Childhood Victimization

Childhood Victimization:

EARLY Adversity, LATER Psychopathology

by Cathy Spatz Widom about the author

hildhood physical abuse, Cathy Spatz Widom is a professor of criminal justice and psychology at the State sexual abuse, and neglect University of New York at Albany. She is widely recognized for her work on the cycle of violence. Chave both immediate and long-term effects. Different types of The author wishes to thank Patricia J. Glynn and Suzanne Luu for their help in the abuse have a range of consequences preparation of this article. for a child’s later physical and psy- The research described in this article was supported by grants from the U.S. chological well-being, cognitive Department of Justice, National Institute of Justice (86–IJ–CX–0033, development, and behavior. But 89–IJ–CX–0007, and 94–IJ–CX–0031), and the U.S. Department of Health and there is another side to the issue: Human Services, National Institute on Alcohol Abuse and Alcoholism (AA09238) and National Institute of Mental Health (MH49467). Because these crimes often occur against a background of more chronic adversity, in families with multiple problems, it may not be from poor peer relations to physical Family and community—the reasonable to assume that before aggression and antisocial behavior context. The long-term impact of being victimized the child enjoyed to violent behavior. These conse- childhood trauma may depend on “well-being.” Parental alcoholism, quences are influenced by such fac- the larger––family or community–– drug problems, and other inade- tors as gender differences and the context.6 In a study of children kid- quate social and family functioning context in which victimization naped and held underground, pre- are among the factors affecting the occurs. existing family pathology was iden- child’s response to victimization. tified as a factor in the victims’ long- Gender differences. Differences term adjustment. Four years after Gender differences add to the com- between men and women in mani- the incident, the children from trou- plexity. Disentangling all these fac- festing the effects of childhood vic- bled families were more maladjusted tors is difficult, as researchers have timization have received only limit- than those from healthier families.7 found. ed attention from scholars. Some The findings of other research were researchers, exploring how men and Clearly, more needs to be learned not as clear; rather, subsequent mal- women differ in showing distress, about the long-term consequences adjustment was linked more to have suggested there is some confor- of childhood victimization and the whether victimized children mity to traditional notions of male processes linking it to outcomes received appropriate play materials and female behavior.2 Some have later in life. This article discusses and maternal involvement than to noted that differences between men what is known from earlier studies whether they were abused.8 Parental and women in manifesting the con- and also presents the findings of alcoholism is another contextual 1 sequences of abuse may parallel more recent research. factor linked to child abuse9 and gender differences in the way psy- to alcoholism later in life in the chopathology is expressed. Thus, offspring.10 Consequences aggression (in males) and depres- and What Gives sion (in females) may express the In the same way, practices of the Rise to Them same underlying distress, perhaps community and the justice and reflecting gender-specific strategies social service systems may have Child maltreatment has physical, for maintaining self-esteem in the long-term effects. Researchers have psychological, cognitive, and face of perceived rejection.3 called attention to the ways in which behavioral consequences. Physical children who are members of racial Differences in the way boys and girls consequences range from minor and ethnic minorities encounter dis- react to abuse have been reported in injuries to brain damage and even crimination, which diminishes their a few studies. In one, boys were death. Psychological consequences self-esteem and exacerbates the found to have more externalizing range from chronic low self-esteem, effects of victimization.11 Elsewhere, and girls to have more internalizing anxiety, and depression to substance researchers have suggested that vic- symptoms.4 An examination of abuse and other self-destructive timized children are more likely to depression and conduct disorders in behavior and suicide attempts. develop problem behavior in adoles- sexually abused children revealed Cognitive effects include atten- cence partly because of juvenile jus- that girls were more likely than boys tion problems, learning disorders, tice system practices that dispropor- to develop depressive disorders and and poor school performance. tionately label them as juvenile less likely to develop conduct disor- Behavioral consequences range offenders and adjudicate them as ders.5 such.12

