycle of maltreatment: a popular concept obscured by methodological limitations☆
Introduction
The maltreatment of children by parents is a major social problem that has been an important focus of researchers. Clinical lore, anecdotal accounts, and research evidence have linked parents’ maltreatment of children to their own maltreatment experiences as children. Early investigators of the intergenerational transmission of child maltreatment set the stage for a violence breeds violence perspective (Curtis, 1963). This was reinforced and conceptually expanded in subsequent work (e.g., Kaufman and Zigler 1987, Steel and Pollack 1968), with contemporary data suggesting that about one-third of individuals who were maltreated as children will, as parents, abuse or neglect their own children Belsky 1993, Kaufman and Zigler 1987.
However, this intergenerational transmission hypothesis has also received a good deal of criticism, and concern has been raised about the adequacy of relevant data and research (Ertem, 2000). Widom (1989b) asserts that there is little evidence supporting the claim that abuse leads to abuse. And controversy remains surrounding whether causal pathways determining the perpetuation of child maltreatment by parents have been sufficiently explored. In fact, much of the existing research is plagued with at least three fundamental weaknesses that may account for some of the discrepancies and part of the resulting controversy. These weaknesses include: (1) identifying subjects by looking at case status only; (2) a lack of continuous measurement of maltreatment variables; and (3) insufficient operational definitions that fail to account for the unique and combined effects of both abuse and neglect.
One way to examine the intergenerational expression of child maltreatment is to study the influence of maltreatment experiences on later parenting practices. As the majority of maltreatment experiences occur within a family or parenting context, we can examine the most direct expression of transmission by studying the relationships between these two processes. While obvious, this approach is seldom used and upon reviewing the bodies of literature on maltreatment and parenting practices, it is apparent that these two domains have rarely been integrated or analyzed simultaneously.
We begin with a brief discussion of the three general limitations with the maltreatment literature, then discuss parenting practices. Following this we present our conceptual model and describe the current study.
A frequently cited problem in the salient literature surrounds the operational definitions of childhood abuse, neglect, and maltreatment. Most researchers define maltreatment in terms of cases adjudicated by Child Protective Service (CPS) agencies. For example, in her review of the neglect literature, Zuravin (1999) found that 18 of the 25 identified studies had used this means to operationalize neglect, using it either as the exclusive criterion (n = 14) or as one used in combination with the researchers’ own definition (n = 4). This method of classification is problematic as agencies differ across governmental jurisdictions with respect to their criteria for determining maltreatment or for interpreting their particular criteria. But even within the same jurisdiction, many factors affect the labeling of a particular case. Belsky (1993) has noted that classifications of maltreatment often “emerge after much negotiation and consultation with the family, judicial authorities, and others. Exactly how these formal labels map onto actual behavior is not always clear” (p. 413).
Clinical judgment, socio-cultural norms for expected behavior, and other factors also play a role (Garbarino, Guttmann, & Seeley, 1986). Miller-Perrin and Perrin (1999) state that “the definitional criteria… deemed important, vary from one audience to the next, as do the distinctions between illegitimate and legitimate aggression” (p. 49). Therefore “abuse, maltreatment, and neglect are merely labels used to describe an indeterminate range of negative behaviors” (Miller-Perrin & Perrin, 1999, p. 49). And, since both abuse and neglect tend to be heterogeneous and have comorbid presentation, arbitrarily defining a case as an abuse case may have the tendency to overlook other critical aspects, such as neglect. This definitional variability does not allow for an understanding of the multiple types of child maltreatment and replicable conclusions to be drawn. Further, definitional variable and the consequent ambiguity has delayed the implementation of effective interventions and solutions (Ohlin & Tonry, 1989).
When neglect is operationalized in terms of adjudicated cases, a dichotomized, either-or criterion is adopted, blindly avoiding the reality that childhood maltreatment exists to some degree, along various continua. Manly, Cicchetti, and Barnett (1994) state that most research is conducted in this manner, focusing on “maltreatment as a dichotomous variable” with “scant empirical attention being devoted to the multidimensional nature within the broad rubric of maltreatment” (p. 121). It is obvious that instances of child maltreatment range in severity and need to be conceptualized accordingly. When this range is not considered, the nuances and variance (both statistical and practical) of child maltreatment are lost.
