Psychological consequences of child maltreatment in homeless adolescents: untangling the unique effects of maltreatment and family environment

Abstract

Objective: This study examined the differential effects of various forms of abuse, as well as their combined effects. The study also sought to separate the factors uniquely associated with abuse from those associated with the more general problems present in an abusive family environment.

Methodology: Data were collected from 329 homeless adolescents. Preliminary analyses indicated some degree of matching for family background variables. Any differences in demographic or family characteristics were controlled for statistically.

Results: Chi-square analyses revealed significant differences across groups for rates of assault, rape, depression/dysthymia, and attempted suicide. Multivariate analyses (MANOVAs) indicated significant differences in severity of internalizing problems and cognitive problems. Without exception, the group with histories of both physical and sexual abuse exhibited the most severe symptomatology and was at greatest risk for revictimization. Multiple regression analyses suggested that abuse histories were predictive of internalizing problems while family characteristics were more predictive of externalizing problems.

Conclusions: The findings indicate that both abuse type and family characteristics contribute to the development of symptomatology. Future prospective and longitudinal studies are needed to clarify the sequelae of abuse, as well as the possible cause and effect relations between abuse, family characteristics, and psychological outcome.

Introduction

WHILE THERE HAS been substantial progress documenting the direct effects of selected factors in the past 30 years, the potential additive and interactive contributions in determining the consequences of various forms of child abuse remain obscure (National Research Council, 1993). As of yet, we cannot clearly discern the effects of physical abuse from those of sexual abuse or physical and sexual abuse combined, nor can we determine whether the effects are primarily due to associated family disruption. Part of the source of these difficulties appears to be difficulties with sampling strategies and with inadequate comparison groups. Common sampling strategies tend to exaggerate the effects of abuse, while the use of comparison groups with relatively few characteristics in common with the abused youth tend to downplay the effects of those associated characteristics Leventhal 1998, Rind et al 1998, Widom 1989. The challenge, it seems, is to find a sample of youth among whom abuse is commonplace yet is not the criteria for participation, and within which an adequate comparison group can be identified. The present study addressed this issue by considering distinct effects of sexual abuse, physical abuse, and combined physical and sexual abuse among homeless youth. The participation of homeless youth allows us to consider the unique effects of different forms of abuse among a sample of youth with high rates of abuse, as well as high levels of family disruption.

Child abuse occurs in families of all races and socioeconomic levels, however, certain segments of the population appear more likely to report child abuse histories or experiences. Homeless adolescents have been shown to report particularly high rates of physical and sexual abuse. Although incidence rates vary across samples, it is commonly reported that between one half and three quarters of homeless youth have been physically abused, sexually abused or both Warren et al 1994, Feitel et al 1992. A recent study of homeless youth in the Midwest reported that 23% of youth had been forced to engage in sexual behavior, and 43% had been beaten up by a caregiver (Whitbeck, Hoyt, & Ackley, 1997). Indeed, several studies have reported that one of the primary reasons that adolescents leave home is to escape parental abuse McCormack et al 1986, Whitbeck and Simons 1990. Moreover, even when abuse is not present, the homes of runaways are characterized by high levels of parent/child conflict (Warren et al., 1994), and low levels of parental emotional support Dadds et al 1993, Englander 1984.

A growing body of literature has addressed the consequences of abuse and shows that child maltreatment has many deleterious consequences for children and youth Aber and Allen 1987, Carlson et al 1989, Cryan 1985, Lamphear 1985, Leventhal 1998, Schneider-Rosen and Cicchetti 1991, Toth et al 1992. Research on the effects of sexual abuse reveals a wide variety of symptoms, with levels of distress ranging from asymptomatic to severe Browne and Finkelhor 1986, Conte and Berliner 1988, Finkelhor 1990, Finkelhor 1994, McLeer et al 1994. Individuals reporting a history of sexual abuse often exhibit internalizing problems Cryan 1985, Finkelhor 1990, Jaffe et al 1986, Ross and Faustini 1989. Externalizing behavior problems, including aggression, delinquency, inappropriate sexual behavior, and academic failure are also common Finkelhor 1990, Friedrich et al 1986, White et al 1988.

Physical abuse impacts a child across development (Hampton, Gullotta, Adams, Potter, & Weissberg, 1993). By middle childhood, physically abused youth are more likely to have affective and behavioral difficulties including anxiety, depression, self-destructive behavior, low self-esteem, social detachment, hyperactivity, excessive aggression, and noncompliance Egeland and Stroufe 1981, Kaufman and Cicchetti 1989. Physical abuse has been shown to be predictive of delinquency, adult criminality, and violent criminal behavior, as well as runaway behavior Kufeldt and Nimmo 1987, Widom 1989. While the most frequently cited behavioral sequelae following physical abuse is the prevalence of externalizing behavior problems, physically abused youth also show high rates of internalizing behavior problems as well (Kaufman & Cicchetti, 1989).