National Institute of Justice Journal ■ January 2000 3 both scored at the lower levels of the How the Study Is IQ scale, with the majority in both groups below the standard mean of Being Conducted 100 (see figure 1, next page). Those who were abused or neglected, how- The study is based on a “prospective cohorts design,” so-called because it follows a ever, scored significantly lower than group of people (a cohort) for an extended period, enabling researchers to examine the comparison group, and these sequences of development over time. In the case of this study, the design helps sort lower levels persisted irrespective of out the effects of childhood victimization from other, potentially confounding effects age, sex, race, and criminal history. traceable to different causes. The subjects were told they were part of a study of the characteristics of people who had grown up in the area in the late 1960’s and early Overall, both groups averaged 11.5 1970’s. years of schooling, but the abused and neglected group completed sig- The cases of children who were abused and/or neglected were drawn from county nificantly fewer years. Thus, the juvenile and adult criminal court records in a metropolitan area of the Midwest childhood victims were less likely to between 1967 and 1971. The children were young–age 11 or younger–at the time have completed high school: Fewer of the incident. than half, in contrast to two-thirds of the people in the control group. The comparison group. To create a control group against which to compare the abused and neglected children, a group of children who had not been reported as victimized Behavioral and social develop- but who were similar in other respects to the study subjects were identified. To match ment. The occupations of both children younger than school age at the time of the incident, county birth records were groups ranged from laborer through used. To match school-age children, records of more than 100 elementary schools were professional. In the sample overall, used. the median job level was that of semiskilled worker, with fewer than Sample size and characteristics. The original sample consisted of 1,575 people, of 7 percent in the two groups holding whom 908 were study subjects and 667 were controls. Of these, 1,196 were interviewed managerial or professional jobs for the study. Just under half the interviewees were female, about two-thirds were white, (see figure 2, next page). The abused and the mean age at the time of the interview was 28.7. There were no differences and neglected individuals had not between the abused/neglected group and the controls in gender, race/ethnicity, or age. done as well as the control group: Significantly more of them held Some caveats. Because the study findings were based on court cases, they most likely menial and semiskilled jobs. Con- represent the most extreme incidents of childhood abuse and neglect. What is more, versely, a larger proportion of peo- they were processed before enactment of child abuse laws, when many cases went ple in the control group held higher unreported and thus never came to the attention of the authorities. The findings are level jobs, ranging from skilled therefore not generalizable to unreported or unsubstantiated cases of abuse and neglect. worker through professional. Because cases brought before the courts disproportionately represent people at the Unemployment and underemploy- lower end of the socioeconomic spectrum, the study’s subjects and controls were ment disproportionately affected the drawn from that stratum. For this reason, it would be inappropriate to generalize to abused and neglected group (see cases involving people from other socioeconomic strata. figure 3, page 6). In both groups, more than one-fifth had been unemployed in the 5-year period before they were interviewed for the Studying the Long- involved in delinquency and crime, study. Not surprisingly, people in 13 Term Effects in Depth including violent crime. The cur- the control group were more likely rent focus is on how their intellectu- than the victims to be employed. In a systematic study of the long- al, behavioral, social, and psycholog- For underemployment, the story is term consequences of early child- ical development was affected. This similar: Significantly more victims hood abuse and neglect, the author second phase began in 1989, more of childhood abuse and neglect were is examining the experiences of than 20 years after the victimization. underemployed in the 5 years before more than 900 people who were (See “How the Study Is Being Con- the interview than were controls. victimized in childhood. Begun in ducted.”) 1986, the study first focused on the The quality of interpersonal rela- extent to which, as the victims grew Intellectual performance. When tions also is affected by childhood into adulthood, they became tested at about age 29, the study victimization, and here again there subjects and the comparison group

