Identification and Definitions
Despite vigorous debate over the last two decades, little progress has been made in constructing clear, reliable, valid, and useful definitions of child abuse and neglect. A 1977 report commissioned by the National Center on Child Abuse and Neglect summed up the situation as follows (Martin, 1978:1):
The issue of defining abuse and neglect is one of central importance and logically precedes a discussion of incidence, etiology and treatment. The vagueness and ambiguities that surround the definition of this particular social problem touch every aspect of the fieldreporting system, treatment program, research and policy planning.
The difficulties facing the field in constructing definitions are described elsewhere (Aber and Zigler, 1981; Cicchetti and Carlson, 1989; Giovannoni, 1989, 1992; Giovannoni and Becerra, 1979; McGee and Wolfe, 1991; Straus and Gelles, 1986; Zuravin, 1991). They include such factors as:
lack of social consensus over what constitutes dangerous or unacceptable forms of parenting.
uncertainty about whether to define maltreatment based on adult characteristics, adult behavior, child outcome, environmental context, or some combination.
conflict over whether standards of endangerment or harm should be used in constructing definitions.
confusion over the multiple purposes to which definitions are to be put: scientific, legal, and clinical.
the fact that the meaning of an act toward a child may vary greatly with the child's age, gender, relation to the actor, ethnicity, and contextual factors.
variations in age definitions of a ''child" (which may range from 0 to 16 or 0 to 18 depending on state definitions) and the existence of a special subpopulation (commonly termed teens, youth, adolescents, teenagers) whose physical, developmental, and social characteristics are significantly different from those of children (Office of Technology Assessment, 1987).
uncertainty as to whether definitions should reflect discrete categories of maltreating behaviors or draw on dimensions of maltreatment within a broader spectrum of "normal" behaviors.
difficulties in developing definitions that are both meaningful and capable of being operationalized. An ideal theoretical construct that cannot be applied to real populations is of limited usefulness.
Given these difficulties, the development of standardized definitions is a challenging task. However, consistent definitions are necessary for better measurement and instrumentation in the field. Attempts to reach consensus on clear operational measures need to be made, and limitations recognized and researched. From these attempts, more refined measures can be developed. To take an analogy from the field of mental health, the system of classification of mental disorders contained in the Diagnostic and Statistical Manual of Mental Disorders, published by the American Psychiatric Association, currently undergoing its fourth revision (DSM-III-R, APA, 1987), has played a major role in modern advances in psychiatric research, despite its limitations. Research on child maltreatment could benefit greatly from the development of an analogous diagnostic system for child abuse and neglect. Cicchetti and Barnett (1991) wrote:
Perhaps investigators in the field of child maltreatment can learn a valuable lesson from research in the area of psychopathology (Askikal and Webb, 1978; Eaton and Kessler, 1985; Goodwin and Guze, 1984; Meehl, 1959/1973). Research in the diagnosis of mental disorders for research purposes has been greatly enhanced by two improvements: (1) the establishment of more operational, explicit, hence reliable criteria for each "type" of mental disorder; and (2) the development of standardized, structured diagnostic interviews. The first advances the criterion or nosological variance in diagnosis, while the second significantly lowers the diagnostician variationvariance owing to different diagnosticians who may employ different clinical interviewing styles, disparate assumptions about signs and symptoms of disorders, and so on.
A basic requirement for scientific progress on research on child maltreatment is the availability of authoritative, valid and operational measures of child abuse and neglect. Definitions are essential to the development of such measures in order to conduct and interpret studies of prevalence and incidence of child maltreatment. In particular, issues of definition obscure comparisons of incidence rates from two governmental national incidence studies that rely heavily on administrative reports of child abuse and neglect (commonly referred to as NIS-1 and NIS-2) (National Center for Abuse and Neglect, 1981; 1988b). Technical issues such as the procedures used to generalize the findings of the NIS studies to the total U.S. population are complex and troublesome. The Healthy People 2000 (NCCAN, 1990) report avoids specific targets for reducing the level of child maltreatment "because of difficulties in obtaining valid and reliable measurement of child maltreatment" (Institute of Medicine, 1990). Administrative data on child maltreatment from child protective service agencies are very difficult to interpret given the wide differences in the way definitions are operationalized in different localities.
Definitions are also essential in order to develop measures to compare and generalize results of different studies on the effects of maltreatment and on the primary, secondary or tertiary prevention of maltreatment. In the absence of consensus about child maltreatment measures, existing studies employ an array of different measures that can yield results that conflict or are hard to interpret.
Review Of Definitions
A 1989 conference convened by the National Institute of Child Health and Human Development recommended that maltreatment be defined as "behavior towards another person, which (a) is outside the norms of conduct, and (b) entails a substantial risk of causing physical or emotional harm. Behaviors included will consist of actions and omissions, ones that are intentional and ones that are unintentional" (Christoffel et al., 1992). The term child maltreatment refers to a broad range of behaviors that involve risk for the child.
Four general categories of child maltreatment are now generally recognized: (1) physical abuse, (2) sexual abuse, (3) neglect, and (4) emotional maltreatment. Each category, in turn, covers a range of behaviors. Physical abuse includes scalding, beatings with an object, severe physical punishment, and a rare form of the abuse called Munchausen by proxy, wherein an adult will feign or induce illness in a child in order to attract medical attention and support. Sexual abuse includes incest, sexual assault by a relative or stranger, fondling of genital areas, exposure to indecent acts, sexual rituals, or involvement in child pornography. Child neglect is the
presence of certain deficiencies in caretaker obligations (usually the parent, although neglect can be found in residential centers or foster care homes) that harm the child's psychological and/or physical health. Child neglect covers a range of behaviors including educational, supervisory, medical, physical, and emotional neglect, and abandonment, often complicated by cultural and contextual factors. Several authors (Mrazek and Mrazek, 1985; Zuravin, 1991) have noted the relative lack of attention to definitional issues of child neglect, particularly given its greater reported prevalence (NCCAN, 1981, 1988b; Wolock and Horowitz, 1984). Emotional maltreatment, a recently recognized form of child victimization, includes such acts as verbal abuse and belittlement, symbolic acts designed to terrorize a child, and lack of nurturance or emotional availability by caregivers.1
Aber and Zigler (1981) distinguished medical-diagnostic, legal, sociological, and research approaches to the definition of child maltreatment. The medical-diagnostic approach "aims to identify a pathological process or condition underlying a symptom pattern in a way that enables a therapeutic intervention" (Aber and Zigler, 1981). It dates from published accounts of bone fractures to children of unspecified cause (Caffey, 1946). The seminal paper by Kempe and colleagues (1962) coined the term battered child syndrome to "characterize a clinical condition in young children who have received serious physical abuse, generally from a parent or foster parent."
