This report presents the results of the congressionally mandated Third National Incidence Study of Child Abuse and Neglect (NIS-3). The NIS is the single most comprehensive source of information about the current incidence of child abuse and neglect in the United States. The NIS-3 findings are based on a nationally representative sample of over 5,600 professionals in 842 agencies serving 42 counties. The study used two sets of standardized definitions of abuse and neglect. Under the Harm Standard, children identified to the study were considered to be maltreated only if they had already experienced harm from abuse or neglect.

Under the Endangerment Standard, children who experienced abuse or neglect that put them at risk of harm were included in the set of those considered to be maltreated, together with the already-harmed children.

The NIS-3 provides us with important insights about the incidence and distribution of child abuse and neglect and about changes in incidence since the previous studies.

Incidence

 
• There have been substantial and significant increases in the incidence of child abuse and neglect since the last national incidence study was conducted in 1986.  
• Under the Harm Standard definitions, the total number of abused and neglected children was two-thirds higher in the NIS-3 than in the NIS-2. This means that a child’s risk of experiencing harm-causing abuse or neglect in 1993 was one and one-half times the child’s risk in 1986.  
• Under the Endangerment Standard, the number of abused and neglected children nearly doubled from 1986 to 1993. Physical abuse nearly doubled, sexual abuse more than doubled, and emotional abuse, physical neglect, and emotional neglect were all more than two and one-half times their NIS-2 levels.  
• The total number of children seriously injured and the total number endangered both quadrupled during this time.

Child Characteristics

 
• Girls were sexually abused three times more often than boys.  
• Boys had a greater risk of emotional neglect and of serious injury than girls.  
• Children are consistently vulnerable to sexual abuse from age three on.  
• There were no significant race differences in the incidence of maltreatment or maltreatment-related injuries uncovered in either the NIS-2 or the NIS-3.

Family Characteristics

 
• Children of single parents had a 77% greater risk of being harmed by physical abuse, an 87% greater risk of being harmed by physical neglect, and an 80% greater risk of suffering serious injury or harm from abuse or neglect than children living with both parents.  
• Children in the largest families were physically neglected at nearly three times the rate of those who came from single-child families.  
• Children from families with annual incomes below $15,000 as compared to children from families with annual incomes above $30,000 per year were over 22 times more likely to experience some form of maltreatment that fit the Harm Standard and over 25 times more likely to suffer some form of maltreatment as defined by the Endangerment Standard.  
• Children from the lowest income families were 18 times more likely to be sexually abused, almost 56 times more likely to be educationally neglected, and over 22 times more likely to be seriously injured from maltreatment as defined under the Harm Standard than children from the higher income families.

Child Protective Services (CPS) Investigation

 
• CPS investigated only 28% of the recognized children who met the Harm Standard. This was a significant decrease from the 44% investigated in 1986.  
• Although the percentage of children whose abuse or neglect was investigated declined, the actual number of children investigated remained constant.  
• CPS investigated less than one-half of all Harm Standard children recognized by any source and less than one-half of all Endangerment Standard children recognized by any source except police and sheriffs’ departments (52%).  
• Schools recognized the largest number of children maltreated under the Harm Standard, but only 16% of these children were investigated by CPS.  
• CPS investigated only 26% of the seriously injured and 26% of the moderately injured children.

This study would not have been possible without the support of hundreds of agencies and individual caseworkers, teachers, police officers, social workers, probation officers, nurses, and other professionals in the study counties who contributed their enthusiastic support and much of their time in the effort to assess accurately the incidence, nature, and distribution of child abuse and neglect in the United States. I extend my appreciation to these dedicated respondents.

This report summarizes the Third National Incidence Study of Child Abuse and Neglect (NIS-3). It gives a synopsis of the study’s background and objectives, its design and methods, and its key findings and implications.

Background and Objectives

The National Incidence Study (NIS) is a congressionally mandated, periodic effort of the National Center on Child Abuse and Neglect (NCCAN). The first NIS (NIS-1), mandated under P.L. 93-247 (1974), was conducted in 1979 and 1980 and published in 1981. The second NIS (NIS-2), mandated under P.L. 98-457 (1984), was conducted in 1986 and 1987 and published in 1988. The third NIS (NIS-3) was mandated under P.L. 100-294 (as amended). The NIS-3 data were collected in 1993 and 1994, analyses conducted in 1995 and 1996, and these results published in 1996. A key objective of the NIS-3 was to provide updated estimates of the incidence of child abuse and neglect in the United States and measure changes in incidence from the earlier studies.

Design and Methods

The NIS-3 offers an important perspective on the scope of child abuse and neglect. The NIS includes children who were investigated by child protective service (CPS) agencies, but it also obtains data on children seen by community professionals who were not reported to CPS or who were screened out by CPS without investigation. This means that the NIS estimates provide a more comprehensive measure of the scope of child abuse and neglect known to community professionals, including both abused and neglected children who are in the official statistics and those who are not. The NIS follows a nationally representative design, which means that the estimates represent the numbers of abused and neglected children in the United States who come to the attention of community professionals. The fact that there have been three similar national incidence studies that have used comparable methods and definitions means that one can compare NIS-3 estimates with those from the earlier studies in order to identify any changes over time in the incidence and distribution of abused and neglected children.

