Amer. J. Orthopsychiat., 53(4), October 1983, 645-653.

Child Abuse and Accidents in Black Families: A Controlled Comparative Study

Jessica H. Daniel, Ph.D., Robert L. Hampton, Ph.D., Eli H. Newberger, M.D.

Families of child abuse and accident victims were evaluated in a study of 402 families with children up to four years of age. Children were matched on a one-to-one basis with a control according to race, socioeconomic status, and age. Socioeconomic factors that playa significant role in imposing undue stress upon many families are identified, and implications for prevention and for practice are offered.

Child abuse and neglect, accidents, poisonings, and failure to thrive are prominent among the “social illnesses” of pediatrics. They derive from many causes in the child, the family, and the environment. Their clinical classification, however, suggests overly simplified models of etiology.14 For a clinician, for example, a diagnosis of “child abuse” requires the suspicion and verification of parental responsibility for the injuries of the victim. We have previously noted the susceptibility of poor and socially marginal families to receiving this diagnosis13 and to the risks of a preventive approach that emphasizes screening.4 The public clinical settings in which most of these diagnoses are made and the highly value-laden nature of the clinical judgments involved in the formulation that a given child is “abused” appear to account for the special tendency of these families to be reported as families who abuse their children. This reporting exposes children to the public child welfare system which, as the Carnegie Council on Children underlined,8 may have only the crudest implements of protective service available (e.g., placing the child in a foster home).

Accidents, by contrast, are suggested, in the conceptual model implicit in the name, to occur at random. They are numerically prevalent and account for the lion’s share of childhood morbidity and mortality in the United States. Because the name implies chance etiology, however, the causes of the traumatic events, in child, family, and environment, are most often ignored in practice and in the formulation of social policy. Practitioners who care for the children of more affluent families, where there is a smaller social distance between clinician and family, seem less likely to accuse the parents by suggesting in their clinical diagnoses that the injuries they treat derive from some parental responsibility or fault. They are paid, for the most part, on a fee-for-service basis; an economic incentive favors the choice of the least offensive diagnosis for how the child received an injury. The ethics of practice in regard to the sharing of confidential personal data seem also to differ for the affluent in contrast to the poor. Even when child abuse is suspected, it appears less likely to be reported.

Discussions with physicians’ and empirical research suggest that social and economic characteristics of families play an important role in determining which children are labeled as “abused.” Turbett and O’Toole19 found that child abuse recognition among physicians is affected by severity of injury, and by parents’ socioeconomic status and ethnicity. When children were described as having a major injury, black children were nearly twice as likely to be recognized as victims of abuse as were white children. With respect to reporting child abuse, black children were one-third more susceptible than white children suffering from identical major injuries. Lower-class white children suffering from major injuries were more likely to be classified as abused than were upper-class white children.

The National Black Child Development Institute has drawn attention to the unique health risks which accrue to black children:16

Black children are much more likely to suffer from poor health than the majority of their American peers. While poverty, unsafe housing, and poor nutrition expose many black children to harmful and hazardous conditions, their plight is compounded by a systematic inaccessibility to competent health care. Together, these factors help to make many black children a population substantially at risk with no resources for assistance.

The statistics detailing the effects of deteriorating environments are particularly grim. Black infants are almost twice as likely as white infants to die before their first birthday. Even afterwards, a black child has a 30% greater probability of dying by his/her fourteenth birthday than does a white child.

Black children have a higher death rate as a result of accidental injuries than whites. Also, black children appear to receive less medical attention for their injuries than do white children.9

The data on family violence, of which child abuse can be seen as a subset, are inconsistent, and their interpretations can be misleading.18 In an effort to define the extent and nature of family violence among black Americans, Cazanave and Straus3 analyzed a sub-sample of a 1976 survey of violence in 2143 black families and 427 white families. While black husbands tended to be more violent toward their wives, self-reported acts of physical violence toward their children showed little difference between blacks and whites. No conclusions could be drawn with respect to the differential levels of violence in black and white families.