Childhood Victimization: Early Adversity, Later Psychopathology 4 are no surprises (see figure 4, next Figure 1: IQ Scores— page). Using marital stability as the Abused/Neglected Group and Control Group measure of success, child abuse and 40% neglect victims did not do as well as control group members. Almost 20 percent of the controls reported a stable marriage, compared to only 13 percent of the abuse and neglect 30% group. Frequent divorce and separa- tion were also more common among abused and neglected Abuse/Neglect Group Control Group people. 20% As reported in previous research, childhood victimization also increases the risk of criminal behav- ior later in life—as measured by 10% arrests for delinquency and adult criminality, including violent crime.14 The current study confirms these findings. The odds of arrest 0% for a juvenile offense were 1.9 times 0Ð69 70Ð79 80Ð89 90Ð99 100Ð109 110Ð119 120Ð129 130+ higher among abused and neglected individuals than among controls; Note: Number = 1,185 Standardized Scores for crimes committed as an adult, IQ scores are based on the Quick Test. See Ammons, R.B., and Ammons, C.H., “The Quick Test (QT): the odds were 1.6 times higher (see Provisional Manual,” Psychological Reports 11 (1962): 11–162 (monograph supplement 7-VII). table 1, page 7). Childhood abuse or neglect increases the risk of being did not become delinquents or dence were some of the measures arrested for violent crime, whether criminals. of psychopathology. The abused in the juvenile or adult years, as well and neglected individuals were as for crime in general. It is perhaps Psychological and emotional significantly more likely than the most important to note, however, fallout. Suicide attempts, diagnosis controls to have attempted suicide that a substantial proportion of the of antisocial personality disorder, and to have met the criteria for abused and neglected children and alcohol abuse and/or depen- antisocial personality disorder (see table 2, page 7), findings irrespective of age, sex, race, and criminal histo- ry. High rates of alcohol abuse were Figure 2: Occupational Status— found in both groups (more than Abused/Neglected Group and Control Group 50 percent in each), although the 70% abuse/neglect victims were not at greater risk than the controls, a 60% Abuse/Neglect Group Abuse/Neglect Group finding that departs from other Control Group Control Group 50% research but that methodological 15 40% differences might explain.

30% As other research has shown, gender

20% can affect the development of psy- chopathology in abused and 10% neglected children later in life. The 0% current study revealed some of these Menial/ Skilled, Clerical/ Professional/ Semiskilled Semi-Professional Managerial gender-based differences. Females abused and neglected in childhood Note: Number = 1,167 were more likely than controls to Occupational status was coded according to the Hollingshead Occupational Coding Index. attempt suicide, to abuse alcohol or See Hollingshead, A.B., “Four Factor Index of Social Class,” New Haven, CT: Yale University Working Paper, 1975.

National Institute of Justice Journal ■ January 2000 5 Figure 3: Employment History— However, where there was no Abused/Neglected Group and Control Group parental criminality, being abused and/or neglected did increase the 70% risk for this disorder. This compli- 60% cates attempts to understand the consequences of childhood victim- Abuse/Neglect Group 50% Control Group ization and also suggests multiple 40% factors in the development of anti- social personality disorder. 30% A different picture and set of rela- 20% tionships were found for alcohol 10% abuse. When parental alcohol/drug abuse, childhood victimization, and 0% Employed Underemployed Unemployed subsequent alcohol problems in off- Note: Number = 1,196 with Reason spring were analyzed, the parents’ substance abuse problem emerged Employment history findings are based on a measure used in Robins, L.N., and D.A. Regier, eds., as the critical factor in the develop- Psychiatric Disorders in America: The Epidemiological Catchment Area Surveys, New York: Free ment of the same problem in the Press, 1991:103. children, and this held true whether or not the child had been victimized (see table 5, page 8). The study also be dependent on it, or to suffer arrested, and the likelihood of the showed that, as a group, the children from an antisocial personality disor- offspring’s developing antisocial who were abused or neglected were der. Like females, male victims were personality disorder. The analysis no more likely than controls to found at greater risk than controls revealed that among people who develop alcohol problems, whether of attempting suicide and develop- had a parent with a history of arrest, or not the parent had the same ing an antisocial personality disor- abuse or neglect in childhood did problem. der, but they were not at greater risk not increase the likelihood of their of developing alcohol problems (see developing an antisocial personality The strong influence of parental table 3, next page). disorder (see table 4, page 8). characteristics on the offspring, regardless of victimization, warrants The findings of males’ higher risk for antisocial personality disorder and females’ higher risk for alcohol problems parallel previous research Figure 4: Marital History— revealing conformity to gender Abused/Neglected Group and Control Group roles. However, the finding that females are, like males, at risk for 35% antisocial personality disorder (as well as criminal behavior)16 may 30% Abuse/Neglect Group call for reconsidering the assump- Control Group 25% tions of externalizing and internaliz- ing as the respective pathways of 20% male and female response. The context of victimization. 15% The findings confirmed earlier 10% research identifying context as a factor influencing the long-term 5% outcome for victims. This became evident in analyzing the relation- 0% ships among childhood victimiza- Stable No Marriage 1 Divorce/ More than Cohabit Only Marriage or Cohabitation Separation 1 Divorce/ tion, having a parent who had been Separation

