FAMILIES WERE ALL FROM ONE LARGE METROPOLITAN CITY. THEIR RECIDIVISM RATE, 120 REPORTS OVER A 7-YEAR PERIOD, WAS THE SAME AS FOR ALL ABUSING FAMILIES. RESULTS SUPPORTED PREVIOUS FINDINGS THAT FAMILIES WITH A CHILD FATALITY, UNLIKE OTHER ABUSING FAMILIES, ABUSE MORE THAN ONE CHILD. HOWEVER, MEAN AGE AND AGE RANGE OF CHILDREN KILLED WAS THE SAME AS FOR ALL ABUSING FAMILIES. ALL THE SAMPLE FAMILIES WERE WELFARE DEPENDENT, PROBABLY DUE TO THE AREA FROM WHICH THE SAMPLE WAS DRAWN. IN ADDITION, FAMILY CONSTELLATIONS WERE CHAOTIC, WITH 109 FATHER FIGURES VISITING THE HOMES SPORADICALLY OVER THE 7-YEAR PERIOD. MOTHERS WERE THE PERPETRATORS IN 82 PERCENT OF THE ABUSE AND 80 PERCENT OF THE FATALITIES, AND COLLUDED WITH THE FATHER OR PARAMOUR IN 13 PERCENT OF THE KILLINGS. LIMITED DATA ON PARENT-CHILD RELATIONSHIPS INDICATED THAT BEHAVIOR CONTROL WAS LARGELY CHAOTIC, WITH ARBITRARY SHIFTS FROM RIGID TO PERMISSIVE CONTROL PATTERNS. RECORDS ALSO SHOW A PREVALENCE OF STRESS FACTORS INCLUDING HUMAN STRESS (CHRONIC HEALTH PROBLEMS IN MOTHERS, FATHER FIGURES, AND CHILDREN), MILIEU STRESS FACTORS (FREQUENT MOVES, INADEQUATE SPACE AND FACILITIES), POOR QUALITY OF CHILD CARE, AND POOR FAMILY SUPPORT SYSTEMS. RESULTS INDICATED THAT A CHILD FATALITY IS MOST LIKELY IN AN ABUSING FAMILY WHERE THE MOTHER IS THE ONLY CONTINUING ADULT IN THE HOME, WHERE MORE THAN ONE CHILD IS ABUSED, WHERE THE FAMILY CONSTELLATION IS CHAOTIC, WHERE THE MOTHER IS THE PERPETRATOR OF THE ABUSE, AND WHERE THERE ARE MANY HUMAN AND MILIEU STRESSES. RESULTS INDICATED THAT AGENCIES SERVICING SUCH FAMILIES LACK ADEQUATE DATA COLLECTION AND DATA USE SYSTEMS AND THEREFORE FAIL TO PREVENT FATALITIES. A PREVENTIVE PROGRAM SHOULD INCLUDE SCREENING OF ALL ABUSING FAMILIES FOLLOWING THE FIRST REPORT OF ABUSE, ATTEMPTS TO CHANGE THE FAMILY MILIEU, REMOVING THE CHILD FROM THE HOME IF NECESSARY, AND TEAM RATHER THAN INDIVIDUAL DECISIONMAKING BY WORKERS INVOLVED. ONE TABLE AND A BIBLIOGRAPHY ARE INCLUDED. (CFW)