The Kempe et al. (1962) article initiated a decade of supremacy of the medical diagnostic approach to the definition of child abuse that focused attention on identifying and treating the abusive parent. However, the approach has important limitations. The diagnosis through manifest physical symptoms in the child emphasizes the more extreme forms of physical abuse over other forms of maltreatment (such as child sexual abuse) and leads to a narrow definition of the problem. Also, the attempt to tie physical injury to the child, a descriptive characteristic with presumably high reliability, to an abusive act by the parent or caretaker is fraught with assumptions and difficulties. From a scientific perspective, the above definition lacks precision, and its reliability and validity are open to question.
Legislative activity in the 1970s, particularly the Child Abuse Prevention and Treatment Act of 1974, placed emphasis on the precise nature of abusive acts themselves and their consequences for the child. It formulated broader legal definitions of child abuse that encompass emotional injury, neglect, parental deprivation of medical services, and factors deleterious to children's moral development (Cicchetti and Barnett, 1991). Legal definitions in the Juvenile Justice Standards Project (Wald, 1977) encompass more than physical abuse, but stringent standards within each category are focused on the issue of legal intervention, and may not be commonly used in other settings:
Physical harm is defined as "disfigurement, impairment of bodily functions or other serious physical injury."
Emotional damage is evidenced by "severe anxiety, depression or withdrawal, or untoward aggressive behavior towards self or others, and the child's parents are unwilling to provide treatment for him/her."
Sexual abuse is limited to those cases in which the child is "seriously harmed physically or emotionally thereby." Such a definition excludes many cases in which consequences are not immediately apparent but emerge at later points in the developmental cycle.
In these legal definitions, coercive state intervention on behalf of the child is weighed against the potential harm of separation from the caregiver and the problems of state intrusion into family autonomy. These standards, although influential in legal circles, were not adopted by the American Bar Association, largely due to opposition that they would tie the hands of judges. Indeed, the desire for judicial discretion conflicts with the precision and clarity needed for scientific definitions of child maltreatment.
The focus in the sociological approach is on the act of maltreatment; the goal is to label and control social deviance (Cicchetti and Barnett, 1991). The sociological view of the problem is broad and encompasses the study of milder forms of maltreatment that may not result in physical trauma requiring medical treatment or form legal grounds for state intervention (Gelles and Straus, 1988). Whereas a fair degree of consensus might be attainable about extreme forms of maltreatment, such as abuse or neglect leading to death of the child, definitional and measurement issues become more thorny when less severe manifestations are considered. Sociologists such as Gil (1975) and Pelton (1978) viewed child maltreatment from a sociocultural perspective and emphasized the need to take into account societal contexts such as overcrowding, poverty, inadequate services, and large family sizes.
Research definitions of child maltreatment aim at developing measures that allow reliable and valid empirical studies of incidence and prevalence rates, longitudinal studies of etiology and sequelae, and comparisons of treatments. Aber and Zigler (1981) recommended the formulation of broad research definitions, in which the main classification principle is the nature of acts, or "descriptive compartmentalization of behaviors" (Zigler and Phillips, 1981), as opposed to child outcomes.
In the absence of explicit definitions, the difficulty of operationalizing concepts of maltreatment in early studies (e.g., Giovannoni and Billingsley, 1970; Green et al., 1974) was usually resolved by simply using the label assigned to the act by responsible agencies, including hospitals, child protection agencies, police, and courts (Giovannoni and Becerra, 1979). This approach continues to be adopted in the use of empirical data based on child protective services records (NCCAN, 1988a,b). Defining child maltreat-
ment by labels assigned by professionals in the field can provide useful descriptive information with minimal data collection efforts, but the absence of precise, objective criteria complicates comparability among measures of maltreatment in subpopulations defined by locality. For example, differences in incidence of case reports over time reported by social service agencies may reflect changes in definition imposed by legislation or resource changes, rather than changes in the incidence of maltreatment itself. Careful studies of the processes leading to case definitions are necessary to quantify and adjust for variations in the manner in which child maltreatment is operationalized.
Some recent studies have developed operational definitions that attempt to define acts of maltreatment more precisely, rather than relying on professional opinion. Zuravin (1991) noted this as a step in the right direction, but argued that comparability between studies is compromised by wide variations in the definitions adopted. Specifically, she analyzed and found differences in definitions of physical abuse, physical neglect, and child neglect from a number of empirical studies. Wyatt and Peters (1986) demonstrated empirically the impact of variations in definitions of sexual abuse on the estimated prevalence rates reported in a number of studies.
Principles Underlying Research Definitions
Zuravin (1991) suggested four general principles for formulating research definitions of child abuse:
Formulation of the specific objectives the definition must serve. Ross and Zigler (1980:294) noted that the current lack of agreement about definitions can be explained, at least in part, by failure to recognize that no single definition is capable of "fulfilling all the functions that social scientists and social service professionals would like." Aber and Zigler (1981) proposed the development of distinct sets of definitions for legal settings, case management settings, and research. Cicchetti and Barnett (1991:352) found this conceptualization useful but observed that "as the quality of data on the sequelae of maltreatment evolves sufficiently to allow stronger causal statements (e.g., about the role of emotional neglect in the development of emotional damage), then the legal thinker, the clinician and the researcher can begin to speak the same language, and the three sets of definitions will begin to converge."