Childhood Victimization: Early Adversity, Later Psychopathology 6 more careful consideration, but is Table 1: Childhood Victimization and Later Criminality consistent with earlier literature on the genetic transmission of alco- Abuse/Neglect Group Control Group holism. (676) (520) % % Arrest as juvenile 31.2*** 19.0 Multiple Mechanisms Arrest as adult 48.4*** 36.2 The study generated more––and Arrest as juvenile 56.5*** 42.5 more systematic––evidence that or adult for any the consequences of childhood crime victimization extend well beyond Arrest as juvenile 21.0* 15.6 childhood and adolescence, persist- or adult for any ing into young adulthood. Such vic- violent crime timization affects many functions * p≤ .05 **p≤ .01 ***p≤ .001 later in life, and what was revealed Note: Numbers in parentheses are numbers of cases. in this study most likely represents only the tip of the iceberg, which further research could bring to light. Table 2: Childhood Victimization and Later Psychopathology On the other hand, some expected outcomes (such as increased risk for Abuse/Neglect Group Control Group alcohol problems in abused and (676) (520) neglected children) did not materi- % % alize, raising questions for further Suicide attempt 18.8*** 7.7 study. Antisocial personality 18.4*** 11.2 Disentangling the pathways. disorder One of the difficulties in assessing Alcohol abuse/ 54.5*** 51.0 risk of negative consequences is dependence sorting out the children’s multiple *p≤ .05 **p≤ .01 ***p≤ .001 problems and those of their parents. Note: Numbers in parentheses are numbers of cases. As previous research has shown, Diagnoses of antisocial personality disorder and alcohol abuse/dependence were determined by using the adverse effects interact, so that the National Institute of Mental Health DIS-III-R diagnostic interview. combined effects of two types of problems may be greater than their sum.17 Whether this interaction Table 3: Childhood Victimization and effect applies to childhood victim- ization is not known, although it is Later Psychopathology, by Gender likely. Abuse/Neglect Group Control Group % % This study has not yet tried to Females (338) (224) distinguish among the many Suicide attempt 24.3*** 8.6 mechanisms by which childhood Antisocial personality 9.8* 4.9 victimization affects development disorder and psychopathology. When it Alcohol abuse/ 43.8** 32.8 dependence comes to the influence of contextual factors, children may simply be Males (338) (276) modeling their parents’ behavior. Suicide attempt 13.4** 6.9 Antisocial personality 27.0** 16.7 But it also is possible that abuse or disorder neglect may produce immediate Alcohol abuse/ 64.4 67.0 effects that then irremediably affect dependence subsequent development, which in * p≤ .05 **p≤ .01 ***p≤ .001 turn may affect still later outcomes. Note: Numbers in parentheses are numbers of cases. Diagnoses of antisocial personality disorder and alcohol abuse/dependence were determined by using the National Institute of Mental Health DIS-III-R diagnostic interview.