In the National Study of the Incidence and Severity of Child Abuse,2 ethnicity and family income did not emerge as major factors in the reporting or non-reporting of identified cases of child maltreatment. Among high income groups (i.e., families with annual incomes of $15,000 or more), incidence rates for all forms of maltreatment were approximately the same for white and nonwhite children. For white children, incidence rates for all forms of maltreatment were much higher in the low income groups than in the higher income groups. For nonwhite children, a similarly strong association between poverty and incidence of maltreatment was found, but only with regard to child neglect. Child abuse incidence rates were low and essentially constant across income groups for nonwhite children.

These findings of ethnically variant patterns of family violence suggest a need for a deeper analysis of risk factors within racial groups rather than the traditional between-group comparisons. This paper presents a comparison of risk indicators for accidents and abuse among the black families that participated in a larger study of pediatric social illnesses.


The Family Development Study is a prospective case-control study of pediatric social illnesses: abuse, neglect, failure to thrive, accidents, and ingestions. These illnesses appear to derive primarily from the child’s physical and social interaction with the nurturing environment. They are distinguishable from illnesses of a more primary biomedical etiology. The goals of the study are to determine which ecological factors place a child at risk for each of the specific social pediatric illnesses, and to identify commonalities and differences in risk indicators across the diagnostic categories.

Two successive samples of “cases” and matched “controls” were identified. The first sample (Phase I), reported previously,14 yielded data that suggested interrelationships among the social illnesses. This report derives from a second wave of observations (Phase II) which built on the main conceptual leads of the earlier study a more focused and rigorous approach to analysis.

With a view to building a structural element which would regularly oblige us to consider the special visibility and vulnerability of minority families, especially in regard to increasing the utility of our research products for clinical practice and social policy, we entered into a contractual agreement with the Boston Community Research Review Committee to meet with us at regular intervals and to monitor continuously our draft instruments and data gathering.


All families with children four years of age or younger who were hospitalized at CHMC with a diagnosis of child abuse, child neglect, accident, ingestion, or failure to thrive were eligible for inclusion in the study. Families of children who did not bear a pediatric social illness diagnosis, but who were hospitalized for medical or surgical illnesses of acute onset, such as pneumonia or meningitis, were eligible for selection as controls. Children suffering from chronic or terminal illnesses were excluded from the control population. Families seen in the outpatient department or emergency room were not eligible for study.

For each case, children were matched on a one-to-one basis with a control according to the following criteria: race, socioeconomic status (i.e., same social class on the Hollingshead Two-Factor Social Index), and age (± 2 months). No violations from these matching criteria were permitted.

All classifications for the purposes of this study are based upon final hospital diagnosis. The diagnosis of child abuse was made by the hospital’s Trauma X Team. This multidisciplinary team, composed of a psychologist, a psychiatrist, pediatricians, a social worker, nurses, and a lawyer, has primary responsibility for consulting on child abuse cases in the hospital.12 All cases of child abuse were reported to the state’s child protective services as mandated by state law.

The sample was ascertained from July 1975 to April 1977. Prior to making contact with a mother to ask permission to interview, the physician responsible for the child’s hospital care and the social worker, if one was assigned, were asked for their permission to interview the mother. Only then was contact made with the mother. After explaining the goals and nature of the study, but before beginning the interview, the mother’s written informed consent was obtained. All families’ participation was voluntary and without remuneration.

Table 1

As summarized in TABLE I, 209 cases and 209 controls were interviewed. Eight cases of failure to thrive were also reported to the state’s department of welfare as being seriously abused or neglected. These children are included in both the Trauma X and F. T. T. categories.

Just under 70% of all eligible families were interviewed. Post-hoc testing reveals no major differences between missed and interviewed families, either across or within any specific attribute of race, sex, social class, urban residence, or duration of hospital stay (a sign of medical severity).


Data on each child and family were obtained from four sources: maternal interview, paternal interview, the child’s medical records, and a Vineland Social Maturity index based on information derived from the maternal interview. The principal instrument was a standardized, precoded maternal interview. It included a wide range of questions about family structure, housing, employment, finances, availability of relatives and friends, mobility, and psychological stresses, as well as questions about the respondent’s history, focusing on the demographic characteristics of the family of origin, events and care in childhood, and prior and current experiences in caring for the child. The mother was asked questions about her child’s temperament (based on Carey’s interview adaptation of the Thomas, Chess and Birch Infant Scales) and health, as well as her knowledge of developmental norms for children. Finally, questions regarding the frequency and methods used to punish and praise the child were asked. The same interview was administered to the father or any regular male caretaker. Both parents were asked questions about themselves and their spouse. From the medical records, physical characteristics were ascertained, i.e., height, weight, hematocrit, head circumferences, and duration of hospitalization.