Division into homogeneous subtypes. Cicchetti and Barnett (1991) delineated five major child maltreatment categories: physical abuse, physical neglect, sexual abuse, emotional maltreatment, and moral/legal/educational maltreatment. Emotional maltreatment was considered the least developed of the major subtypes. They recognized, however, that each of
these major categories displays considerable heterogeneity in character, severity, and potential consequences. Further subtyping into more homogeneous categories is an important consideration in developing a classification system.
An example of detailed subtyping was presented by Zuravin (1991), who proposed the 14 subtypes of physical neglect (which she defined as encompassing the category of moral/legal/educational maltreatment in Cicchetti and Barnett's  schema): (1) refusal to provide physical health care; (2) delay in providing physical health care; (3) refusal to provide mental health care; (4) delay in providing mental health care; (5) supervisory neglect; (6) custody refusal; (7) custody-related neglect; (8) abandonment/desertion; (9) failure to provide a permanent home; (10) personal hygiene neglect; (11) housing standards; (12) housing sanitation; (13) nutritional neglect; (14) educational neglect. This scheme is presented as an illustration and is not necessarily endorsed by the panel.
Conceptual clarity. For a classification system to be reliable and valid, it is important to specify clearly every criterion that a behavior must meet in order to be classified in a category, and that categories are distinctly defined.
Measurability of operational translations. A classification scheme of great conceptual elegance is of limited utility in empirical research unless it can be operationalized. That is, the conceptual definition has to be converted into specific behaviors that can be measured by observation, interview, or some other practical means. The panel believes strongly that progress on child maltreatment research requires not only the development of intelligent classification schemes, but also the development of standardized field instruments, such as clinical checklists or structured survey questionnaires, with documented psychometric properties. In other words, instruments need to be developed, together with associated documentation such as training manuals; interrater reliability studies need to be conducted to document consistency together with studies of construct validity and (in the longer term) predictive validity; culturally sensitive versions of these instruments need to be developed for ethnic subpopulations;2 and consideration for literacy or English as a foreign language needs attention. As noted above, work in psychiatric classification provides a useful paradigm for this work, although the approach needs to move beyond the diagnosis of a syndrome toward an effort to describe dimensions of maltreatment that may be embedded within a wide spectrum of behaviors.
Specific Definitional Issues
In addition to these general principles, the following specific issues in the operationalization of child maltreatment definitions need to be consid-
ered. Our discussion borrows heavily from the useful discussions of Cicchetti and Barnett (1991) and Zuravin (1991).
Endangerment Versus Demonstrable Harm
Zuravin (1991) observed that there is general conceptual agreement that abuse represents acts of commission and neglect represents acts of omission (Giovannoni, 1971) by parents and/or caretakers that are "judged by a mixture of community values and professional expertise to be inappropriate and damaging" (Garbarino and Gilliam, 1988:7). An important distinction among existing operational definitions of child maltreatment is whether they require endangerment of the child or demonstrable harm to the child. Zuravin illustrates the difference in the context of definitions of physical abuse.
Altemeier et al. (1984) provide an example of a demonstrable harm definition as one in which an abused child is one who has incurred injuries (bruises, abrasions, cuts, burns, fractures, bites, and loss of hair) as a result of parental actions. The Second National Incidence Study (NIS-2) of 1986 provides an example of an endangerment definition: "Child abuse is physical assault with or without a weapon by a parent or temporary caretaker. It includes hitting with a stick, strap, or other hard object, as well as scalding, burning, poisoning, suffocating, and drowning. It also includes slapping, spanking with hand, hitting with fist, biting, kicking, shoving, shaking, throwing, nonaccidental dropping, stabbing, and choking" (NCCAN, 1988b). This definition clearly includes acts that may not result in physical injury. Demonstrable harm may be a useful standard in legal settings, but for research purposes we agree with Zuravin (1991) that endangerment is the more appropriate criterion, since it places emphasis on the act itself rather than the uncertain consequence of the act.
Severity of Acts
Acts of maltreatment can differ markedly and vary with respect to severity and the relative likelihood of injury. For example, acts of similar severitya blow to the headcan result in significantly different injuries, such as a broken skull for a 6-month-old infant and a bruise for a 4-year-old. These developmental differences need to be considered in considering the severity of specific acts. The endangerment definition of Straus and Gelles (1986) is restricted to physical assaults. Although their definition of abuse and their cutoff level for abuse excludes less severe actions such as pushing and spanking that are included in the NIS-2 definition quoted above, their instrument (the Conflict Tactics Scales or CTS) records both mild and severe forms of physical attacks on a child and provides the capability of generating a variety of definitions at the analysis stage, depending on con-
text. Other indices of physical abuse, along the lines of the Conflict Tactics Scales (Straus, 1979, 1990a,b) might be developed from the answers to a sequence of questions concerning the occurrence of abusive acts, giving higher weight to more severe acts. Similar indices could be developed for other forms of maltreatment and used in various ways according to context. By analogy, the National Institute of Mental Health developed a widely used index of depression, the Center for Epidemiologic Studies Depression Scale (CES-D). A particular cut off value of CES-D is often used to operationalize the concept of clinical depression, but other transformations of the scale may be of interest if subclinical depression is under study. Existing scales need to be calibrated so that they yield comparable conclusions on false positives and negatives.
Frequency of Acts
In addition to severity, measures of child maltreatment need to take into account the frequency of acts. Definitions need to distinguish between "chronic behavioral patterns" and "infrequent explosive episodes" (Widom, 1988:263). Wolock and Horowitz (1977) required a recurrent pattern of minor injury to establish abuse, but a single episode if the injury is severe. Zuravin and Taylor (1987) employed similar approaches. Many definitions do not take account of frequency, and definitions that do so often differ with respect to the number of times an act must occur before a recurrent or chronic pattern is established (Zuravin, 1991). Since frequency is often dependent on self-reports, a single reported incident may in fact represent repeated assaults, including the first assault that caused recognized harm. A consensus strategy for defining chronicity would enhance the comparability of results across studies.