The NIS-3 was conducted in a nationally representative sample of 42 counties. In every county, the CPS agency was a key participant, providing basic demographic data on all the children who were reported and accepted for investigation during the 3-month study data period, September 5 through December 4, 1993. Further details about the child’s maltreatment and the outcome of the CPS investigation were obtained for a representative sample of these cases.

Like the NIS-1 and NIS-2 before it, the NIS-3 employed a sentinel survey methodology, in which community professionals serving children and families in various categories of non-CPS agencies were also recruited into the study. In each county, these sentinels were a representative sample of all professional staff who were likely to come into contact with maltreated children in police and sheriffs’ departments, public schools, day-care centers, hospitals, voluntary social service agencies, mental health agencies, and the county juvenile probation and public health departments. The participating sentinels in the NIS-3 were 5,612 professionals in 800 non-CPS agencies who remained on the lookout for maltreated children during the study period. They were trained in the standard NIS definitions of abuse and neglect at the outset, and they submitted data forms on any children they encountered who were maltreated during the study data period. The NIS-3 collected a total of 50,729 data forms: 4,711 from non-CPS sentinels; 3,154 on the investigation outcomes and the abuse and neglect involved in cases sampled at participating CPS agencies; and 42,864 capturing the basic demographic data on all cases reported to participating CPS agencies during the study period.

Children who were submitted to the study by non-CPS sentinels and those who were investigated in the CPS sampled cases were evaluated according to standard study definitions of abuse and neglect, and only children who fit the standards were used in generating the national estimates. The definitional standards used in the NIS-3 were identical to those used in the NIS-2. These standards imposed a number of requirements, including the restriction that the abuse or neglect be within the jurisdiction of CPS (i.e., perpetrated or permitted by a parent or caretaker), and they applied uniform classification systems to index the type of maltreatment and the severity and type of injury or harm.

Two sets of definitional standards were applied: the Harm Standard and the Endangerment Standard. The Harm Standard was developed for the NIS-1, and it has been used in all three national incidence studies. It is relatively stringent in that it generally requires that an act or omission result in demonstrable harm in order to be classified as abuse or neglect. Exceptions are made in only a few categories where the nature of the maltreatment itself is so egregious that the standard permits harm to be inferred when direct evidence of it is not available. The chief advantage of the Harm Standard is that it is strongly objective in character. Its principal disadvantage is that it is so stringent that it provides a view of abuse and neglect that is too narrow for many purposes, excluding even many children whose maltreatment is substantiated or indicated as abuse or neglect by CPS.

To meet the need to include the full set of substantiated/indicated children in the incidence statistics, the Endangerment Standard was developed as a definitional standard during the NIS-2 to supplement the perspective provided by the Harm Standard. The Endangerment Standard includes all children who meet the Harm Standard but adds others as well. The central feature of the Endangerment Standard is that it allows children who were not yet harmed by maltreatment to be counted in the abused and neglected estimates if a non-CPS sentinel considered them to be endangered by maltreatment or if their maltreatment was substantiated or indicated in a CPS investigation. In addition, the Endangerment Standard is slightly more lenient than the Harm Standard concerning the identity of allowable perpetrators in that it includes maltreatment by adult caretakers other than parents in certain categories as well as sexual abuse perpetrated by teenage caretakers. The Endangerment Standard was used in both the NIS-2 and the NIS-3.

Duplicate forms about the same child were identified and unduplicated, so that each child was included in the database only once. Finally, the data were weighted to represent the total number of children maltreated in the United States and annualized to transform the information from the 3-month data period into estimates reflecting a full year.

The National Incidence of Child Abuse and Neglect

The findings of the Third National Incidence Study of Child Abuse and Neglect (NIS-3) show a sharp increase in the scope of the problem, whether maltreatment is defined using the Harm Standard or the Endangerment Standard.

Estimated Incidence As Defined by the Harm Standard. An estimated 1,553,800 children in the United States were abused or neglected under the Harm Standard in 1993. The NIS-3 total reflects a 67% increase since the NIS-2 estimate, which indicated that the total was 93 1,000 children in 1986, and it corresponds to a 149% increase since the NIS-1 estimate for 1980 of 625,100 children. Significant or close-to-significant increases were found in both abuse and neglect. The number of abused children who were countable under the Harm Standard rose by 46% from an estimated 507,700 in the NIS-2 to 743,200 in the NIS-3. The number of neglected children who fit the Harm Standard increased significantly from 474,800 during the NIS-2 data collection in 1986 to 879,000 at the time of the NIS-3 data period in 1993. In the estimates given here and below, children are included in all categories that apply to them (i.e., those who were both abused and neglected are included in both estimates).