All interviews were conducted by specially trained, sensitive interviewers. Each interview lasted approximately one hour. All were conducted in the hospital, either during the child’s hospitalization or upon the child’s first visit to a follow-up clinic. All data were checked by a supervisor immediately following the interviews to insure that they were appropriately coded and recorded. Given the emphasis on environmental realities during the interview, it was felt that the project had an obligation to offer assistance to ameliorate the identified problems. To this end a family advocacy program was developed and was available to all participants. Designated initially to help families obtain such essential supports as adequate housing, child care, legal services, and adult health care, the program evolved into an organized service available to all hospital patients. Personnel with no formal professional education were trained and supervised in helping families deal with contemporary life stresses and in gaining access to essential services.11

Demographic Characteristics

No differences in race were found across the case categories. By race the sample was 73% white, 22% black, and 5% Hispanic; these proportions did not vary markedly from category to category. No sex differences across the case categories were found. The majority of the sample was male; this is consistent with the generally observed higher morbidity for males at these ages.

On age, however, cases differed. The children bearing abuse or failure to thrive diagnoses were significantly younger than those with accidents and ingestions: 85% of both child abuse and failure to thrive victims were under two years of age; 55% were under one year of age. Those with accidents were almost evenly distributed in all the age groups. Children with ingestions tended to be older; none were below one year of age.

Families differed with respect to welfare dependency. A high percentage of abused children (four out of five) were members of families on welfare. The families of abuse victims were, as a group, very poor. The families of victims of accidents and ingestions were more evenly distributed across all five of the Hollingshead social classes. Excluding all of the child abuse cases from the failure to thrive sample brings its social class distribution into line with the accidents and poisonings. The prevalence of poverty in the families of abuse victims is one of the striking demographic findings of this study.


The subsample consisted of 94 black families. There were 15 cases of child abuse, nine of failure to thrive, 17 with accidents, and six with ingestions. Each of these was matched with a control. The following analysis addresses the differences across the accident, abuse, and control categories in the black subsample.

Case-Control Differences

Black child abuse cases differed from their controls in several ways, including a higher level of social isolation, more geographic mobility, maternal childhood history of corporal punishment extending through adolescence, and a generally more stressful present living situation.

The mothers of the black child abuse cases tended to describe their childhoods as having been less happy and cold in tone. These mothers were struck more frequently and spanked more severely than were the mothers in the matched control group. Mothers of the child abuse victims reported having been spanked on parts of the body other than the hand and bottom through adolescence. But they had more prior experience with children, and were happy during their pregnancies. The pressures on these mothers seemed extreme. They suffered more losses due to recent deaths in their families, more recent changes in their life situations, and, overall, more negative family stress. It is not surprising that they described their living situation as being less happy than did the controls.

Fathers of black child abuse cases, like case mothers, were spanked more frequently than the controls, and their childhood punishments were more severe. These fathers reported that they liked infants less than did the fathers of the matched controls (note that black child abuse cases had a mean age of 14 months).

The victims of accidents were older (mean age of 23.3 months) and relatively more healthy as a group in comparison to their controls. While these families were highly mobile, they appeared to have greater access to shopping and to recreation than did the comparison group. They expressed greater satisfaction with social service agencies. The fathers in the families of accident victims had moved less frequently during childhood. They left their parents’ homes relatively shortly before the birth of their first child. The mothers of accident victims, however, appeared to be more depressed as a group than did the control mothers.

Fewer factors differentiated black families whose children suffered accidents from their controls than did black families bearing the diagnosis of child abuse. Although they appeared as a group to require fewer social services, they seemed more capable of establishing relationships with service providers. The fathers had experienced fewer changes in childhood and had been on their own relatively fewer years before assuming the responsibilities of parenthood.

