Determining when maltreatment has occurred continues to be challenging, as it can result from acts of commission (abuse) or omission (neglect) and can be acute or chronic, subtle or extreme (see Brassard and Hardy, 1997). Additional confusion has arisen in the field over defining abuse and neglect according to the perpetrator’s behavior or injury to the victim or whether a single act or a pattern of repeated actions is required. For example, the original term “battered woman syndrome” implies that a pattern is required, but the more recent approach when collecting data is to obtain information on the type, severity, and frequency of the violence.
The confusion is exacerbated by the development of multiple definitions of maltreatment developed for different purposes (e.g., research, judicial action, clinical investigation; NRC, 1993). Some studies rely on reported cases and others estimate cases. The definition of a case varies from state to state, with some requiring that there be “reason to believe” abuse has occurred before “substantiating” abuse or neglect, others requiring that abuse “probably occurred,” and still others requiring “clear and convincing” evidence. Interstate variability exists in the number of cases reported per population at risk. The heterogeneity of definitions and evidentiary requirements makes accuracy in incidence data extremely difficult to achieve.
Investigative methods and data sources also affect research results. For example, rates tend to be very low if only severe physical injury (e.g., fracture) is included, somewhat higher if milder injury (e.g., bruises) is included, and can increase even more if psychological violence and emotional violence are included (Wilt and Olson, 1996). Surveys of patients in health care settings tend to find higher rates than do surveys of the general population (Campbell et al., 2000; Wilt and Olson, 1996). Self-administered written questionnaires tend to result in higher rates than do personal interviews (Canterino et al., 1999; Thompson et al., 2000), although this is not always the case (Gazmararian et al., 1996). Higher rates tend to be obtained when victim and perpetrator are assessed separately (Bohannon et al., 1995). Rates also depend on how the survey is framed or introduced (Campbell et al., 2000).
All of this diversity can result in tremendous variation in research results and difficulty in comparing and interpreting data. Despite these constraints, the data do generally indicate the extent of the problem. In the following sections, the committee describes the definitions of types of family violence that informed its analysis in this report and what is known about the magnitude of each type. Table 2.1 presents the definitions of family violence terminology used in the report. Table 2.2 presents common elements of the three types of maltreatment.
The current understanding of child abuse and neglect has expanded greatly since Kempe and colleagues in 1962 first coined the term battered child syn-