Considering specific types of abuse and neglect as defined by the Harm Standard, significant increases since the NIS-2 were found in the incidence of sexual abuse, physical neglect, and emotional neglect, and a close-to-significant (i.e., statistically marginal) increase was observed in the incidence of physical abuse:

 
• The estimated number of sexually abused children under the Harm Standard rose from II 9,200 in 1986 to 217,700 in 1993 (an 83% increase);  
• The number of physically neglected children under the Harm Standard increased from an estimated 167,800 at the time of the NIS-2 to an estimated 338,900 in the NIS-3 (a 102% rise in incidence);  
• There was a 333% increase in the estimated number of emotionally neglected children using the Harm Standard, from 49,200 in the NIS-2 to 212,800 in the NIS-3; and  
• The estimated number of physically abused children under the Harm Standard was 269,700 at the time of the NIS-2, but it had increased to 381,700 during the NIS-3 (a 42% increase).

When these abused and neglected children were classified according to the injury or harm they suffered from maltreatment that fit the Harm Standard, there was a substantial and significant increase in the incidence of children who were seriously harmed and a statistically marginal increase in the number for whom injury could be infeffed due to the severe nature of their maltreatment. The estimated number of seriously injured children essentially quadrupled from 141,700 to 565,000 in the intervening 7 years between the NIS-2 and the NIS-3 (a 299% increase). The number for whom injury could be infeffed increased from an estimated 105,500 children in the NIS-2 to an estimated 165,300 children in the NIS-3 (a 57% increase).

Estimated Incidence Using the Endangerment Standard. Between 1986 and 1993, the total estimated number of abused and neglected children in the United States who fit the Endangerment Standard nearly doubled: in 1986, there were an estimated 1,424,400 abused and neglected children in the United States. The NIS-3 estimate of 2,815,600 reflects a 98% increase over the NIS-2 figure. Significant increases were found in both abuse and neglect. The number of abused children more than doubled from an estimated 590,800 to 1,221,800 (a 107% increase), while the estimated number of neglected children also more than doubled from 917,200 to 1,961,300 (a 114% increase).

 
• The increases were substantial and significant in all types of abuse and neglect except educational neglect;  
• The estimated number of physically abused children rose from 311,500 to 614,100 (a 97% increase);  
• The estimated number of sexually abused children increased from an estimated 133,600 children to 300,200 (a 125% increase);  
• The more recent estimate of the number of emotionally abused children was 183% higher than the previous estimate (188,100 in 1986 versus 532,200 in 1993);  
• The estimated number of physically neglected children increased from 507,700 to 1,335, 100 (a 163% increase); and  
• The estimated number of emotionally neglected children nearly tripled in the interval between the studies, rising from 203,000 in 1986 to 585,100 in 1993 (a 188% increase).

When the children whose abuse or neglect met the Endangerment Standard were classified according to the injury or harm they suffered, significant increases were evident in two categories. First, the 1993 estimate of the number of children who were endangered by their maltreatment (but not yet harmed) was more than four times the corresponding 1986 estimate. That is, the number of endangered children rose from an estimated 254,000 in 1986 to an estimated 1,032,000 in 1993 (a 306% increase). Second, the number of children who were seriously injured or harmed by abuse or neglect that fit the Endangerment Standard in 1993 was well over one-half million, which is nearly quadruple the 1986 estimate for this category. In 1986, an estimated 143,300 children had been seriously injured by abuse or neglect; in 1993, the figure was 569,900 children (a 298% increase). Note that nearly all (99%) of the children who counted as seriously injured here were also countable under the Harm Standard, so the near-quadrupling of their numbers since 1986 essentially reiterates what was reported above in connection with the Harm Standard.

Distribution of Child Abuse and Neglect by the Child’s Characteristics

The child’s sex and age were related to the rate of maltreatment, but race was not.

Child’s Sex. Girls were sexually abused about three times more often than boys, under both the Harm Standard and the Endangerment Standard. This finding reiterates the NIS-2 result, so females’ disproportionately greater risk of sexual abuse has been stable over time. This sex difference in incidence rates of sexual abuse leads to higher rates of abuse in general among girls. Also, because the definitional guidelines permit the inference that injury or harm occurred in connection with the more extreme forms of sexual abuse, girls’ greater risk of sexual abuse also accounts for their higher incidence rates for inferred injury.

At the same time, boys had higher incidence rates than girls in some arenas, and boys’ maltreatment risks also demonstrated some increases since the NIS-2. Boys were at somewhat greater risk of serious injury (24% higher than girls’ risk under both definitional standards), and boys were significantly more likely to be emotionally neglected (boys’ risk was 18% greater than girls’). Also, boys’ rates of physical neglect defined by the Harm Standard and of emotional abuse using the Endangerment Standard increased more since the NIS-2 than girls’ rates did. Moreover, trends in the incidence of fatal injuries from maltreatment moved in opposite directions for girls and boys-the incidence of fatally injured girls declined slightly since the NIS-2, while the incidence of fatally injured boys rose.