Though the literature examining maltreatment in homeless youth is neither extensive nor comprehensive (often comparing abused and non-abused in univariate analyses), it finds many of the same outcomes associated with maltreatment in the general population. Homeless youth with abuse histories are more likely to have internalizing problems like depression, suicidality and self-injurious behavior as well as externalizing problems McCormack et al 1986, Unger et al 1997. Indeed, the research conducted by Warren and colleagues (1994) points to abuse in the family of origin as pivotal in the runaway youth’s decision to leave home and seek refuge elsewhere. Perhaps most compelling is the suggestion by Whitbeck and Simons (1990) that a history of abuse predisposes youth to involvement with highly deviant peers, increasing the likelihood that they will engage in deviant subsistence strategies and increased, and prolonged, episodes of running away. These factors are implicated in the higher rates of on-street victimization suffered by youth who were abused during childhood Whitbeck et al 1997, Whitbeck and Simons 1990.

Our understanding of the effects of different forms of abuse is not as uncomplicated as it may appear. For instance, physically abused youth have been shown to be more likely to have problems with social cognition; showing less understanding of social roles, hostile attribution styles, and aggressive approaches to problem solving (Dodge, Bates, & Pettit, 1990). However, a study by Wolfe and Mosk (1983) found no differences in social and behavior problems for abused and nonabused youth from high distress families. These investigators posit that a child’s social and behavioral development may be more strongly influenced by general family interaction than it is by specific abuse episodes.

Similarly, Briere (1988) presents strong evidence for the idea that nonsexual abuse (physical, psychological, and emotional) and neglect are particularly important predictors of psychological outcome following a maltreatment situation involving both sexual and physical abuse. Indeed, his findings suggested that only 1–3% of the adjustment variance was accounted for by sexual abuse, while nonsexual abuse was shown to account for 13–24% of the adjustment variance. As of yet, research has not identified whether the consequences of types of abuse are due to the specific form of abuse, or due to associated stressors.

One method of examining the common and distinct consequences of different types of abuse has been to compare children who have suffered from physical abuse with those who have suffered from sexual abuse. Such studies suggest that there are both similar and differential effects. The commonality of the effects of physical and sexual abuse, including both internalizing and externalizing behaviors, has been attributed to the shared elements of powerlessness, terror, betrayal, and overstimulation experienced in abusive situations (Wind & Silvern, 1992). Less consensus exists with regard to the possible differential effects of physical and sexual abuse. Briere and Runtz (1990) reported unique relations between physical abuse and aggression, sexual abuse and sexual dysfunction, as well as between sexual abuse and acting out behaviors. These findings were not replicated, however, in a later study examining internalizing problems among women with histories of physical abuse, sexual abuse, or both forms of abuse (Wind & Silvern, 1992). Indeed, although having experienced both forms of abuse was associated with higher levels of symptomatology, the investigators found no differences between physically and sexually abused women (Wind & Silvern, 1992). Dykman and colleagues (1997) found that while externalizing scores were significantly higher than internalizing scores in the abused children they studied, there were no significant differences in behavioral consequences for children experiencing sexual abuse, physical abuse or both forms of abuse. These findings were further replicated in a population of psychiatrically hospitalized adolescents (Cohen et al., 1996). This later investigation revealed only one difference across abuse groups, with sexually abused youth using negative coping strategies more often than physically abused youth.

A recent study of homeless youth found several statistically significant relations between abuse type and psychopathology (Feitel et al., 1992). Youth who reported having been beaten in childhood and adolescence were more likely than those youth who did not report this form of physically abusive behavior to have a major affective disorder. They were also more likely to have attempted suicide than homeless youth who did not report being beaten. Sexually abused youth also reported higher rates of suicide attempts than those not reporting a sexually abusive experience (Feitel et al., 1992).

The difficulties encountered in reliably identifying the consequences of different forms of abuse is in part due to participant sampling strategies. The majority of the abuse literature reports on findings from research conducted with clinical and legal samples Briere 1992, National Research Council 1993. The use of such samples may exaggerate the effects of abuse and limits generalizability due to selection bias Leventhal 1998, National Research Council 1993, Rind and Tromovitch 1997, Widom 1989. Members of these samples not only self-select in that they exhibit or perceive higher levels of disturbance, but they are also often selected based on the disclosure of abuse or symptoms thought to be characteristic of abuse victims. Additionally, they rarely reflect the characteristics of the general population (Okami, 1991). Indeed, studies with college student participants have found that, although students who experienced sexual abuse present as less well adjusted than controls, the effect is often small (Rind et al., 1998). Although the use of college samples represents an improvement over the use of clinical samples, they may be only slightly more representative. For instance, it is likely that college students are fairly high functioning and that they have adequate to well developed coping skills and social skills; such characteristics may obscure differences among abused and nonabused students (Rind et al., 1998). The present investigation did not select youth for participation in the study based on the presence or absence of abuse. Adolescents were recruited into the study to assess the general characteristics and service needs of this population.