Class of Potential Perpetrators
Abuse and neglect definitions vary with respect to whether they are applied to acts committed by a wide or narrow range of perpetrators. In definitions of neglect, the perpetrator should generally be restricted to the child's parent or legal caretaker, although there are exceptions to this rulefor example, when a parent believes she or he is leaving a child with a responsible temporary caretaker, but the child turns out to be at risk or is actually harmed by that caretaker. With child abuse, no consensus appears to exist as to how broadly the set of potential perpetrators should be defined.
Intent to Harm and Culpability of the Perpetrator
Zuravin (1991) concludes that intention to harm, a criterion that has been used mainly for physical abuse, is not included in most recent research definitions of abuse or neglect. However, culpability is used as a qualifier for definitions of physical neglect in NIS-2 (NCCAN, 1988b). For example, hazardous housing conditions are not considered neglect when they are attributed to a lack of financial resources.
Several researchers have raised the link between a child's age or developmental stage when maltreated and long-term sequelae (Aber and Zigler, 1981; Cicchetti and Manly, 1990; McGee and Wolfe, 1991; Polansky et al., 1972). Age of child is clearly a key consideration in the development of valid operational criteria for child maltreatment. For example, leaving a 3-year old alone for 5 hours is clearly inappropriate, whereas leaving a 12-year old alone for the same period of time is not (Polansky et al., 1972). Aber and Zigler (1981) noted that "severe emotional damage due to separation and loss may be particularly easy to inflict during the child's earliest years." Clearly, successful measurement of maltreatment needs to be based on a developmental perspective of child rearing. Operationally, instruments developed to measure child maltreatment need to be carefully tuned to the age group and literacy level of the child. Documentation needs to be supplied to ensure that instruments are not administered to inappropriate age groups.
Culturally Informed Definitions
Although definitions of child abuse and neglect vary across time and across cultures, consensus exists around definitions of severe forms of child maltreatmentfor example, all members of society would probably agree that battering a child to death is morally repugnant. Less obviously substandard behaviors seem more subject to genuine cultural differences. The challenge of formulating culturally informed definitions of child maltreatment is to accommodate cultural variability in child care beliefs and practices while taking care not to promote different standards of care for children on the basis of race, ethnicity, or economic class.
We lack information on the cultural parameters of child maltreatment. Research must differentiate cultural norms from individual deviations from those norms in defining and identifying maltreatment. Child maltreatment is a highly charged issuea "diagnosis" of abusing or neglecting one's child constitutes a serious moral and value judgment not present in a strictly
medical diagnosis. Hence definitions need to take into account sociological and ethnographic aspects of the problem and balance the scientific desire for comparability and uniformity of definition with the need to be sensitive to differences in cultural viewpoints.
Consistent with the need to identify incidents of abuse that are related to culturally condoned practices, we need to encourage research that identifies the long- and short-term consequences of these incidents and their prevalence. The prevalence and effects of practices such as "coining" (a curing practice that involves the forceful pressing of coins on a child's body that results in bruises) among the Vietnamese cultures or punishment with tree switches, cords, or ropes among African Americans also need to be assessed. Although such practices may meet some criteria for abuse in the United States, it is critical for researchers who identify them to understand the context of those practices, at least in the perceptions of the parents, in order to influence behavioral changes in parenting.
Finally, vignette studies have indicated that various cultures judge the seriousness of child maltreatment incidents differently (Giovannoni and Becerra, 1979; Hong and Hong, 1991). These differences are not always in the expected directions. Although ethnic minorities are often overrepresented in official reports of maltreatment, blacks and Hispanics judged maltreatment vignettes as more serious than did whites (Giovannoni and Becerra, 1979).
Identification Of Child Maltreatment
Detection in Medical Settings
Spurred by pediatric advocacy, reporting laws adopted in the mid-1960s were narrowly focused on encouraging physicians to recognize and initiate protective action for children who were victims of physical abuse inflicted by their caretakers. Medical professionals primarily identified multiple bruises and fractures at different stages of healing resulting from abuse (Krugman, 1984). Over the years, the laws have broadened in scope to include child neglect, emotional injury, parental deprivation of medical care, and factors injurious to a child's mental development (Dubowitz and Newberger, 1989). Current reporting laws require health professionals to identify a wide array of physical conditions resulting from child maltreatment. The diagnosis of maltreatment by physicians is complicated by the absence of a universal medical definition.3
Research and technological developments have enabled health professionals to become increasingly sophisticated in their ability to detect and diagnose abuse. Several diagnostic indicators of child maltreatment have emerged through clinical experiences in medical settings, including: dis-
crepant, partial, or vague history; delay in seeking care; a family crisis; trigger behavior by the child; unrealistic expectations of the child by the parents; isolation of the family; and/or a history of the parent being abused as a child (Krugman, 1984).
Medical reports have described how to recognize specific patterns of immersion burns that result from intentional injuries (Purdue et al., 1988), and studies have determined the effects of the duration of heat exposure, water temperature, and thickness of skin (Moritz and Henriques, 1947). Medical research determining both the necessary force and resulting injury patterns of specific types of falls has aided medical professionals in assessing whether histories given by caretakers are consistent with a child's injury (Chadwick et al., 1991, Helfer et al., 1977; Nimityongskul and Anderson, 1987). Medical understanding of the ''shaken baby syndrome" has benefited from technological developments, including computerized topography and magnetic resonance imaging. Numerous medical consequences, particularly damage to the central nervous system, brain hemorrhages, skull fractures, and respiratory problems have been identified (Ludwig, 1983; Levitt, 1992). Research developments have also strengthened diagnosis of abuse in skin injuries (Wilson, 1977) and fractures (King et al., 1988). Likely locations for nonaccidental injuries have been identified, as well as injuries that would require children to have certain developmental motor skills in order to be self-inflicted (Ludwig, 1983; Johnson, 1990).