Accident-Abuse Differences

Table 2
Trauma-X, White
Trauma-X, Black
FTT, White
FTT, Black
Accidents, White
Accidents, Black
Ingestions, White
Ingestions, Black
1 Hollingshead Scale, Category 5


Table 3
Child Factors
Vineland Social Maturity Quotient*(—)1
Child perceived as easy to care for*(—)
Communication score*(—)
Mother-Child Interaction
Mother felt guilty when she punished child*(—)
Mother-child contact in the first few weeks*(—)
Frequency of spanking*
Family Factors
Mother unable to “get going” each day*(+)
Mother felt no one was interested in her problems*(+)
Mother described situation as happy**(—)
Father made rural-urban move*(+)
Recent death in the family**(+)
Recent family stress*(+)
* p<.05; ** p<.001.1 Indicates direction of difference.

As noted in TABLE 2, 86.6% of the Trauma X families and 64.7% of the accident families were in the lowest social class. To control for confounding on this variable, the cases were matched on social class. The significant variables summarized in TABLE 3 suggest that, among black families, victims of child abuse and accidents differ in several ways. The children bearing the diagnosis of abuse scored lower on the Vineland Social Maturity Index. Their mothers’ occupational status, as a group, was lower. Maternal reports of the number of times per week they spanked their child indicated no difference between the two case subsamples. Mothers of accident victims reported that they were more likely to feel guilty if they punished the child than did mothers of abuse victims, They were, as a group, also more likely to describe their relationship with their infant during the first few weeks as pleasant.

The parental social environment and support network differed between the two groups. Our data strongly suggest an important degree of social isolation and of maternal depression for black mothers of child abuse victims. Although the mothers’ marital status did not differentiate at a statistically significant level in this study, it is noteworthy that the abuse sample contained more single parents. Relatively more of these mothers had moved to the greater Boston area from rural areas, which may have separated them from their familial support systems. The fathers, on the other hand, tended to be from urban areas.

The mothers of child abuse victims described their living situations as unhappy. They expressed as a group the sad feeling that no one was interested in their problems, and they described themselves as being unable to “get going” each day. An unexpected finding was that these differences existed in such a magnitude. Although there was no significant difference in the number of times per week the mothers reported seeing relatives, there appeared to be a substantial qualitative difference in the extent to which they felt nurtured by their contacts with kin.

Stress appeared to bear onerously on families with children diagnosed as abused. Recent losses of loved ones exerted a strong impact, and the cumulative stress levels for families of abuse victims was extremely high.


In the present study, black families who abuse their children appear to suffer from poverty, social isolation, and stressful relationships with and among kin. Maternal depression and poor mobility were noted more frequently in black families whose children’s injuries were seen as having been accidental, yet they had many strengths in comparison to the families whose children were diagnosed as abused. The accident-abuse comparison suggests that although a family’s ability to protect a child from an environmental hazard may be enhanced by the association of a mother’s sense of well-being and connection to kin and community, the crushing burdens associated with child abuse seem qualitatively and quantitatively to erode the family’s parental competency. For the black families in this study, the contexts of child abuse appear to be those of severe economic adversity, no one to turn to for help, a death in the family, a history of having suffered serious personal violence, and a child who may be delayed in social and cognitive development. Any parent would be severely affected by such circumstances and, indeed, the parents of victims of child abuse in this study could easily be seen as victims themselves.

The implications of these data for prevention and for practice seem clear, but, in the light of present trends in social policy, they would not seem to be able to translate into action quickly. As a matter of priority, to strengthen these most vulnerable families, we feel that society must provide, at least, and in this order: 1) financial support; 2) a line of contact with a professionally trained or supervised individual who can give help in times of personal distress; 3) diagnostic consultation; and 4) therapeutic intervention to promote the development and health of the children. Now is, perhaps, the least propitious time in recent years, however, to urge these supports for the prevention of child abuse in the black families that seem to bear a disproportionate brunt of a stagnant economy and the general withdrawal of fiscal support, social services, and medical services for the poor. Moreover, the racial prejudice implicit in present social policies, which blame the victim for his or her life circumstances, visits additional cruelty and violence upon the black family and its children. When such auxiliary supports to social work counseling as child development services are not available for the treatment of child abuse, the only choice may be to separate child from parent and to place the child in a foster home. This final protective “service” is well-documented to affect black families disproportionately. For them, too frequently the last resort becomes the first resort when society’s resources—in the workplace; in the housing market; and in the education, health, and social service systems—are in short supply.