National Institute of Justice Journal ■ January 2000 7 Table 4: Antisocial Personality Disorder in Offspring— quences may be due not so much Relation to Parental Criminality to the abuse or neglect, but to the chain of events it triggers. Abuse/Neglect Control Row Group Group Significance No doubt there are many other % % mechanisms by which abuse and Either parent arrested 21.9 18.8 n.s. neglect affect a child. Hopefully, (365) (170) future models that explain long- Neither parent arrested 14.2 7.4 *** term consequences will examine (365) (350) some of them, because finding a single mechanism that explains all Column significance * *** cases of abuse and neglect is highly * p≤ .05 ** p≤ .01 *** p≤ .001 n.s. = not statistically significant. unlikely. Note: Numbers in parentheses are numbers of cases. Diagnoses of antisocial personality disorder and alcohol abuse/dependence were determined by using the NCJ 180077 National Institute of Mental Health DIS-III-R diagnostic interview. Notes Table 5: Alcohol Abuse/Dependence in Offspring— 1. This article summarizes the Relation to Parental Alcohol/Drug Problems author’s “Childhood Victimization: Early Adversity Abuse/Neglect Control Row Group Group Significance and Subsequent Psychopath- ology,” in Adversity, Stress, and % % Psychopathology, ed. B.P. Either parent alcohol/ 63.2 56.6 n.s. drug problem (389) (196) Dohrenwend, New York: Oxford University Press, 1998: 81–95. Neither parent alcohol/ 42.6 47.5 n.s. drug problem (284) (324) 2. Downey, G., et al., “Maltreat- Column significance *** * ment and Childhood Depres- sion,” in Handbook of Depression * p≤ .05 ** p≤ .01 *** p≤ .001 n.s. = not statistically significant. in Children, ed. W.M. Reynolds Note: Numbers in parentheses are numbers of cases. and H.F. Johnson, New York: Diagnoses of antisocial personality disorder and alcohol abuse/dependence were determined by using the Plenum, 1994; Dohrenwend, National Institute of Mental Health DIS-III-R diagnostic interview. B.P., and B.S. Dohrenwend, “Sex Differences in Psychiatric Direct and indirect pathways. school performance, less than ade- Disorders,” American Journal Some pathways may be direct—per- quate functioning on the job, or of Sociology 81 (1976):1447–54; sisting into adulthood. Abused and antisocial personality disorder. Horwitz, A.V., and H.R. White, neglected children may show aggres- Adaptations that might serve well “Gender Role Orientations and siveness and behavior problems in at one stage of development may Styles of Pathology Among childhood, delinquency in adoles- no longer do so at a later stage, Adolescents,” Journal of Health cence, and antisocial and criminal placing the person at risk for further and Social Behavior 28 behavior in adulthood. It also is unfavorable situations or subse- (1987):158–170; and Widom, likely that this path leads to abusive quent victimization that may C.S., “Sex Roles, Criminality, behavior in the home, manifested in trigger psychopathology. and Psychopathology,” in Sex spouse or child abuse. In other Roles and Psychopathology, ed. Some early, adverse experiences may instances there may be a delayed C.S. Widom, New York: Plenum, be indirect, creating byproducts. reaction, occurring years later. 1984: 87–213. They may change the environment Abuse or neglect may encourage or the family situation, which in 3. Downey et al., “Maltreatment certain dysfunctional ways of cop- turn may predispose a person to and Childhood Depression.” ing. An example is impulsive behav- problem behavior. They also may 4. Friedrich, W.H., A.J. Urquiza, ior that in turn gives rise to defi- expose the child to further harmful and R.L. Beilke, “Behavior ciencies in problem solving or in experiences. In this way, the conse-