The medical literature describing physical findings associated with sexual abuse has also grown rapidly. Although physical findings alone are rarely conclusive in the absence of a history of sexual abuse or specific lab findings (American Academy of Pediatrics, 1991),4 recent studies have indicated that chafing, abrasions, or bruising of the inner thighs and genitalia, scarring in specific genital areas, and specific abnormalities of the hymen are "consistent with but not diagnostic of sexual abuse" (American Academy of Pediatrics, 1991:256). Documenting physical evidence of sexual abuse is complicated by variations in observations and descriptions of normal and abnormal genital appearance (Paradise, 1990).5 Accurate diagnosis of sexual abuse in children is hindered by frequent delays between alleged molestations, disclosure, and medical examinations. Currently, physicians retrospectively interpret changes in anogenital anatomy without the benefit of clinical research describing the healing chronology of acute genital and anal trauma (Finkel, 1989).
Medical research in the area of sexual abuse is constantly changing, and many early studies have not been validated (Finkel and DeJong, in press). For example, specific physical findings, previously thought to result from sexual abuse, are now questioned because recent studies have revealed similar findings in "nonabused" children (Krugman, 1990). The presence of certain sexually transmitted diseases as indicators of sexual abuse is also ques-
tioned since some sexually transmitted diseases, such as condylomata, chlamydia, and genital herpes, can be transmitted nonsexually (Ludwig, 1983; American Academy of Pediatrics, 1991). Without precise understanding of the mode of transmission of each sexually transmitted disease, physicians cannot rely simply on the presence of these diseases as definitive evidence of sexual abuse.
Diagnosing emotional abuse and physical and emotional neglect is also problematic for medical professionals. Diagnosis of these conditions is complicated by social factors such as poverty and differing cultural definitions of adequate care. In the absence of specific definitive physical manifestations, diagnoses often depend on the value judgments of individual physicians (Ludwig, 1983; Bross, 1982).6
Although the medical diagnosis of child abuse has improved (particularly in the area of physical abuse), most medical research in child maltreatment has been retrospective and clinical, consisting primarily of observations of patients at one institution. Additional empirical research on physical indicators of child maltreatment could improve diagnostic assessments of child maltreatment victims by physicians (Johnson, 1990). Applying research from other public health concerns, such as automotive safety technology, to child maltreatment may be helpful.
Reports to Child Protection Agencies
Aside from the question of how child maltreatment is defined, the interpretation of results must take into account aspects of the research design, and in particular the method by which subjects in the study are sampled. Ideally, incidence and prevalence would be determined by a population-based survey, in which a random sample of the general population is studied to determine patterns of maltreatment.7
Such population-based studies are expensive and difficult to implement. In contrast, studies may instead be based on the system of reports of child maltreatment to legal and social service agencies.8 Reports to child protective services or other agencies and their disposition are subject to biases from uncontrolled methods of detection, yielding problems analogous to those of clinic-based samples in medical studies. Comparatively mild forms of child maltreatment may not appear in such samples since they are never reported. Particular socioeconomic groups, such as the poor, are likely to be relatively overrepresented, since they often have more extensive contact with social service or other agencies liable to report maltreatment. Abusers who are capable of covering up their acts may go undetected, and certain ethnic groups may be overrepresented because of prejudice on the part of the reporters. Regional differences in maltreatment rates may reflect different patterns of reporting rather than true differences in underlying rates
for example, a region with good child protection agencies may paradoxically appear to have higher rates of abuse, since there is an incentive to report cases when services are available to address the problem. For all these reasons and more, data based on reports and their disposition need careful analysis of the effects of known or potential detection biases.
Despite their limitations, data on reports of child maltreatment and their disposition provide important descriptive information. The scientific quality of the information may be enhanced by more detailed studies of the detection and selection processes involved, perhaps using stratified random samples of case reports, as well as the addition of data on types, severity, and frequency of maltreatment from multiple sources both within child protective services and the family (Cicchetti and Barnett, 1991).
Improvements on a research-based classification of child maltreatment can greatly enhance the scientist's ability to correctly identify child maltreatment cases. Research efforts in a variety of disciplines are currently severely restrained by a lack of good diagnostic instruments. The development of standardized instruments will lead to more consistency in the identification of cases and strengthen the scientific quality of data obtained.
2-1. Research Recommendation: Recognizing that the absence of consistent research definitions seriously impedes the development of an integrated research base in child abuse and neglect, a series of expert multidisciplinary panels should be convened to review existing work and to develop a consensus on research definitions of each form of abuse and neglect.
If definitions of abuse and neglect continue to be influenced by fluctuations in community norms and shifting administrative and legislative requirements, the incidence of "occurrence" will remain ambiguous and variations will continue over time and between individuals, agencies, and communities. Research definitions derived from scientific criteria rather than legal classification systems can reduce subjective variations and can improve the quality of objective standards of measurement. Research definitions should be coordinated with case-report and legal definitions, be developmentally appropriate and culturally sensitive, provide clear inclusion and exclusion criteria, adopt unified subtyping schemes, and provide clear guidelines on issues of severity, duration, and frequency of acts of maltreatment. Definitions must be developed in consultation with existing agencies.
2-2. Research Recommendation: Sound clinical-diagnostic and research instruments for the measurement of child maltreatment are needed to operationalize the definitions discussed under Recommendation 1.
The clinically based instruments need to classify children into categories of maltreatment based on the clinical picture they present. The reliability and validity of these instruments must be established by sound testing with economically and culturally diverse populations. The generalizability of these instruments then needs to be tested on nonclinical populations to strengthen the instruments' abilities to classify children demonstrating similar behaviors who have not been reported in abusive incidents.
2-3. Research Recommendation: Research should be conducted on the detection processes that lead to the definition of cases identified in child protective services records and other social agencies that handle child maltreatment.
Better understanding of the case assessment, investigation, and substantiation processes will lead to more consensus, and the development of useful standardized instrumentation for case assessment. In addition, programs must be developed to train clinicians in appropriate techniques for obtaining abuse histories. Pilot studies of clinically based instruments to determine the nature, incidence, and prevalence of abuse experiences in children and adolescents could be conducted in medical and educational settings commonly frequented by children. Survey instruments that incorporate clinical findings must also be developed. Recognizing that improved instruments may lead to detection of previously unreported cases of abuse, ways must be devised to enable clinicians and other service providers to refer potentially abusive parents for direct assistance without requiring clear evidence of maltreatment prior to the delivery of services.