A reasonable approach, that is seldom used, would be to incorporate psychometrically reliable and valid measures of child maltreatment that yield continuous scores for the different commonly accepted subtypes of maltreatment. This would capture more of the true expression or continua of child maltreatment, but is rarely used. Straus (1992) found that of 617 articles published in Child Abuse & Neglect, 86% used no instrument to measure any variable whatsoever. Concern about the lack of measurement is reinforced by the National Research Council Panel on Research on Child Abuse and Neglect’s (1993) call for the development of “reliable and valid clinical-diagnostic and research instruments for the measurement of child maltreatment… ” (p. 345).
Most studies fail to differentiate between children who have been abused and those who have been neglected (Widom, 1989a). Knutson (1995) states that “One of the major problems in the literature on maltreatment is… the aggregating of physical abuse, neglect, and sexual abuse into a single category of child maltreatment” (p. 401). While there are common aspects, there are important differences. There are six commonly accepted subtypes of child maltreatment—emotional neglect, physical neglect, educational neglect, emotional abuse, physical abuse, and sexual abuse (Miller-Perrin & Perrin, 1999). These categories were adopted in the National Incidence Studies (NIS; e.g., Sedlak & Broadhurst, 1996), and should be empirically studied as both related and unique maltreatment subtypes. Discerning the consequences of particular types of child maltreatment is an important advance needed in the contemporary literature.
If there is a cycle or intergenerational transmission of maltreatment, the conduit is parenting practices. These are the behavioral expressions that pass on maltreatment from one generation to another. This is a logical perspective, for childhood maltreatment almost inevitably is perpetrated by the child’s parent or caretaker (National Research Council Panel on Research on Child Abuse and Neglect, 1993, p. 65). In fact, many problems of children are known to arise from their early environment (Cicchetti & Toth, 1998) and family interactions (Cowan, Powell, & Cowan, 1998). The family’s influence on the child is often thought to exert itself through specific parenting practices Capaldi and Clark 1998, Darling and Steinberg 1993, Rohner 1991, Rutter 1995. Parenting style affects the quality of parent-child interactions as well as the “pervasive interactional climate” (Mize & Pettit, 1997, p. 291). For example, parents who experienced low levels of nurturance during their own childhood endorse the use of physical punishment (Ringwalt, Browne, Rosenbloom, Evans, & Kotch, 1989). And mothers who perceived more rejection during their own childhood display more negative affect toward their own children (Belsky, Youngblood, & Pensky, 1989). Conversely, mothers who had positive relationships with their mothers in childhood are more likely to be securely attached and responsive to their infants (Gara, Rosenberg, & Herzog, 1996). Given these patterns and findings, it is critical to integrate parenting practices in studies of the intergenerational cycle of child maltreatment.
When considering our conceptual and analytic framework several concepts are important to understand. First, all variables are conceptualized and operationalized using continuous measures, allowing for a more complete understanding of the continua of relationships, rather than using arbitrary cutoff points. To understand maltreatment and parenting practices as continuous variables not only respects what occurs in real-life, but also increases variance and therefore the sensitivity in statistical analyses employed.
The relationships between Family Abuse, Family Neglect, and Child Maltreatment will be expressed in terms of common or shared variance, and specific or nonshared variance. Child Maltreatment is statistically conceptualized as the shared variance between the latent factors representing Family Abuse and Family Neglect (see Fig. 1). The shared relationships are statistically captured as latent constructs, as are those which are unique to both Family Abuse and Family Neglect (see small circles). This structure allows us to separate the effects of abuse and neglect from the more general influence of Child Maltreatment, allowing for a more complete understanding of relationships, predictors, and outcomes (Newcomb, 1994).
Likewise, Poor Parenting is statistically conceptualized as the shared variance between the variables of warmth, aggression, undifferentiated rejection, and neglect (Myers, Newcomb, Richardson, & Alvy, 1997). This structure (see Fig. 1) allows us to separate the effects of specific forms of parenting from the more general concept of Poor Parenting practices.
To study the intergenerational transmission of Child Maltreatment, we are looking at how childhood maltreatment experiences predict or lead to aspects of parenting. The constructs utilized provide reliable and valid accounts of both processes and reasonably represent what we are investigating.
Therefore, our focus is not on the incidence of intergenerational transmission, but rather on a relatively neglected area of research, its structure and linkages. Specifically, we examine which childhood experiences predict later parenting behaviors in an ethnically diverse, longitudinal community sample from the Los Angeles area. We examine the structure of both fathers’ and mothers’ experiences independently, since we are hypothesizing different structures by gender.