Child’s Age. A consistent feature of the age differences in incidence rates within the NIS-3 was the lower incidence of maltreatment among the younger children under both definitional standards. In most cases, the differentiation was between the 0- to 2-year-olds and older children or between the 0- to 5-year-olds and older children. It is possible that the lower rates at these younger ages reflect undercoverage of these age groups. That is, prior to attaining school age, children are less observable to community professionals.

Another recurring theme in connection with age is that of disproportionate increases in the incidence of maltreatment among the younger children (under 12 years old) and especially among children in their middle-cbildhood years (ages 6 to I 1). Note that as circumstances deteriorate and maltreatment becomes more prevalent and more severe, older children have greater opportunities for escape. Also, older children are more able to defend themselves andIor retaliate. These factors may have moderated the increases in maltreatment that were observed among the older age groups.

The disproportionate increases during the younger and middle-childhood years mean that the overall profiles of age differences in maltreatment were different in the NIS-3 than they had been in the NIS-2. During the NIS-2, the risk of maltreatment generally increased with the age of the child in a close-to-linear fashion. With the lopsided increases among the younger children and among children in their middle-childhood years, the profile has changed toward a curvilinear configuration-where the middle-years of childhood are associated with the maximum risk of maltreatment-and toward a somewhat flatter distribution-where age differences are somewhat attenuated overall compared to their NIS-2 patterns.

One of the most striking findings is the age distribution of sexual abuse, which combined the general flattening of the age differences in incidence rates with a very low age transition in the distribution of incidence rates. The rate of sexual abuse as defined under the Endangerment Standard was very low for 0- to 2-year-olds, but then relatively constant for children ages 3 and older, indicating a very broad age range of vulnerability from preschool age on.

Race. The NIS-3 found no race differences in maltreatment incidence. The NIS-3 reiterates the findings of the earlier national incidence studies in this regard. That is, the NIS-1 and the NIS-2 also found no significant race differences in the incidence of maltreatment or maltreatment-related injuries.

Service providers may find these results somewhat surprising in view of the disproportionate representation of children of color in the child welfare population and in the clientele of other public agencies. However, it should be recognized that the NIS methodology identifies a much broader range of children than those who come to the attention of any one type of service agency or the even smaller subset who receive child protective and other child welfare services. The NIS findings suggest that the different races receive differential attention somewhere during the process of referral, investigation, and service allocation, and that the differential representation of minorities in the child welfare population does not derive from inherent differences in the rates at which they are abused or neglected. It is also important to recognize that while there are no overall race differences in the incidence of child abuse and neglect in the NIS-3 findings, subsequent analyses that simultaneously consider multiple characteristics may reveal race differences in maltreatment incidence among specific subsets of children (e.g., for children of certain ages, for one sex but not the other, etc.).

Distribution of Child Abuse and Neglect by Family Characteristics

The incidence of child maltreatment varied as a function of family income, family structure, family size, and the metropolitan status of the county.

Family Structure. Children of single parents were at higher risk of physical abuse and of all types of neglect and were overrepresented among seriously injured, moderately injured, and endangered children. Compared with their counterparts living with both parents, children in singleparent families had:

 
• a 77% greater risk of being harmed by physical abuse (using the stringent Harm Standard) and a 63% greater risk of experiencing any countable physical abuse (using the Endangerment Standard);  
• an 87% greater risk of being harmed by physical neglect and a 165% greater risk of experiencing any countable physical neglect;  
• a 74% greater risk of being harmed by emotional neglect and a 64% greater risk of experiencing any countable emotional neglect;  
• a 220% (or more than three times) greater risk of being educationally neglected;  
• an approximately 80% greater risk of suffering serious injury or harm from abuse or neglect;  
• an approximately 90% greater risk of receiving moderate injury or harm as a result of child maltreatment; and  
• a 120% (or more than two times) greater risk of being endangered by some type of child abuse or neglect.

Among children in single-parent households, those living with only their fathers were approximately one and two-thirds times more likely to be physically abused than those living with only their mothers.

Although parents are not necessarily, nor even most frequently, the perpetrators of maltreatment, the relationship between parent structure and maltreatment incidence is understandable, considering the added responsibilities and stresses of single-parenting together with the likelihood that surrounding social and practical support may be inadequate.

Family Size. The incidence of maltreatment was related to the number of dependent children in the family, especially in the categories of physical and educational neglect. For educational neglect, and for physical neglect according to the Harm Standard, the pattern was nonlinear: the incidence rates were highest for children in the largest families (those with four or more children), intermediate for “only” children, and lowest for children in families with two to three children. Children in the largest families were almost three times more likely to be educationally neglected, and nearly two and two-fifths times more likely to be physically neglected under the Harm Standard, compared to children in families with two or three children. Under the Endangerment Standard, the pattern was one of increasing incidence of physical neglect with greater numbers of children. Children in the largest families were physically neglected at nearly three times the rate of those who came from “only” child families.