Childhood Victimization: Early Adversity, Later Psychopathology 8 Problems in Sexually Abused atry 56 (1986):481–485; Reider, 13. Widom, C.S., “The Cycle of Young Children,” Journal of E.E., et al., “Alcohol Involvement Violence,” Science 244 Pediatric Psychology 11 and Violence Toward Children (1989):160–166. (1986):47–57. Among High-Risk Families,” 14. These findings, based on the Paper presented at the annual 5. Livingston, R., “Sexually and study of 1,196 of the original meeting of the American Physically Abused Children,” 1,575 subjects (the 908 Psychological Association, New Journal of the American abuse/neglect victims plus the Orleans, Louisiana, August Academy of Child and 667 in the control group), 11–15, 1989. Adolescent Psychiatry 26 should not be confused with (1987):413–415. 10. Goodwin, D.W., et al., “Alcohol findings from studies published Problems in Adoptees Raised previously (Widom, “Cycle of 6. Briere, J., and M. Runtz, Apart From Alcoholic Biological Violence,” and Maxfield, M.G., “Symptomatology Associated Parents,” Archives of General and C.S. Widom, “The Cycle of With Childhood Sexual Psychiatry 28 (1973):238–243; Violence: Revisited Six Years Victimization in a Nonclinical Goodwin, D.W., et al., “Alcohol- Later,” Archives of Pediatrics Adult Sample,” Child Abuse and ism and Depression in Adopted- and Adolescent Medicine 150 Neglect 12 (1988):51–60; Harris, Out Daughters of Alcoholics,” [1996]:390–395), which report T., G.W. Brown, and A. Bifulco, Archives of General Psychiatry on the entire original sample of “Loss of Parent in Childhood 34 (1977):751–755; Cloninger, 1,575. and Adult Psychiatric Disorder: C.R., et al., “Psychopathology in A Tentative Overall Model,” 15. See Widom, C.S., T. Ireland, and Adopted-Out Children of Development and Psychopath- P.J. Glynn, “Alcohol Abuse in Alcoholics: The Stockholm ology 2 (1990): 311–328; Terr, Abused and Neglected Children Adoption Study,” in Recent L.A., “Chowchilla Revisited: The Followed-Up: Are They at Developments in Alcoholism, Effects of Psychiatric Trauma Increased Risk?” Journal of Vol. 3, M. Galanter, ed., New Four Years After a School-Bus Studies on Alcohol 56 York: Plenum, 1985. Kidnaping,” American Journal (1995):207–217. of Psychiatry 140 (1983): 11. Wyatt, G.E., “Sexual Abuse of 16. These findings are not shown 1543–1550. Ethnic Minority Children: here. See Maxfield and Widom, Identifying Dimensions of 7. Terr, “Chowchila Revisited.” “The Cycle of Violence: Victimization,” Professional Revisited.” 8. Gibbin, P.T., R.H. Starr, and Psychology: Research and S.W. Agronow, “Affective Practice 21 (1990):338–343. 17. Rutter, M., “Protective Factors in Behavior of Abused and Children’s Response to Stress 12. Smith, C.P., D.J. Berkman, Controlled Children: and Disadvantage,” in Primary and W.M. Fraser, A Preliminary Comparison of Parent-Child Prevention of Psychopathology: National Assessment of Child Interactions and the Influence Social Competence in Children, Abuse and Neglect and the of Home Environment Var- Vol. 3, ed. M.V. Kent and J.E. Juvenile Justice System: The iables,” Journal of Genetic Rolf, Hanover, NH: New Shadows of Distress, Washington, Psychology 144 (1984):69–82. England Press, 1979: 49–74 . D.C.: U.S. Department of 9. Famularo, R., et al., “Alcoholism Justice: Office of Juvenile and Severe Child Maltreatment,” Justice and Delinquency American Journal of Orthopschi- Prevention, 1980.

National Institute of Justice Journal ■ January 2000 9