To address definitional problems, we operationalize childhood maltreatment and parenting practices with two multiscale, empirically validated, and reliable psychometric instruments, the Childhood Trauma Questionnaire (CTQ; Bernstein et al., 1999) and the Parental Acceptance-Rejection Questionnaire (PARQ; Rohner, 1991). These instruments have scales that yield scores for multiple types of childhood maltreatment and parenting expressions along several continua. Structural Equation Modeling will be used to explore the complex relationships.
Section snippets
Participants
Participants in this study (N = 383) are parents from the UCLA Longitudinal Study of Growth and Development (Newcomb, 1997a). Beginning in 1976, data were collected from a community sample of 7th through 9th grade students from 11 Los Angeles County schools, with oversampling to appropriately include representative proportions of minority and lower socioeconomic groups. Data used in this study are cross-sectional from study Year 21 and include only those individuals who have children, and where
Variable characteristics
An examination of the univariate kurtosis and skewness estimates revealed that most of the variables were normally distributed (see Table 2). Therefore, we used standard maximum likelihood (ML) estimation for most of the models. However, three variables in the female data and two variables in the male data had moderate kurtosis. To account for this violation of normality, we retested our final path models using ML-robust, which adjusts for nonnormality of the data (Bentler, 1995).
Gender differences on means and correlations
One important
Discussion
This study supports the intergenerational cycle of child maltreatment in such a way that several weaknesses in much of existing research were controlled. Specifically, we used reliable and valid psychometric instruments to measure various continua of child maltreatment severity and subsequent parenting practices within an ethnically diverse community sample of mothers and fathers. We begin by summarizing our findings, suggesting some clinical implications, reviewing hypothesized mechanisms of
References (63)
- et al.
Validity of the Childhood Trauma Questionnaire in an adolescent psychiatric population
Journal of the American Academy of Child & Adolescent Psychiatry
(1997) - et al.
The abused child as parent
Child Abuse & Neglect
(1996) - et al.
The measurement of maltreatmenta comparison of approaches
Child Abuse & Neglect
(1995) - et al.
Long-term effects of childhood maltreatment history on gender-related personality characteristics
Child Abuse & Neglect
(1998) Intergenerational transmission of parentinga review of studies in nonclinical populations
Developmental Review
(1992)Etiology of child maltreatment a developmental-ecological analysis
Psychological Bulletin
(1993)- et al.
Child-rearing history, marital quality, and maternal affectintergenerational transmission in a low risk sample
Development and Psychopathology
(1989) Multivariate analysis with latent variablescausal modeling
Annual Review of Psychology
(1980)Comparative fit indexes in structural models
Psychological Bulletin
(1990)EQS structural equations program manual
(1995)
Covariance structure analysisstatistical practice, theory and directions
Annual Review of Psychology
Initial reliability and validity of a new retrospective measure of child abuse and neglect
American Journal of Psychiatry
A self-trauma model for treating adult survivors of severe child abuse
Structural equational modeling with EQS and EQS/Windows Basic Concepts, Applications, and Programming
Prospective family predictors of aggression toward female partners for at-risk young men
Developmental Psychology
Perspectives on research and practice in developmental psychopathology
Parenting interventionsa family systems perspective
Violence breeds violence-perhaps?
American Journal of Psychiatry
Social networks, quality of child rearing, and child development
Child Development
Parenting style as contextan integrative model
Psychological Bulletin
Intergenerational continuity of child physical abusehow good is the evidence?
Lancet
The psychologically battered childstrategies for identification, assessment and intervention
Predictors of Latino men’s paternity in teen pregnancytest of a mediational model of childhood experiences, gender role attitudes, and behaviors
Journal of Counseling Psychology
On the “remembrance of things past” a longitudinal evaluation of the retrospective method
Psychological Assessment
Do abused children become abusive parents?
American Journal of Orthopsychiatry
Psychological characteristics of maltreated childrenputative risk factors and consequences
Annual Review of Psychology
Treatment and intervention for child victims of violence
Child sexual and physical abuse and alcoholisma review
Journal of Studies on Alcohol
Perceived abuse and neglect as risk factors for suicidal behavior in adolescent inpatients
Journal of Nervous and Mental Disease
The impact of subtype, frequency, chronicity, and severity of child maltreatment on social competence and behavior problems
Development and Psychopathology
Cited by (0)
- ☆
This study was supported by Grant No. DA01070 from the National Institute on Drug Abuse.
Note: This article is being published without the benefit of the author’s review.