Additional children in a household mean additional tasks and responsibilities, so it is understandable why incidence rates of child abuse and neglect may be higher when there are more children. Accounting for why “only” children have higher rates of educational neglect and of physical neglect under the Harm Standard than children in families with two or three children requires a different explanation. One possibility is that there may be too many expectations focused on “only” children, whereas expectations (and disappointments) are diffused over multiple children in the larger families. Another possibility is that many “only” child households represent the early stages in their families’ development, since a number of these families will have additional children, in time. Thus, many “only” children are in families with relatively young and inexperienced parents and caretakers.

County Metropolitan Status. The incidence of children who had been moderately harmed by maltreatment was significantly lower among children in large urban counties than among children who lived in other urban counties. This was interpreted as reflecting a general undercoverage of moderately injured maltreated children in the large urban counties. It was not clear whether this was because the moderately injured children are less likely to be encountered by community professionals in the large urban centers, because community professionals in these locales are less likely to identify these children as maltreated, or because the NIS information sources in these counties are less likely to submit data about these maltreated children.

Family Income. Despite the fact that only a rather gross index of family income was available, and despite a substantial percentage of cases with missing data on this factor, family income was significantly related to incidence rates in nearly every category of maltreatment. Compared to children whose families earned $30,000 per year or more, those in families with annual incomes below $15,000 per year were:

 
• more than 22 times more likely to experience some form of maltreatment under the Harm Standard and over 25 times more likely to suffer maltreatment of some type using the Endangerment Standard;  
• almost 14 times more likely to be harmed by some variety of abuse and nearly 15 times more likely to be abused using the Endangerment Standard criteria;  
• more than 44 times more likely to be neglected, by either definitional standard;  
• almost 16 times more likely to be a victim of physical abuse under the Harm Standard and nearly 12 times more likely to be a victim of physical abuse using the Endangerment Standard;  
• almost 18 times more likely to be sexually abused by either definitional standard;  
• thirteen times more likely to be emotionally abused under the Harm Standard criteria and more than 18 times more likely to be emotionally abused in a manner that fit Endangerment Standard requirements;  
• forty times more likely to experience physical neglect under the Harm Standard and over 48 times more likely to be a victim of physical neglect using the Endangerment Standard;  
• over 29 times more likely to be emotionally neglected under the Harm Standard definitions and over 27 times more likely to be emotionally neglected by Endangerment Standard criteria;  
• nearly 56 times more likely to be educationally neglected, by either definitional standard;  
• sixty times more likely to die from maltreatment of some type under the Harm Standard and over 22 times more likely to die from abuse or neglect using the Endangerment Standard;  
• over 22 times more likely to be seriously injured by maltreatment under the Harm Standard and almost 22 times more likely to be seriously injured by maltreatment that fit the Endangerment Standard requirements;  
• about 18 times more likely to be moderately injured by abuse or neglect under the Harm Standard and nearly 20 times more likely to have a moderate injury from maltreatment as defined by the Endangerment Standard;  
• fifty-seven times more likely to be classified as having an inferred injury under the Harm Standard and 39 times more likely to meet the criteria for inferred injury as defined by the Endangerment Standard; and  
• over 31 times more likely to be considered endangered, although not yet injured, by some type of abusive or neglectful treatment.

The NIS-3 findings on the correlation between family income and child maltreatment are entirely consistent with the earlier findings of the NIS-2. Moreover, they cannot be plausibly explained on the basis of the higher visibility of lower-income families to community professionals.

On the one hand, the NIS sentinels observe substantial numbers of children and families at the middle- and upper-income levels. The large majority of maltreated children were recognized by professionals likely to encounter children and fan-filies at all Income levels, such as sentinels in hospitals, schools, day-care centers, mental health agencies, voluntary social service agencies; by professionals not represented by NIS sentinel categories; and by the general public. Sentinels in schools alone recognized the majority of the maltreated children. Although the NIS design includes only public schools, approximately 89% of the U.S. population of school-age children attend public schools, so children attending the public schools represent a broad spectrum of family income levels. Moreover, the private schools not reflected in the NIS include religiously affiliated schools, which have sliding scales for poorer children, so children who attend private schools are not necessarily from better economic circumstances than children enrolled in public schools.

On the other hand, if the income finding is interpreted as an artifact of selective observation of low-income families, then it would mean that there have to be enough undetected abused and neglected children in the middle- and upper-income brackets used here to equalize the incidence rates across different income categories. That would require an astounding number of still-undetected children in the nation who experience countable maltreatment. Specifically, it would mean that an additional 2,138,700 children suffered maltreatment according to the Harm Standard yet remained hidden to the NIS. Similarly, it would mean there were an additional 4,500,700 children in 1993 who experienced maltreatment under the Endangerment Standard but who escaped observation by community professionals. To add some perspective as to what this would entail, consider that almost seven% of the total U.S. child population would be maltreated in countable ways yet entirely escape the attention of the spectrum of community professionals who serve as NIS sentinels, and all of these additional children would have to be in families with incomes of $15,000 per year or more.