A word of caution must be added. Because of the hazards of erroneous identification, pilot screening studies must incorporate measures to protect families from the possible consequences of misdiagnosis and labeling.
2-4. Research Recommendation: Empirical research that builds on existing medical knowledge of the physical indicators of child sexual and physical abuse would assist physicians in the identification of child maltreatment. Such identification would also be facilitated by the development of training programs that integrate research findings from child maltreatment studies into the education of health professionals.
Research should focus specifically on children because, although research on burns, bruises, and trauma conducted with adults may be revealing, the results may not directly apply to children. Studies of indicators of physical abuse should specifically include studies with large populations and attention to physical presentation and healing (such as bruises) in different ethnic groups. Studies should also evaluate the relationship of bone density to fracturability in children. Research on diseases and conditions that mimic abusive head trauma or malnutrition, such as metabolic diseases
and genetic diseases, may also be revealing. Longitudinal studies of the genital anatomy of nonabused children as they progress through puberty as well as studies that document changes in the appearance of preexisting trauma (including the effect of hormonal influences during puberty) would be helpful to physicians in identifying victims of sexual abuse.
Such studies, however, may raise important ethical dilemmas, as discussed in Chapter 9. For example, what consideration should be given to the psychological impacts associated with studies of the physical characteristics of sexuality (such as hymenal anatomy) in young children? Is it appropriate to remind a child who has been sexually abused at a young age, or the child's parents, of this event annually?
1. There are many reasons why emotional maltreatment has not received more attention, including lack of agreement as to its many forms and the difficulty of operationalizing the definition once the form of emotional maltreatment has been agreed upon. Most forms of emotional maltreatment are very subtle and are usually overlooked, especially in the presence of physical abuse and neglect. Hart and Brassard (1987, 1990) and others argue that all children who are physically abused and neglected are also emotionally maltreated at least to some extent. However, the converse is not always true. Egeland and Erickson (1987) found in their high-risk sample a number of children who were emotionally maltreated but not physically abused or neglected. They were well fed, clothed, and received proper health care, but their caregivers did not respond to their emotional needs at an early age, and as a consequence they displayed severe forms of developmental maladaptation. Another reason it has not received proper attention is that emotional maltreatment leaves no physical marks, whereas physical abuse often leaves obvious physical signs. One definitional controversy centers on whether it is necessary for the parental action (or inaction) to have an apparent impact on the child. Some argue that any definition must include substantial observable impairment in the child's ability to perform and behave within a normal range. Others would argue that actions can be considered maltreatment regardless of their immediate observable impact on the child's functioning (see Schakel, 1987). Other definitional issues include intentionality and cultural context (Korbin, 1980). (See McGee and Wolfe  for a proposed model for a definition and criteria for emotional maltreatment.)
2. In a commentary on research with diverse populations, Brown et al. (1992) note that the selection of questions to be asked and instruments of assessment can all be subject to cultural influences. For example, parents and children may be asked to plan a family vacation or the child is presented with a story which involves the child breaking a mother's valued vase. How universal are these situations?
3. State laws mandating physician reporting define abuse and neglect with ambiguous terms such as "substantial, unjustified, and allowable" (Johnson, 1990). The American Medical Association's Diagnostic and Treatment Guidelines Concerning Child Abuse and Neglect (1985) have vague definitions of what physicians should report and suggest peer consultation in questionable cases.
4. Sexual abuse, particularly when there is no penetration, rarely results in physical trauma, and the elasticity of the internal genital anatomy also inhibits the detection of evidence of sexual or physical penetration (DeJong, 1992).
According to the American Academy of Pediatrics (1991), the presence of semen/sperm/acid phosphatase, a positive culture for gonorrhea, or a positive serologic test for syphilis are the only physical findings that make the diagnosis of sexual abuse a "medical certainty" (p. 257).
5. Even in clinical texts and dictionaries, definitions of "normal" genital anatomy in children differ. Physicians are also often unfamiliar with the genital anatomy of children. In one study, 77 percent of a surveyed group of physicians routinely examined genitalia 50 percent of the time; 17 percent of these physicians routinely examined genitalia less than 10 percent of the time (Finkel and DeJong, in press).
6. Neglect is classified by medical practitioners as either nonorganic failure to thrive, medical neglect, or abandonment (Ludwig, 1983). Failure to thrive is indicated by decelerating growth rates, poor hygiene, excessive oral stimulation, and developmental delays but medical distinctions between organic and nonorganic causes of failure to thrive are not clear (Drotar, 1992). Medical neglect results when parents deny manifestations of serious illnesses and refuse appropriate medicine or surgical treatment, or when caretakers fail to provide minimal well-child care (Mrazek and Mrazek, 1985). Diagnosing abandonment, which may include symptoms such as excessive dirty diapers, poor hygiene, excessive hunger, and dehydration as well as burns, ingestions, and repeated accidents, also presents difficulty for the health professional.
7. Rates from such a survey may be distorted by nonresponse and response errors such as arise when respondents do not accurately report their experiences, but they are not subject to systematic bias from the method of ascertainment.
8. Over the years, reporting laws have expanded in scope and in the categories of professionals and lay persons who are mandated to report maltreatment. Although the original proponents of these policies had as one of their purposes the creation of a repository of reports for statistical purposes, this purpose has receded into the background given the wide variation in the states' laws and the mechanisms to implement them (Giovannoni, 1992).
Aber, J.L., and E.F. Zigler
1981 Developmental considerations in the definition of child maltreatment. In D. Cicchetti and R. Rizley, eds., New Directions for Child Development. San Francisco: Jossey-Bass.
Altemeier, W., S. O'Connor, P. Vietze, H. Sandler, and K. Sherrod
1984 Prediction of child abuse: A prospective study of feasibility. Child Abuse and Neglect 8:393-400.
American Academy of Pediatrics
1991 Committee on Child Abuse and Neglect. Guidelines for the Evaluation of Sexual Abuse of Children. Pediatrics 87(2)(February):254-259.