Thus, child abuse in the black family is conceptualized best as a symptom, rather than as a disease. To address the origins of the symptom in individual families, for both children and parents, requires a thoughtful professional response. Nevertheless, the professional community cannot ignore the social action imperatives to work on the problem of child abuse. Responsibility for child abuse in the black family resides as much in the society as in the relationships among family members.


1. ALLEN, W. 1978. The search for applicable theories of black family life. J. Marr. Fam. 40: 117-128.

2. BURGDORF, K. 1980. Recognition and reporting of child maltreatment: findings from the national study of the incidence and severity of child abuse and neglect. Prepared for the National Center on Child Abuse and Neglect, Washington, D.C.

3. CAZENAVE. N. AND STRAUS, M. 1979. Race, class, network embeddedness and family violence: a search for potent support systems. J. Compar. Fam. Stud. 10:281-300.

4. DANIEL, J. ET AL. 1978. Child abuse screening: implications of the limited predictive power of child abuse discriminants in a controlled family study of pediatric social illness. Inter. J. Child Abuse and Neglect 2:247-259.

5. FONTANA, V. 1973. The Maltreated Child: The Maltreatment Syndrome in Children. Charles C Thomas, Springfield. III.

6. GELLES, R. 1980. Violence in the family: a review of research in the seventies. J. Marr. Fam. 42:873-885.

7. GIL, D. 1970. Violence Against Children: Physical Child Abuse in the United States. Harvard University Press, Cambridge, Mass.

8. KENISTON, K. 1977. All Our Children: The American Family Under Pressure. Harcourt, Brace and Jovanovich, New York.

9. MANHEIMER, D., DEWEY, J. AND MELLINGER, G. 1966. 50,000 child years of accidental injury. Pub. Hlth Rep. 81:519-532.

10. MARX, T., NEWBERGER, E. AND WHITE, K. 1981. When an injury is a symptom. Presented to the National Conference for Family Violence Researchers, University of New Hampshire.

11. MORSE, A. ET AL. 1977. Environmental correlates of pediatric social illness: preventive implications of an advocacy approach. Amer. J. Pub. Hlth 67(7):612-615.

12. NEWBERGER, E. ET AL. 1973. Reducing the literal and human cast of child abuse: impact of a new hospital management system. Pediatrics 51(5):840-848.

13. NEWBERGER, E. AND DANIEL, J. 1976. Knowledge and epidemiology of child abuse: a critical review of concepts. Ped. Annals 5(3):140-145.

14. NEWBERGER, E. ET AL. 1977. Pediatric social illness: toward an etiologic classification. Pediatrics 60: 179-185.

15. One child dies daily from abuse: parent probably was abuser. 1975. Ped. News 9(3).

16. The Status of Black Children 1980. Report from the National Black Child Development Institutes, Inc.

17. STAPLES, R. AND MIRANDE, A. 1980. Racial and cultural, variations among American families: a decennial review of the literature on minority families. J. Marr. Fam. 42:887-903.

18. STRAUS, M., GELLES, R. AND STEINMETZ, S. 1980. Behind Closed Doors: Violence in the American Family. Doubleday, Garden City, N.Y.

19. TURBETT, P. AND O’TOOLE, R. 1980. Physicians’ recognition of child abuse. Presented to the American Sociological Association, New York.

20. United States Senate Hearing Before the Subcommittee on Children and Youth of the Committee on Labor and Public Welfare. 1973 United States Senate 93rd Congress, First Session on S. 1191, Child Abuse Prevention Act. Washington, D.C., U.S. Senate Government Printing Office.

21. Vital Statistics of the United States. 1980. National Center for Health Statistics, U.S. Department of Health and Human Services.

Submitted to the Journal in September 1982. Authors are at: Harvard Medical School and Judge Baker Guidance Center and Children’s Hospital Medical Center, Boston (Daniel); Children’s Hospital Medical Center, Boston, and Connecticut College, New London, Conn. (Hampton); and Harvard Medical School and Children’s Hospital Medical Center, Boston (Newberger). Research was supported by grants from the Office of Child Development (OCD-CB-/4/) and from the Center for Studies of Crime and Delinquency, National Institute of Mental Health (1 TO1 MH15117) 1A2 CD and 1 732 MH17063-01 RERC).