Considering the implications of the alternative, it appears more plausible to assume that the income-related differences in incidence found in the NIS reflect real differences in the extent to which children in different income levels are being abused or neglected. Note that. there are a number of problems associated with poverty that may contribute to child maltreatment: more transient residence, poorer education, and higher rates of substance abuse and emotional disorders. Moreover, families at the lower socioecononu’c levels have less adequate social support systems to assist parents in their child care responsibilities.

Distribution of Child Abuse and Neglect by Perpetrator Characteristics

Children who had been maltreated as defined by the Harm Standard were categorized according to their relationship to the most closely related perpetrator and according to this perpetrator’s sex, age, and employment status; these categorizations were examined in relation to the type of maltreatment and the severity of the child’s injury or harm. Perpetrators’ relationships to the children also were examined in relation to the children’s race. The findings represent only a preliminary exploration of perpetrator characteristics in the NIS-3 data, since they lack significance tests concerning potential relationships and substantial percentages of the children were missing information concerning certain of the perpetrator characteristics.

Perpetrator’s Relationship to the Child. The majority of all children countable under the Harm Standard (78%) were maltreated by their birth parents, and this held true both for children who were abused (62% were maltreated by birth parents) and for those who were neglected (91 % experienced neglect by birth parents).

Birth parents were the most closely related perpetrators for 72% of the physically abused children and for 81% of the emotionally abused children. The pattern was dis tinctly different for sexual abuse. Nearly one-half of the sexually abused children were sexually abused by someone other than a parent or parent-substitute, while just over one-fourth were sexually abused by a birth parent, and one-fourth were sexually abused by other than a birth parent or parent-substitute. In addition, a sexually abused child was most likely to sustain a serious injury or impairment when a birth parent was the perpetrator.

Perpetrator’s Sex. Children were somewhat more likely to be maltreated by female perpetrators than by males: 65% of the maltreated children had been maltreated by a female, whereas 54% had been maltreated by a male. Of children who were maltreated by their birth parents, the majority (75%) were maltreated by their mothers and a sizable minority (46%) were maltreated by their fathers (some children were maltreated by both parents). In contrast, children who were maltreated by other parents or parent-substitutes, or by other persons, were more likely to have been maltreated by a male than by a female (80 to 85% were maltreated by males; 14 to 41 % by females).

Abused children presented a different pattern in connection with the sex of their perpetrators than did the neglected children. Children were more often neglected by female perpetrators (87% by females versus 43% by males). This finding is congruent with the fact that mothers and mothersubstitutes tend to be the primary caretakers and are the primary persons held accountable for any omissions and/or failings in caretaking. In contrast, children were more often abused by males (67% were abused by males versus 40% by females). The prevalence of male perpetrators was strongest in the category of sexual abuse, where 89% of the children were abused by a male compared to only 12% by a female.

Among all abused children, those abused by their birth parents were about equally likely to have been abused by mothers as by fathers (50% and 58%, respectively), but those abused by other parents, parent-substitutes, or other, nonparental perpetrators were much more likely to be abused by males (80 to 90% by males versus 14 to 15% by females). This general pattern held for emotional abuse, but was slightly different in the area of physical abuse. Children who had been physically abused by their birth parents were more likely to have suffered at the hands of their mothers than their fathers (60% versus 48%), while those who had been physically abused by other parents or parent-substitutes were much more likely to have been abused by their fathers or father-substitutes (90% by their fathers versus 19% by their mothers). For sexual abuse, the child’s relationship to the perpetrator made very little difference, since males clearly predominated as perpetrators, whatever their relationship to the child. Moreover, the severity of the injury or impairment that the child experienced as a result of maltreatment did not appear to bear any relationship to the sex of the perpetrator.

Perpetrator’s Age. The perpetrator’s age was entirely unknown for one-third of the children who were countable under the Harm Standard. Given the prevalence of children maltreated by perpetrators of unknown age, the findings here are tentative, since they could easily be eradicated if all perpetrators’ ages were known.

Among all maltreated children, only a small percentage (13%) had been maltreated by a perpetrator in the youngest age bracket (under 26 years of age). However, younger perpetrators were slightly more predominant among children who had been sexually abused (where 22% had been sexually abused by a perpetrator under 26 years of age) and among children who had been maltreated in any way by someone who was not their parent or parent-substitute (among whom 40% had been maltreated by a perpetrator in the youngest age bracket).

A child’s severity of injury or harm from maltreatment appeared not to be associated with the age of the perpetrator.

Perpetrator’s Employment Status. Perpetrator’s employment status was unknown for more than one-third of the maltreated children, limiting the value of the findings on this issue. Nearly one-half of all maltreated children were abused by a perpetrator who was employed, and this held true for both abuse and neglect. Of the children who sustained serious injury, the majority were maltreated by an employed perpetrator. In no category were the majority of children maltreated by a perpetrator who was unemployed.