Another limitation of the existing research is that few studies have included appropriate comparison groups of nonabused persons. When comparison groups have been included, they have tended to consist of nonabused individuals drawn from clinical and legal samples (Rind & Tromovitch, 1997). These comparison groups do not adequately control for other environmental influences such as the dysfunctional family environment because, at best, matching occurs only on demographic, social status, and treatment history variables. As such, these studies may not convey an accurate picture of the relation between abuse and psychological symptomatology. Better comparison groups are necessary to separate the factors uniquely associated with abuse from those associated with the more general problems present in abusive families (i.e., family disruption, discord, and neglect).

Some evidence suggests that the symptomatology evident in many abuse victims maybe a function of the family disorganization, deprivation, and violence that a child experiences in an abusive household, rather than the direct result of the abuse (Herman & Hirschman, 1977). More recently, Nash, Hulsey, Sexton, Harralson, and Lambert (1993) were unable to detect any significant differences in adjustment for abused and control subjects after statistically controlling for family environment. These findings have been replicated in several studies using college samples Cole 1988, Higgins and McCabe 1994, Pallotta 1992, Wisniewski 1990. A disruptive family environment may impact child development and family interaction patterns leading to subsequent socioemotional and psychological difficulties Ge et al 1996, Simons et al 1991, Skinner et al 1992. For instance, rates of disruptive behavior disorders are significantly higher in families where parenting is aggressive and coercive (Patterson, 1992).

The present study examined the consequences of abuse within a homeless adolescent sample. Runaway and homeless youth are highly likely to be from dysfunctional and/or abusive families Adams et al 1985, Burgess et al 1987, Whitbeck and Simons 1993. In a study of Midwest homeless youth, Whitbeck and colleagues (1997) found that the parents of the youth, and the youth themselves, report parent-child relationships with little parental warmth and supportiveness, low levels of parental monitoring and high levels of parental rejection. In addition, many of the parents of homeless youth have alcohol problems. Moreover, these families have often had numerous contacts with the social service system, including the foster care system and residential treatment facilities. The portrait that emerges is one of family dysfunction with or without sexual exploitation and physical violence. Adolescents on the streets, therefore, are likely to be developmentally disadvantaged, and at increased risk for negative psychological outcomes. For example, one study (Yates, MacKenzie, Pennbridge, & Cohen, 1988) found a much higher rate of depression and attempted suicide among homeless and runaway youth, and another (Robertson, 1986) reported that nearly a quarter of homeless youth had received inpatient mental health treatment at some point in their life. Moreover, the family characteristics of these adolescents have often been associated with externalizing behavior problems Patterson 1986, Widom 1994. In using a homeless sample, the study achieved some degree of matching for family background variables, including family caring and emotional support, family upsetting influences, and mother criminal justice history. Any other differences were controlled for statistically.

Another significant limitation of the existing research examining the consequences of maltreatment has been the lack of focus on potential mediators such as age at first and last abuse, number of abusers, frequency of abuse, and severity of abuse. Recent investigations suggest that these, and other, abuse characteristics may play a significant role in the development of emotional and psychological problems following the abuse experience.

The present study builds on the previous research examining the consequences of maltreatment among homeless youth (Feitel et al 1992, McCormack et al 1986, Powers et al 1990; Stiffman, 1989a, 1989b) by fully examining the differential effects of both physical and sexual abuse, as well as their combined effects, and by taking into account family characteristics as well as characteristics of the abuse context. Given the unusually high base rates of abuse in this population as well as the large sample size, it was possible to compare three distinct groups of maltreated youth: (1) physically abused (PA), (2) sexually abused (SA), and (3) physically and sexually abused (PSA). A group reporting no abuse (NA) was also included.

It was hypothesized that homeless adolescents across all groups would show high levels of internalizing and externalizing problems, as well as difficulties with cognitive and social functioning. Furthermore, it was expected that adolescents reporting maltreatment would be at greater risk for negative outcomes than the non-maltreated, runaway youth; the result of cumulative and particularly severe family dysfunction. Sexual abuse was expected to lead to more internalizing problems, and physical abuse to more externalizing problems. The group of adolescents experiencing both physical and sexual abuse were expected to show the most severe psychopathology. Drawing from Patterson’s social interaction perspective, it was hypothesized that for maltreated youth externalizing outcomes would be related more significantly to family environment factors than would internalizing problems. Finally, it was hypothesized that early and multiple incidents of abuse would lead to more negative outcomes.