American Medical Association
1985 Council report: AMA diagnostic and treatment guidelines concerning child abuse and neglect. Council on Scientific Affairs. Journal of the American Medical Association 254(6)(August 9):796-800.
Askikal H., and W. Webb, eds.
1978 Psychiatric Diagnosis. New York: Spectrum.
1982 Medical care neglect. Child Abuse and Neglect 6:375-381.
Brown, E., P. Martinez, and M. Radke-Yarrow
1992 Diversity: Research with diverse populations. Society for Research on Child Development Newsletter (Fall).
1946 Multiple fractures in the long bones of infants suffering from chronic subdural hematoma. American Journal of Roentgenology 56(August):163-173.
Chadwick, D., S. Chin, C. Slarno, J. Landsverk, and L. Kitchen
1991 Deaths from falls in children: How far is fatal? Journal of Trauma 31(10):1353-1355.
Christoffel, K.K., P.C. Scheidt, P.F. Agran, J.F. Kraus, E. McLoughlin, and J.A. Paulson
1992 Standard definitions for childhood injury research: Excerpts of a conference report. Pediatrics 89:1027-1034.
Cicchetti, D., and Barnett
1991 Toward the development of a scientific nosology of child maltreatment. Pp. 346-377 in D. Cicchetti and W. Grove, eds., Thinking Clearly About Psychology: Essays in Honor of Paul E. Meehl. Minneapolis: University of Minnesota Press.
Cicchetti, D., and V. Carlson, eds.
1989 Child Maltreatment: Theory and Research on the Causes and Consequences of Child Abuse and Neglect. New York: Cambridge University Press.
Cicchetti, D., and J.T. Manly
1990 Problems and solutions to conducting research in maltreating families: An auto-biographical perspective. In I.E. Sigel and G.H. Brody, eds. Methods of Family Research: Biographies of Research Projects, Vol. 2. Hillsdale, NJ: Erlbaum Press.
1992 Medical detection and effects of the sexual abuse of children. Pp. 71-99 in W. O'Donohue and J.H. Geer, eds., The Sexual Abuse of Children: Clinical Issues Volume 2. Hillsdale, NJ: Lawrence Erlbaum Press.
1992 Prevention of neglect and nonorganic failure to thrive. Pp. 115-149 in D.J. Willis, E.W. Holden, and M. Rosenberg, eds., Prevention of Child Maltreatment: Developmental and Ecological Perspectives. New York: Wiley.
1987 Diagnostic and Statistical Manual of Mental Disorders. 3rd edition revised. Washington, D.C.: American Psychiatric Association.
Dubowitz, H., and E. Newberger
1989 Pediatrics and child abuse. Pp. 76-94 in D. Cicchetti and V. Carlson, eds., Child Maltreatment: Theory and Research on the Causes and Consequences of Child Abuse and Neglect. New York: Cambridge University Press.
Eaton, W., and L. Kessler, eds.
1985 Epidemiologic Field Methods in Psychiatry. Orlando, FL: Academic Press.
Egeland, B., and M.F. Erickson
1987 Psychologically unavailable caregiving. Pp. 110-120 in M.R. Brassard, R. Germain, and S.N. Hart, eds., Psychological Maltreatment of Children and Youth. New York: Pergamon Press.
1989 Anogenital trauma in sexually abused children. Pediatrics 84(2)(August):317-322.
Finkel, M.A., and A.R. DeJong
in press Draft chapter. Medical Findings in Sexual Abuse: Genital Findings, Anal Findings, and Sexually Transmitted Diseases. In R.M. Reece, ed., Child Abuse Diagnosis and Treatment. Philadelphia: Lee and Febiger.
Garbarino, J., and G. Gilliam
1988 Understanding Abusive Families. Lexington, MA: Lexington Books.
Gelles, R.J., and M.A. Straus
1988 Intimate Violence. New York: Simon and Schuster.
1975 Unraveling child abuse. American Journal of Orthopsychiatry 45(3):346-356.
1971 Parental mistreatment: Perpetrators and victims. Journal of Marriage and the Family 33(November): 637-638.
1989 Definitional issues in child maltreatment. In D. Cicchetti and V. Carlson, eds, Child Maltreatment: Theory and Research on the Causes and Consequences of Child Abuse and Neglect. Cambridge: Cambridge University Press.
1992 Issues in the Definition of Child Maltreatment. Background paper prepared for the National Research Council's Panel on Research on Child Abuse and Neglect.
Giovannoni, J.M., and R.M. Becerra
1979 Defining Child Abuse. New York: Free Press.
Giovannoni, J.M., and A. Billingsley
1970 Child neglect among the poor: A study of parental adequacy in families of three ethnic groups. Child Welfare 49(April):196-204.
Goodwin, D., and S. Guze
1984 Psychiatic Diagnosis. Third edition. New York: Oxford University Press.
Green, A.H., R.W. Gaines, and A. Sandgrund
1974 Child abuse: Pathological syndrome of family interaction. American Journal of Psychiatry 131 (8): 882-886.
Hart, S.N., and M.R. Brassard
1987 A major threat to children's mental health: Psychological maltreatment. American Psychologist 43(2):160-165.
1990 Psychological maltreatment of children. Pp. 77-112 in R.T. Ammerman and M. Hersen, eds., Treatment of Family Violence: A Sourcebook. New York: John Wiley.
Helfer, R., T.L. Slovis, and M. Black
1977 Injuries resulting when small children fall out of bed. Pediatrics 60:533-35.
Hong, G.K., and L.K. Hong
1991 Comparative perspectives on child abuse and neglect: Chinese versus Hispanics and whites. Child Welfare 70(4):463-475.
Institute of Medicine
1990 Healthy People 2000: Citizens Chart the Course. Committee on Health Objectives for the Year 2000. Washington, DC: National Academy Press.