Child’s Race and Relationship to the Perpetrator. Because the perpetrator’s race was not known for children submitted to the study solely through non-CPS sources, the child’s race was examined in connection with the relationship to the perpetrator and with the nature and severity of the maltreatment.

For overall abuse, child’s race reflected no notable connection to the relationship with the perpetrator. However, among sexually abused children, white children constituted a greater proportion of children who were sexually abused by their birth parents than of those sexually abused by other parents and parent-substitutes, and by others. Among physically abused children, white children were more prevalent among those who were physically abused by other parents and parent-substitutes than among those who were physically abused by their birth parents or among those physically abused by other types of perpetrators. Although non-white children were the minority of victims in all categories, they were more prevalent among children who were physically or sexually abused by perpetrators other than parents or parent-substitutes.

White children are a larger majority of those who suffered serious injury, whereas nonwhite children’s representation was strongest among those who experienced moderate injury and among those for whom injury could be infeffed based on the severity of their maltreatment.

Sources of Recognition for Maltreated Children

School staff predominated as a source of recognition for maltreated children. School sentinels recognized 59% of the children who suffered maltreatment as defined by the Harm Standard and 54% of the Endangerment Standard total. Other important sources of abused and neglected children were hospitals, police departments, social service agencies, and the general public. For maltreatment defined under the Endangerment Standard, day-care centers also joined in the group of agency categories that encountered more than 100,000 abused and neglected children.

Since the NIS-2, hospitals more than tripled the rate at which they recognized maltreated children; mental health agencies nearly quadrupled their rate of recognition of children who met the Harm Standard and increased their recognition fivefold of children who met the Endangerment Standard; schools more than doubled their rate of recognition of children who met the Endangerment Standard, which included a 70% increase in their recognition rate for the Harm Standard sector. Endangerment Standard recognition more than doubled in law enforcement agencies. Interestingly, there were no changes in the contributions of sources that are tapped in the NIS only through their reports to CPS (e.g., private physicians and the general public). This last finding probably reflects the relatively stable level of CPS involvement with the abused and neglected children countable in the NIS over the time period, as noted below.

Official Reporting of Maltreated Children and Their Investigation by Child Protective Services

The NIS methodology provides information that speaks only to the end result of several processes, indicating whether or not a given maltreated child was or was not among the children whose maltreatment was investigated by CPS. Children who do not receive CPS investigation of their maltreatment represent an enigma to the study, as it cannot be determined whether this was because they were not reported to CPS or because CPS screened their reports out without an investigation.

Despite that limitation, the NIS-3 findings concerning the percentages of abused and neglected children whose maltreatment received CPS investigation are cause for serious concern. Only a minority of the children who were abused or neglected, by either definitional standard, received CPS attention for their maltreatment. CPS investigated the maltreatment of only 28% of children who were countable under the Harm Standard and of only 33% of those whose maltreatment fit the Endangerment Standard. Moreover, the percentages of those who received CPS investigation represented less than one-half of the maltreated children in all categories of maltreatment except fatalities, and across nearly all recognition sources. Especially remarkable was the finding that CPS investigation extended to only slightly more than one-fourth of the children who were seriously harmed or injured by abuse or neglect.

Another important finding was that the percentages of maltreated children who receive CPS investigation have decreased significantly since the NIS-2. The percentage of children receiving investigation among those who met the Harm Standard dropped from 44% to 28%, while the percentage of CPS investigation of children who met the Endangerment Standard fell from 51% to 33%. Although the decline was significant only among children recognized in law enforcement agencies and hospitals, it nevertheless cut across every type of recognition source. The decline in rates of CPS investigation affected abuse under the Harm Standard, all categories of maltreatment under the Endangerment Standard, and all levels of outcomes except fatalities.

At the same time, the actual numbers of countable children investigated by CPS remained stable (when considering Harm Standard totals) or even slightly increased (considering the Endangerment Standard totals). Thus, as the total number of maltreated children has risen, it means that a larger percentage of them have not had access to CPS investigation of their maltreatment. This picture suggests that the CPS system has reached its capacity to respond to the maltreated child population.

Implications

Are the observed increases in the incidence of child abuse and neglect, especially the quadrupling of the numbers of children who were seriously injured or endangered by maltreatment, real increases in the scope of the problem, or do they instead reflect improved recognition on the part of sentinels and other reporters to CPS? The fact that the increases occurred where they did-among children who were seriously injured and among children who were endangered-suggests that both of these dynamics contributed to the observed increases, each dynamic affecting a different sector of the abused and neglected population.

More Children Are Now Being Abused and Neglected Than in 1986, and Their Injuries Are More Serious. The rise in the number of seriously injured children probably reflects a real increase in child abuse and neglect, because it cannot plausibly be explained on the basis of heightened sensitivity. It is unreasonable to suppose that quadruple the number of seriously injured victims of abuse and neglect existed at the time of the NIS-2 and somehow escaped notice by community professionals. The fact that the seriously injured group has quadrupled during the 7 years since the NIS-2, and now comprises more than one-half million children, appears to herald a true rise in the scope and severity of child abuse and neglect in the United States.

Although the NIS does not address the causes of abuse and neglect, it was striking how often illicit drug use was noted in the narrative descriptions on the NIS data forms. The increase in illicit drug use since the fall of 1986 when the NIS-2 data were collected may have contributed to the rise in incidence observed in the NIS-3. Economics is another factor that may have enlarged the problem. Family income is the strongest correlate of incidence in nearly all categories of abuse and neglect, with the lowest income families evidencing the highest rates of maltreatment. Increases in incidence since 1986 may partially derive from decreased economic resources among the poorer families and the increase in the number of children living in poverty.

Community Professionals Are Better at Recognizing Abused and Neglected Children, Especially Those Endangered but Not Yet Harmed by Maltreatment. The rise in the number of endangered children probably stems from improved recognition of more subtle cues-those that indicate abusive and neglectful behaviors that have not yet resulted in harm or injury. It is quite plausible to suppose that some (even sizable) portion of the endangered children escaped attention in the NIS-2, but that by the time of the NIS-3, community professionals had learned to pay better attention to information that might indicate endangering maltreatment. Note that this explanation also completes an account of consistent progression in recognition across the three national incidence studies. The NIS-2 demonstrated an increase in the number of moderately injured children. In interpreting that finding, it was considered likely to have derived from improved attentiveness to moderate-injury indicators of abuse and neglect. The NIS-3 found no statistical change in the numbers of moderately injured children, which suggests that professionals had reached close-to-maximum recognition rates for this category of children at the time of the NIS-2. The fourfold increase in the number of endangered children in the NIS-3 implies that the subsequent further improvements in recognition have now shifted toward even subtler cues:those associated with not-yet-injurious abusive actions and neglectful omissions.

Better Targeting Is Needed To Ensure CPS Investigation for the Children Who Most Need It. The number of NIS-countable children who are investigated by CPS has remained fairly stable, or risen slightly, since the last national incidence study in 1986. As a result, CPS investigation has not kept up with the dramatic rise in the incidence of these children, so the percentages who receive CPS investigation of their maltreatment have fallen significantly. The low rates of CPS investigation of the maltreated children, especially of those already seriously injured by maltreatment, warrant immediate attention.

These findings emphasize the need for better targeting, whether by reporters in referring children to CPS, by CPS screening practices in connection with reports, or by both. One possibility is that, although reporters now demonstrate considerable perceptiveness in identifying maltreated children, they have not reliably translated this into reports to CPS, or are unclear as to how to do so. Another possibility is that CPS, which has increasingly turned to screening cases in order to keep its workload manageably within the range of its resources, has not been using effective screening criteria or has been unclear or inconsistent about the criteria to be applied. Note that these are not independent dynamics, because the response of CPS to a report provides feedback that has consequences for future reporting behaviors. Information bearing on these issues is provided by reports on two of the NIS-3 special substudies: the Sentinel Questionnaire Follow-up Study, which asked school sentinels about their decisions to report cases to CPS, and the CPS Screening Policy and Recordkeeping Study, which examined the screening policies and practices of CPS agencies that participated in the NIS-3.

The main NIS-3 data can offer some guidance in targeting. Neglect warrants more attention. It affects the greatest number of maltreated children, and their injuries are often serious. Children from the poorest families are at the greatest risk of maltreatment, so these children may warrant increased CPS attention as well. Children in single-parent families also experienced higher rates of maltreatment. A number of characteristics explored here are not unrelated to each other-for instance, single-parent families often have lower incomes. Further analyses of the NIS-3 data can address the independent contributions of different characteristics to better clarify risk factors that can guide CPS screening activities. Narratives on the NIS data forms can also be more systematically explored. The narratives often included spontaneous comments about illegal drug use, indicated whether the perpetrator had a history (sometimes a criminal record) of sexually or physically abusing/assaulting other children or adults, or noted that the incident described was not the first time the child had been abused or neglected.

As part of improving CPS targeting of the more serious cases, efforts should also focus on achieving better consensus about what types of cases should not receive CPS investigation. Very few of the educationally neglected children currently have their maltreatment investigated by CPS, and those who do may have been maltreated in multiple ways, with the CPS investigation focusing on abuse or other types of neglect. The current role of CPS in relation to educational neglect might be the centerpiece of an emerging consensus on what specific forms of abuse or neglect should not receive CPS investigation.

Forging Working Relationships Between CPS Agencies and Schools. The NIS has consistently demonstrated that professionals in schools play a central and critical role in identifying children who are abused and neglected. As policies are developed to address the burgeoning problem of child abuse and neglect, they should capitalize on the unique role of school professionals as front-line observers.

Information provided by the DHHS.