1990 Inflicted injury versus accidental injury. Pediatric Clinics of North America 37(4)(August):791-814.
Kempe, C.H., F.N. Silverman, B. Steele, W. Droegemueller, and H.R. Silver
1962 The battered child syndrome. Journal of the American Medical Association 181(1):17-24.
King, J., D. Diefendorf, J. Apthorp, V.F. Negrete, and M. Carlson
1988 Analysis of 429 fractures in 189 battered children. Journal of Pediatric Orthopedics 8(5):585-589.
1980 The cultural context of child abuse and neglect. Child Abuse and Neglect 4:3-13.
1984 Child abuse and neglect: The role of the primary care physician in recognition, treatment, and prevention. Primary Care 11(3)(September):527-534.
1990 Future role of the pediatrician in child abuse and neglect. Pediatric Clinics of North America 37(4):1003-1011.
1992 Shaken Baby Syndrome (Inflicted Cerebral Trauma). Draft statement in American Academy of Pediatrics, Committee on Child Abuse and Neglect.
1983 Child abuse. Pp. 1127-1163 in G. Fleisher and S. Ludwig, eds., Textbook of Pediatric Emergency Medicine. Baltimore: Williams and Wilkins.
1978 1977 Analysis of Child Abuse and Neglect Research. National Center on Child Abuse and Neglect, U.S, Children's Bureau, Administration for Children, Youth and Families, U.S. Department of Health, Education and Welfare.
McGee, R.A., and D.A. Wolfe
1991 Psychological maltreatment: Toward an operational definition. Development and Psychopathology 3:3-18.
1959/1973 Some ruminations on the validation of clinical procedures. Reprinted in P. Meehl, Psychodiagnosis. Minneapolis: University of Minnesota Press.
Moritz A.R., and F.C. Henriques
1947 Studies on thermal injuries: Pathology and pathogenesis of cutaneous burns experimental study. American Journal of Pathology 23:915-941.
Mrazek, D.A., and P.J. Mrazek
1985 Child maltreatment. Pp. 679-697 in M. Rutter and L. Hersor, eds., Child and Adolescent Psychiatry: Modern Approaches 2nd Edition. London: Blackwell Scientific Publications.
National Center for Child Abuse and Neglect
1981 National Study of the Incidence and Severity of Child Abuse and Neglect. 81-30325. Washington, DC: U.S. Department of Health and Human Services. [NIS 1]
1988a Report on Data Collection: Study of National Incidence and Prevalence of Child Abuse and Neglect. Washington, DC: U.S. Department of Health and Human Services.
1988b Study Findings: Study of National Incidence and Prevalence of Child Abuse and Neglect. Washington, D.C.: U.S. Department of Health and Human Services. [NIS 2]
Nimityongskul A., and A. Anderson
1987 Likelihood of injuries when children fall out of bed. Journal of Pediatric Orthopedics 7(2):184-186.
Office of Technology Assessment
1987 Healthy Children: Investing in the Future. Washington, DC: U.S. Congress, Office of Technology Assessment.
1990 The medical evaluation of the sexually abused child. Pediatric Clinics of North America 37(4):839-862.
1978 Child abuse and neglect: The myth of classlessness. American Journal of Orthopsychiatry 48(4)(October):608-617.
Polansky, N., R. Borgman, and C. DeSaix.
1972 Roots of Futility. San Francisco: Jossey Bass.
Purdue, G., J. Hunt, and P. Prescott.
1988 Child abuse by burning: An index of suspicion. The Journal of Trauma 28(2):221-224.
Ross, C.J., and E. Zigler
1980 An agenda for action. Pp. 293-305 in G. Gerbner, C.J. Ross, and E. Zigler, eds., Child Abuse: An Agenda for Action. New York: Oxford University Press.
1987 Emotional neglect and stimulus deprivation. Pp. 100-109 in M.R. Brassard, R. Germain, and S.N. Hart, eds., Psychological Maltreatment of Children and Youth. New York: Pergamon Press.
1979 Measuring intrafamily conflict and violence: The Conflict Tactics Scales. Journal of Marriage and the Family 48:75-88.
1990a The Conflict Tactics Scales and its critics: An evaluation and new data on validity and reliability. In M.A. Straus and R.J. Gelles, eds., Physical Violence in American Families: Risk Factors and Adaptations to Violence in 8145 Families. New Brunswick, NJ: Transaction Publications.
1990b Measuring physical and psychological maltreatment of children with the Conflict Tactics Scales. Family Research Labratory, University of New Hampshire.
Straus, M.A., and R.J. Gelles
1986 Societal change in family violence from 1975 to 1985 as revealed by two national surveys. Journal of Marriage and the Family 48(August):465-480.
1977 Juvenile Justice Standards Project. In M. Wald, Standards Relating to Abuse and Neglect. Cambridge, MA: Ballinger.
1988 Sampling biases and implications for child abuse research. American Orthopsychiatric Association 58(2):260-270.
1977 Estimation of age of cutaneous contusions in child abuse. Pediatrics 60:750-752.
Wolock I., and B. Horowitz
1977 Factors Relating to Levels of Child Care Among Families Receiving Public Assistance in New Jersey. Final Report to the National Center on Child Abuse and Neglect. (DHEW Grant 90-C-418.) Washington, DC: Clearinghouse on Child Abuse and Neglect Information.
1984 Child maltreatment as a social problem: The neglect of neglect. American Journal of Orthopsychiatry 54(4)(October):530-543.
Wyatt, G.E., and S.D. Peters
1986 Issues in the definition of child sexual abuse in prevalence research. Child Abuse and Neglect 10:231-240.
Zigler, E.F., and L. Phillips
1981 Psychiatric diagnosis. Journal of Abnormal and Social Psychology 63:607-18.
1991 Research definitions of child physical abuse and neglect: Current problems. In R. Starr and D. Wolfe, eds., The Effects of Child Abuse and Neglect: Issues and Research. New York: Guilford Press.
Zuravin, S., and R. Taylor
1987 Family Planning Behaviors and Child Care Adequacy. Final report submitted to the U.S. Department of Health and Human Services, Office of Population Affairs (Grant FPR 000028-01-1).